Datasets:

librarian-bots

/

arxiv-metadata-snapshot

like 3

2407.13153

Daniel Platnick

Daniel Platnick, Bishoy Abdelnour, Eamon Earl, Rahul Kumar, Zahra Rezaei, Thomas Tsangaris, Faraj Lagum

Preset-Voice Matching for Privacy Regulated Speech-to-Speech Translation Systems

Accepted to the ACL PrivateNLP 2024 Workshop, 7 pages, 2 figures

cs.CL cs.CR cs.LG cs.SD eess.AS

http://creativecommons.org/licenses/by/4.0/

In recent years, there has been increased demand for speech-to-speech translation (S2ST) systems in industry settings. Although successfully commercialized, cloning-based S2ST systems expose their distributors to liabilities when misused by individuals and can infringe on personality rights when exploited by media organizations. This work proposes a regulated S2ST framework called Preset-Voice Matching (PVM). PVM removes cross-lingual voice cloning in S2ST by first matching the input voice to a similar prior consenting speaker voice in the target-language. With this separation, PVM avoids cloning the input speaker, ensuring PVM systems comply with regulations and reduce risk of misuse. Our results demonstrate PVM can significantly improve S2ST system run-time in multi-speaker settings and the naturalness of S2ST synthesized speech. To our knowledge, PVM is the first explicitly regulated S2ST framework leveraging similarly-matched preset-voices for dynamic S2ST tasks.

2407.13154

Bethlee Lindor

Bethlee Lindor, Eric Agol

Modeling the Solar System as an Observed Multi-Transit System I: Characterization Limits from Analytic Timing Variations

19 pages, 15 figures

astro-ph.EP

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Planetary systems with multiple transiting planets are beneficial for understanding planet occurrence rates and system architectures. Although we have yet to find a solar system (SS) analog, future surveys may detect multiple terrestrial planets transiting a Sun-like star. In this work, we simulate transit timing observations of our Solar System as viewed from a distance and based on the actual orbital motions of Venus and the Earth-Moon Barycenter, as influenced by the other SS bodies, with varying noise levels and observing durations. We then retrieve the system's dynamical parameters using an approximate N-body model for transit timing shifts while considering four possible plane-parallel configurations: two planets, three planets, four planets, and five planets. We demonstrate that -- with the retrieval applied to simulated transit times of Venus and EMB -- we can: 1) detect Jupiter at high significance (up to 90-s timing noise); 2) measure the masses and orbits of both transiting planets (mass-ratios are down to 4-8% uncertainty for the 3-planet model) ; 3) detect Mars with more than $5\sigma$ given very high level precision (10s of seconds). To accurately characterize Jupiter, we require timing precisions of better than 30 seconds and survey durations longer than 22 years. Accurate retrieval of Mars is possible when the survey baseline is longer than 25 years. Additionally, while Jupiter's mass is underestimated in most of our simulated cases, the addition of Mars improves the posterior mass, suggesting that unseen terrestrials could interfere in the characterization of multi-planetary systems if they are nearly resonant to transiting planets. Ultimately, these simulations will help to guide future missions -- such as PLATO, Nautilus, LUVOIR, and Ariel -- in detecting and characterizing exoplanet systems analogous to our Solar System.

2407.13155

Yulin He

Yulin He, Wei Chen, Tianci Xun, Yusong Tan

Real-Time 3D Occupancy Prediction via Geometric-Semantic Disentanglement

cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Occupancy prediction plays a pivotal role in autonomous driving (AD) due to the fine-grained geometric perception and general object recognition capabilities. However, existing methods often incur high computational costs, which contradicts the real-time demands of AD. To this end, we first evaluate the speed and memory usage of most public available methods, aiming to redirect the focus from solely prioritizing accuracy to also considering efficiency. We then identify a core challenge in achieving both fast and accurate performance: \textbf{the strong coupling between geometry and semantic}. To address this issue, 1) we propose a Geometric-Semantic Dual-Branch Network (GSDBN) with a hybrid BEV-Voxel representation. In the BEV branch, a BEV-level temporal fusion module and a U-Net encoder is introduced to extract dense semantic features. In the voxel branch, a large-kernel re-parameterized 3D convolution is proposed to refine sparse 3D geometry and reduce computation. Moreover, we propose a novel BEV-Voxel lifting module that projects BEV features into voxel space for feature fusion of the two branches. In addition to the network design, 2) we also propose a Geometric-Semantic Decoupled Learning (GSDL) strategy. This strategy initially learns semantics with accurate geometry using ground-truth depth, and then gradually mixes predicted depth to adapt the model to the predicted geometry. Extensive experiments on the widely-used Occ3D-nuScenes benchmark demonstrate the superiority of our method, which achieves a 39.4 mIoU with 20.0 FPS. This result is $\sim 3 \times$ faster and +1.9 mIoU higher compared to FB-OCC, the winner of CVPR2023 3D Occupancy Prediction Challenge. Our code will be made open-source.

2407.13156

Pankaj Kumar

Pankaj Kumar, Neel Kanth Kundu, and Binayak Kar

Quantum Key Distribution Routing Protocol in Quantum Networks: Overview and Challenges

cs.CR

http://creativecommons.org/licenses/by/4.0/

The use of quantum cryptography in everyday applications has gained attention in both industrial and academic fields. Due to advancements in quantum electronics, practical quantum devices are already available in the market, and ready for wider use. Quantum Key Distribution (QKD) is a crucial aspect of quantum cryptography, which involves generating and distributing symmetric cryptographic keys between geographically separated users using principles of quantum physics. Many successful QKD networks have been established to test different solutions. The objective of this paper is to delve into the potential of utilizing established routing design techniques in the context of quantum key distribution, a field distinguished by its unique properties rooted in the principles of quantum mechanics. However, the implementation of these techniques poses substantial challenges, including quantum memory decoherence, key rate generation, latency delays, inherent noise in quantum systems, limited communication ranges, and the necessity for highly specialized hardware. This paper conducts an in-depth examination of essential research pertaining to the design methodologies for quantum key distribution. It also explores the fundamental aspects of quantum routing and the associated properties inherent to quantum QKD. This paper elucidates the necessary steps for constructing efficient and resilient QKD networks. In summarizing the techniques relevant to QKD networking and routing, including their underlying principles, protocols, and challenges, this paper sheds light on potential applications and delineates future research directions in this burgeoning field.

2407.13157

Jin Zhang

Jin Zhang and Ruiheng Zhang and Yanjiao Shi and Zhe Cao and Nian Liu and Fahad Shahbaz Khan

Learning Camouflaged Object Detection from Noisy Pseudo Label

Accepted by ECCV2024

cs.CV cs.AI

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Existing Camouflaged Object Detection (COD) methods rely heavily on large-scale pixel-annotated training sets, which are both time-consuming and labor-intensive. Although weakly supervised methods offer higher annotation efficiency, their performance is far behind due to the unclear visual demarcations between foreground and background in camouflaged images. In this paper, we explore the potential of using boxes as prompts in camouflaged scenes and introduce the first weakly semi-supervised COD method, aiming for budget-efficient and high-precision camouflaged object segmentation with an extremely limited number of fully labeled images. Critically, learning from such limited set inevitably generates pseudo labels with serious noisy pixels. To address this, we propose a noise correction loss that facilitates the model's learning of correct pixels in the early learning stage, and corrects the error risk gradients dominated by noisy pixels in the memorization stage, ultimately achieving accurate segmentation of camouflaged objects from noisy labels. When using only 20% of fully labeled data, our method shows superior performance over the state-of-the-art methods.

2407.13158

Qiuyu Zhu

Qiuyu Zhu, Liang Zhang, Qianxiong Xu, Kaijun Liu, Cheng Long, Xiaoyang Wang

HHGT: Hierarchical Heterogeneous Graph Transformer for Heterogeneous Graph Representation Learning

cs.LG cs.DB

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Despite the success of Heterogeneous Graph Neural Networks (HGNNs) in modeling real-world Heterogeneous Information Networks (HINs), challenges such as expressiveness limitations and over-smoothing have prompted researchers to explore Graph Transformers (GTs) for enhanced HIN representation learning. However, research on GT in HINs remains limited, with two key shortcomings in existing work: (1) A node's neighbors at different distances in HINs convey diverse semantics. Unfortunately, existing methods ignore such differences and uniformly treat neighbors within a given distance in a coarse manner, which results in semantic confusion. (2) Nodes in HINs have various types, each with unique semantics. Nevertheless, existing methods mix nodes of different types during neighbor aggregation, hindering the capture of proper correlations between nodes of diverse types. To bridge these gaps, we design an innovative structure named (k,t)-ring neighborhood, where nodes are initially organized by their distance, forming different non-overlapping k-ring neighborhoods for each distance. Within each k-ring structure, nodes are further categorized into different groups according to their types, thus emphasizing the heterogeneity of both distances and types in HINs naturally. Based on this structure, we propose a novel Hierarchical Heterogeneous Graph Transformer (HHGT) model, which seamlessly integrates a Type-level Transformer for aggregating nodes of different types within each k-ring neighborhood, followed by a Ring-level Transformer for aggregating different k-ring neighborhoods in a hierarchical manner. Extensive experiments are conducted on downstream tasks to verify HHGT's superiority over 14 baselines, with a notable improvement of up to 24.75% in NMI and 29.25% in ARI for node clustering task on the ACM dataset compared to the best baseline.

2407.13159

Bach Nguyen

Bach Nguyen Gia, Chanh Minh Tran, Kamioka Eiji, Tan Phan Xuan

Attenuation-Aware Weighted Optical Flow with Medium Transmission Map for Learning-based Visual Odometry in Underwater terrain

cs.CV

http://creativecommons.org/licenses/by/4.0/

This paper addresses the challenge of improving learning-based monocular visual odometry (VO) in underwater environments by integrating principles of underwater optical imaging to manipulate optical flow estimation. Leveraging the inherent properties of underwater imaging, the novel wflow-TartanVO is introduced, enhancing the accuracy of VO systems for autonomous underwater vehicles (AUVs). The proposed method utilizes a normalized medium transmission map as a weight map to adjust the estimated optical flow for emphasizing regions with lower degradation and suppressing uncertain regions affected by underwater light scattering and absorption. wflow-TartanVO does not require fine-tuning of pre-trained VO models, thus promoting its adaptability to different environments and camera models. Evaluation of different real-world underwater datasets demonstrates the outperformance of wflow-TartanVO over baseline VO methods, as evidenced by the considerably reduced Absolute Trajectory Error (ATE). The implementation code is available at: https://github.com/bachzz/wflow-TartanVO

2407.13160

Kirti Chandra Sahu Prof.

Someshwar Sanjay Ade, Pavan Kumar Kirar, Lakshmana Dora Chandrala and Kirti Chandra Sahu

Droplet breakup and size distribution in an airstream -- effect of inertia

23 pages, 17 figures

physics.flu-dyn

http://creativecommons.org/licenses/by/4.0/

We experimentally investigate the morphology and breakup of a droplet as it descends freely from a height and encounters an airstream. The size distributions of the child droplets are analysed using high-speed shadowgraphy and in-line holography techniques. We found that a droplet falling from various heights exhibits shape oscillations due to the intricate interplay between inertia and surface tension forces, leading to significant variations in the radial deformation of the droplet, influencing the breakup dynamics under an identical airstream condition. Specifically, the droplet undergoes vibrational breakup when introduced at a location slightly above the air nozzle. In contrast, as the release height of the droplet increases, keeping the Weber number defined based on the velocity of the airstream fixed, a dynamic interplay between the inertia of the droplet and the aerodynamic flow field comes into play, resulting in a sequence of breakup modes transitioning from vibrational breakup to retracting bag breakup, bag breakup, bag-stamen, retracting bag-stamen breakup, and eventually returning to vibrational breakup. Our experiments also reveal that the size distribution resulting from retracting bag breakup primarily arises from rim and node fragmentation, leading to a bimodal distribution. In contrast, bag and bag-stamen breakups yield a tri-modal size distribution due to the combined contributions of bag, rim, and node breakup mechanisms. Furthermore, we utilize a theoretical model that incorporates the effective Weber number, considering different release heights. This model accurately predicts the size distribution of the child droplets resulting from the various breakup modes observed in our experiments.

2407.13161

Min Xia

Min Xia, Zhu Su, Weibing Deng, Xiumei Feng and Benwei Zhang

How to quantify an examination? Evidence from physics examinations via complex networks

physics.soc-ph cs.CY physics.ed-ph

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Given the untapped potential for continuous improvement of examinations, quantitative investigations of examinations could guide efforts to considerably improve learning efficiency and evaluation and thus greatly help both learners and educators. However, there is a general lack of quantitative methods for investigating examinations. To address this gap, we propose a new metric via complex networks; i.e., the knowledge point network (KPN) of an examination is constructed by representing the knowledge points (concepts, laws, etc.) as nodes and adding links when these points appear in the same question. Then, the topological quantities of KPNs, such as degree, centrality, and community, can be employed to systematically explore the structural properties and evolution of examinations. In this work, 35 physics examinations from the NCEE examination spanning from 2006 to 2020 were investigated as an evidence. We found that the constructed KPNs are scale-free networks that show strong assortativity and small-world effects in most cases. The communities within the KPNs are obvious, and the key nodes are mainly related to mechanics and electromagnetism. Different question types are related to specific knowledge points, leading to noticeable structural variations in KPNs. Moreover, changes in the KPN topology between examinations administered in different years may offer insights guiding college entrance examination reforms. Based on topological quantities such as the average degree, network density, average clustering coefficient, and network transitivity, the Fd is proposed to evaluate examination difficulty. All the above results show that our approach can comprehensively quantify the knowledge structures and examination characteristics. These networks may elucidate comprehensive examination knowledge graphs for educators and guide improvements in teaching.

2407.13162

Ali A. Nazari

Ali A. Nazari, Jeremy Catania, Soroush Sadeghian, Amir Jalali, Houman Masnavi, Farrokh Janabi-Sharifi, Kourosh Zareinia

A Master-Follower Teleoperation System for Robotic Catheterization: Design, Characterization, and Tracking Control

24 pages, 7 figures, 6 tables

cs.RO cs.CV cs.SY eess.SY

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Minimally invasive robotic surgery has gained significant attention over the past two decades. Telerobotic systems, combined with robot-mediated minimally invasive techniques, have enabled surgeons and clinicians to mitigate radiation exposure for medical staff and extend medical services to remote and hard-to-reach areas. To enhance these services, teleoperated robotic surgery systems incorporating master and follower devices should offer transparency, enabling surgeons and clinicians to remotely experience a force interaction similar to the one the follower device experiences with patients' bodies. This paper presents the design and development of a three-degree-of-freedom master-follower teleoperated system for robotic catheterization. To resemble manual intervention by clinicians, the follower device features a grip-insert-release mechanism to eliminate catheter buckling and torsion during operation. The bidirectionally navigable ablation catheter is statically characterized for force-interactive medical interventions. The system's performance is evaluated through approaching and open-loop path tracking over typical circular, infinity-like, and spiral paths. Path tracking errors are presented as mean Euclidean error (MEE) and mean absolute error (MAE). The MEE ranges from 0.64 cm (infinity-like path) to 1.53 cm (spiral path). The MAE also ranges from 0.81 cm (infinity-like path) to 1.92 cm (spiral path). The results indicate that while the system's precision and accuracy with an open-loop controller meet the design targets, closed-loop controllers are necessary to address the catheter's hysteresis and dead zone, and system nonlinearities.

2407.13163

Yi Zhang

Yi Zhang, Ruihong Qiu, Jiajun Liu and Sen Wang

ROLeR: Effective Reward Shaping in Offline Reinforcement Learning for Recommender Systems

CIKM 2024

cs.IR cs.AI

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Offline reinforcement learning (RL) is an effective tool for real-world recommender systems with its capacity to model the dynamic interest of users and its interactive nature. Most existing offline RL recommender systems focus on model-based RL through learning a world model from offline data and building the recommendation policy by interacting with this model. Although these methods have made progress in the recommendation performance, the effectiveness of model-based offline RL methods is often constrained by the accuracy of the estimation of the reward model and the model uncertainties, primarily due to the extreme discrepancy between offline logged data and real-world data in user interactions with online platforms. To fill this gap, a more accurate reward model and uncertainty estimation are needed for the model-based RL methods. In this paper, a novel model-based Reward Shaping in Offline Reinforcement Learning for Recommender Systems, ROLeR, is proposed for reward and uncertainty estimation in recommendation systems. Specifically, a non-parametric reward shaping method is designed to refine the reward model. In addition, a flexible and more representative uncertainty penalty is designed to fit the needs of recommendation systems. Extensive experiments conducted on four benchmark datasets showcase that ROLeR achieves state-of-the-art performance compared with existing baselines. The source code can be downloaded at https://github.com/ArronDZhang/ROLeR.

2407.13164

Pengcheng Huang

Pengcheng Huang and Yongyu Mu and Yuzhang Wu and Bei Li and Chunyang Xiao and Tong Xiao and Jingbo Zhu

Translate-and-Revise: Boosting Large Language Models for Constrained Translation

16 pages

cs.CL cs.AI

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Imposing constraints on machine translation systems presents a challenging issue because these systems are not trained to make use of constraints in generating adequate, fluent translations. In this paper, we leverage the capabilities of large language models (LLMs) for constrained translation, given that LLMs can easily adapt to this task by taking translation instructions and constraints as prompts. However, LLMs cannot always guarantee the adequacy of translation, and, in some cases, ignore the given constraints. This is in part because LLMs might be overly confident in their predictions, overriding the influence of the constraints. To overcome this overiding behaviour, we propose to add a revision process that encourages LLMs to correct the outputs by prompting them about the constraints that have not yet been met. We evaluate our approach on four constrained translation tasks, encompassing both lexical and structural constraints in multiple constraint domains. Experiments show 15\% improvement in constraint-based translation accuracy over standard LLMs and the approach also significantly outperforms neural machine translation (NMT) state-of-the-art methods.

2407.13165

William Erickson

William Q. Erickson

The Demazure product extended to biwords

18 pages

math.CO

http://creativecommons.org/licenses/by/4.0/

The symmetric group $\mathfrak{S}_n$ (and more generally, any Coxeter group) admits an associative operation known as the Demazure product. In this paper, we first extend the Demazure product to the (infinite) set of all biwords on $\{1, \ldots, n\}$, or equivalently, the set of all $n \times n$ nonnegative integer matrices. We define this product diagrammatically, via braid-like graphs we call kelp beds, since they significantly generalize the seaweeds introduced by Tiskin (2015). Our motivation for this extended Demazure product arises from optimization theory, in particular the semigroup of all $(n+1) \times (n+1)$ simple nonnegative integer Monge matrices equipped with the distance (i.e., min-plus) product. As our main result, we show that this semigroup of Monge matrices is isomorphic to the semigroup of biwords equipped with the extended Demazure product. We exploit this isomorphism to write down generating functions for the growth series of the Monge matrices with respect to certain natural matrix norms.

2407.13166

Hyunwoo Kim

Hyunwoo Kim, Yoonseo Choi, Taehyun Yang, Honggu Lee, Chaneon Park, Yongju Lee, Jin Young Kim, Juho Kim

Using LLMs to Investigate Correlations of Conversational Follow-up Queries with User Satisfaction

Accepted to LLM4Eval @ SIGIR 2024 - The First Workshop on Large Language Models (LLMs) for Evaluation in Information Retrieval

cs.HC cs.IR

http://creativecommons.org/licenses/by/4.0/

With large language models (LLMs), conversational search engines shift how users retrieve information from the web by enabling natural conversations to express their search intents over multiple turns. Users' natural conversation embodies rich but implicit signals of users' search intents and evaluation of search results to understand user experience with the system. However, it is underexplored how and why users ask follow-up queries to continue conversations with conversational search engines and how the follow-up queries signal users' satisfaction. From qualitative analysis of 250 conversational turns from an in-lab user evaluation of Naver Cue:, a commercial conversational search engine, we propose a taxonomy of 18 users' follow-up query patterns from conversational search, comprising two major axes: (1) users' motivations behind continuing conversations (N = 7) and (2) actions of follow-up queries (N = 11). Compared to the existing literature on query reformulations, we uncovered a new set of motivations and actions behind follow-up queries, including asking for subjective opinions or providing natural language feedback on the engine's responses. To analyze conversational search logs with our taxonomy in a scalable and efficient manner, we built an LLM-powered classifier (73% accuracy). With our classifier, we analyzed 2,061 conversational tuples collected from real-world usage logs of Cue: and examined how the conversation patterns from our taxonomy correlates with satisfaction. Our initial findings suggest some signals of dissatisfactions, such as Clarifying Queries, Excluding Condition, and Substituting Condition with follow-up queries. We envision our approach could contribute to automated evaluation of conversation search experience by providing satisfaction signals and grounds for realistic user simulations.

2407.13167

Hiroshi Okada

Takaaki Nomura, Hiroshi Okada

Quasi two-zero texture in Type-II seesaw at fixed points from modular $A_4$ symmetry

14 pages, 6 figures, 1 table

hep-ph

http://creativecommons.org/publicdomain/zero/1.0/

We study a quasi two-zero neutrino texture based on type-II seesaw model in a modular $A_4$ symmetry to evade the cosmological bound on the sum of neutrino mass while keeping some predictability in neutrino sector. Working on three fixed points for modulus, we discuss predictions of the model and show the allowed points satisfying the cosmological bound on neutrino mass from both CMB and CMB+BAO data.

2407.13168

Minyang Tian

Minyang Tian, Luyu Gao, Shizhuo Dylan Zhang, Xinan Chen, Cunwei Fan, Xuefei Guo, Roland Haas, Pan Ji, Kittithat Krongchon, Yao Li, Shengyan Liu, Di Luo, Yutao Ma, Hao Tong, Kha Trinh, Chenyu Tian, Zihan Wang, Bohao Wu, Yanyu Xiong, Shengzhu Yin, Minhui Zhu, Kilian Lieret, Yanxin Lu, Genglin Liu, Yufeng Du, Tianhua Tao, Ofir Press, Jamie Callan, Eliu Huerta, Hao Peng

SciCode: A Research Coding Benchmark Curated by Scientists

25 pages, 9 figures, 7 tables

cs.AI cs.CL

http://creativecommons.org/licenses/by/4.0/

Since language models (LMs) now outperform average humans on many challenging tasks, it has become increasingly difficult to develop challenging, high-quality, and realistic evaluations. We address this issue by examining LMs' capabilities to generate code for solving real scientific research problems. Incorporating input from scientists and AI researchers in 16 diverse natural science sub-fields, including mathematics, physics, chemistry, biology, and materials science, we created a scientist-curated coding benchmark, SciCode. The problems in SciCode naturally factorize into multiple subproblems, each involving knowledge recall, reasoning, and code synthesis. In total, SciCode contains 338 subproblems decomposed from 80 challenging main problems. It offers optional descriptions specifying useful scientific background information and scientist-annotated gold-standard solutions and test cases for evaluation. Claude3.5-Sonnet, the best-performing model among those tested, can solve only 4.6% of the problems in the most realistic setting. We believe that SciCode demonstrates both contemporary LMs' progress towards becoming helpful scientific assistants and sheds light on the development and evaluation of scientific AI in the future.

2407.13169

John Yannotty

John C. Yannotty, Thomas J. Santner, Bo Li, Matthew T. Pratola

Combining Climate Models using Bayesian Regression Trees and Random Paths

52 pages, 18 figures

stat.ME

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Climate models, also known as general circulation models (GCMs), are essential tools for climate studies. Each climate model may have varying accuracy across the input domain, but no single model is uniformly better than the others. One strategy to improving climate model prediction performance is to integrate multiple model outputs using input-dependent weights. Along with this concept, weight functions modeled using Bayesian Additive Regression Trees (BART) were recently shown to be useful for integrating multiple Effective Field Theories in nuclear physics applications. However, a restriction of this approach is that the weights could only be modeled as piecewise constant functions. To smoothly integrate multiple climate models, we propose a new tree-based model, Random Path BART (RPBART), that incorporates random path assignments into the BART model to produce smooth weight functions and smooth predictions of the physical system, all in a matrix-free formulation. The smoothness feature of RPBART requires a more complex prior specification, for which we introduce a semivariogram to guide its hyperparameter selection. This approach is easy to interpret, computationally cheap, and avoids an expensive cross-validation study. Finally, we propose a posterior projection technique to enable detailed analysis of the fitted posterior weight functions. This allows us to identify a sparse set of climate models that can largely recover the underlying system within a given spatial region as well as quantifying model discrepancy within the model set under consideration. Our method is demonstrated on an ensemble of 8 GCMs modeling the average monthly surface temperature.

2407.13170

Eashan Adhikarla

Eashan Adhikarla, Kai Zhang, Rosaura G. VidalMata, Manjushree Aithal, Nikhil Ambha Madhusudhana, John Nicholson, Lichao Sun, Brian D. Davison

Unified-EGformer: Exposure Guided Lightweight Transformer for Mixed-Exposure Image Enhancement

Under submission

cs.CV cs.AI

http://creativecommons.org/licenses/by-sa/4.0/

Despite recent strides made by AI in image processing, the issue of mixed exposure, pivotal in many real-world scenarios like surveillance and photography, remains inadequately addressed. Traditional image enhancement techniques and current transformer models are limited with primary focus on either overexposure or underexposure. To bridge this gap, we introduce the Unified-Exposure Guided Transformer (Unified-EGformer). Our proposed solution is built upon advanced transformer architectures, equipped with local pixel-level refinement and global refinement blocks for color correction and image-wide adjustments. We employ a guided attention mechanism to precisely identify exposure-compromised regions, ensuring its adaptability across various real-world conditions. U-EGformer, with a lightweight design featuring a memory footprint (peak memory) of only $\sim$1134 MB (0.1 Million parameters) and an inference time of 95 ms (9.61x faster than the average), is a viable choice for real-time applications such as surveillance and autonomous navigation. Additionally, our model is highly generalizable, requiring minimal fine-tuning to handle multiple tasks and datasets with a single architecture.

2407.13171

Mashbat Suzuki

Sirin Botan and Angus Ritossa and Mashbat Suzuki and Toby Walsh

Maximin Fair Allocation of Indivisible Items under Cost Utilities

Appeared in SAGT 2023

cs.GT

http://creativecommons.org/licenses/by/4.0/

We study the problem of fairly allocating indivisible goods among a set of agents. Our focus is on the existence of allocations that give each agent their maximin fair share--the value they are guaranteed if they divide the goods into as many bundles as there are agents, and receive their lowest valued bundle. An MMS allocation is one where every agent receives at least their maximin fair share. We examine the existence of such allocations when agents have cost utilities. In this setting, each item has an associated cost, and an agent's valuation for an item is the cost of the item if it is useful to them, and zero otherwise. Our main results indicate that cost utilities are a promising restriction for achieving MMS. We show that for the case of three agents with cost utilities, an MMS allocation always exists. We also show that when preferences are restricted slightly further--to what we call laminar set approvals--we can guarantee MMS allocations for any number of agents. Finally, we explore if it is possible to guarantee each agent their maximin fair share while using a strategyproof mechanism.

2407.13172

Worawat Meevasana

Chutchawan Jaisuk, Tanawat Sawasdee, Warakorn Jindata, Thanundon Kongnok, Sirichok Jungthawan, Atsushi Fujimori and Worawat Meevasana

Chemical Potential Shift in Doped Mott-insulators for Energy Storage Applications

27 pages, 4 figures

cond-mat.str-el

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

This work explores the unique character of strongly correlated systems, specifically Mott-insulators, in the context of battery electrode materials. The study investigates the correlation between the proposed chemical potential evolution and charge storage performance in transition metal oxide-based electrodes. The hypothesis suggests that doping a Mott insulator reduces the Hubbard Coulomb interaction, which could slow down the chemical shift and result in enhanced charge storage capabilities compared to classic band insulators. The results support the hypothesis through a systematic comparison of selected transition metal oxide-based electrodes (Cu, Mn, Co, and Fe oxide electrodes). Furthermore, a toy model is employed to investigate the shift in chemical potential with doping-dependent U using DFT+U calculation, aiming to visualize the chemical potential evolution in Mott-insulators relevant to their application as battery electrodes. This study provides valuable insights into how strongly correlated materials, especially Mott-insulators, contribute to the advancement of energy storage technologies.

2407.13173

Litong Xu

Litong Xu and Tingting Xi

High-energy tunable ultraviolet pulses generated by optical leaky wave in filamentation

physics.optics

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Ultraviolet pulses could open up new opportunities for the study of strong-field physics and ultrafast science. However, the existing methods for generating ultra-violet pulses face difficulties in fulfilling the twofold requirements of high energy and wavelength tunability simultaneously. Here, we theoretically demonstrate the generation of high-energy and wavelength tunable ultraviolet pulses in preformed air-plasma channels via the leaky wave emission. The output ultraviolet pulse has a tunable wavelength ranging from 250 nm to 430 nm and an energy level up to sub-mJ. An octave-spanning ultraviolet supercontinuum with a flatness better than 3 dB can also be obtained via longitudinally modulated dispersion. Such a high-energy tunable ultraviolet light source may provide promising opportunities for characterization of ultrafast phenomena such as molecular breakup, and also an important driving source for the generation of high-energy attosecond pulses.

2407.13174

Rohit Raj Rai

Rohit Raj Rai, Rishant Pal, Amit Awekar

Compressed models are NOT miniature versions of large models

Accepted at the 33rd ACM International Conference on Information and Knowledge Management (CIKM 2024) for the Short Research Paper track, 5 pages

cs.LG cs.IR

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Large neural models are often compressed before deployment. Model compression is necessary for many practical reasons, such as inference latency, memory footprint, and energy consumption. Compressed models are assumed to be miniature versions of corresponding large neural models. However, we question this belief in our work. We compare compressed models with corresponding large neural models using four model characteristics: prediction errors, data representation, data distribution, and vulnerability to adversarial attack. We perform experiments using the BERT-large model and its five compressed versions. For all four model characteristics, compressed models significantly differ from the BERT-large model. Even among compressed models, they differ from each other on all four model characteristics. Apart from the expected loss in model performance, there are major side effects of using compressed models to replace large neural models.

2407.13175

Li Meng

Li Meng, Zhao Qi, Lyu Shuchang, Wang Chunlei, Ma Yujing, Cheng Guangliang, Yang Chenguang

OVGNet: A Unified Visual-Linguistic Framework for Open-Vocabulary Robotic Grasping

Accepted in IROS2024

cs.RO

http://creativecommons.org/publicdomain/zero/1.0/

Recognizing and grasping novel-category objects remains a crucial yet challenging problem in real-world robotic applications. Despite its significance, limited research has been conducted in this specific domain. To address this, we seamlessly propose a novel framework that integrates open-vocabulary learning into the domain of robotic grasping, empowering robots with the capability to adeptly handle novel objects. Our contributions are threefold. Firstly, we present a large-scale benchmark dataset specifically tailored for evaluating the performance of open-vocabulary grasping tasks. Secondly, we propose a unified visual-linguistic framework that serves as a guide for robots in successfully grasping both base and novel objects. Thirdly, we introduce two alignment modules designed to enhance visual-linguistic perception in the robotic grasping process. Extensive experiments validate the efficacy and utility of our approach. Notably, our framework achieves an average accuracy of 71.2\% and 64.4\% on base and novel categories in our new dataset, respectively.

2407.13176

Yixiao Ge Mr.

Yixiao Ge, Behzad Zamani, Pieter van Goor, Jochen Trumpf, Robert Mahony

Geometric Data Fusion for Collaborative Attitude Estimation

To be presented at IFAC MTNS 2024

eess.SY cs.SY

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

In this paper, we consider the collaborative attitude estimation problem for a multi-agent system. The agents are equipped with sensors that provide directional measurements and relative attitude measurements. We present a bottom-up approach where each agent runs an extended Kalman filter (EKF) locally using directional measurements and augments this with relative attitude measurements provided by neighbouring agents. The covariance estimates of the relative attitude measurements are geometrically corrected to compensate for relative attitude between the agent that makes the measurement and the agent that uses the measurement before being fused with the local estimate using the convex combination ellipsoid (CCE) method to avoid data incest. Simulations are undertaken to numerically evaluate the performance of the proposed algorithm.

2407.13177

Junho Hong

Nikolay Bobev, Sunjin Choi, Junho Hong, and Valentin Reys

Superconformal Indices of 3d $\mathcal{N}=2$ SCFTs and Holography

v1: 31 pages + appendices

hep-th

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

We study the superconformal index of 3d $\mathcal{N}=2$ superconformal field theories on $S^1\times_{\omega} S^2$ in the Cardy-like limit where the radius of the $S^1$ is much smaller than that of the $S^2$. We show that the first two leading terms in this Cardy-like expansion are dictated by the Bethe Ansatz formulation of the topologically twisted index of the same theory. We apply this relation to 3d $\mathcal{N}=2$ holographic superconformal field theories describing the low-energy dynamics of $N$ M2-branes and derive closed form expressions, valid to all orders in the $1/N$ expansion, for the two leading terms in the Cardy-like expansion of the superconformal index. We also discuss the implications of our results for the entropy of supersymmetric Kerr-Newman black holes in AdS$_4$ and the four-derivative corrections to 4d gauged supergravity.

2407.13178

Zeinab Sedaghatjoo

Zeinab Sedaghatjoo, Hossein Hosseinzadeh

The use of the symmetric finite difference in the local binary pattern (symmetric LBP)

cs.CV cs.NA math.NA

http://creativecommons.org/licenses/by/4.0/

The paper provides a mathematical view to the binary numbers presented in the Local Binary Pattern (LBP) feature extraction process. Symmetric finite difference is often applied in numerical analysis to enhance the accuracy of approximations. Then, the paper investigates utilization of the symmetric finite difference in the LBP formulation for face detection and facial expression recognition. It introduces a novel approach that extends the standard LBP, which typically employs eight directional derivatives, to incorporate only four directional derivatives. This approach is named symmetric LBP. The number of LBP features is reduced to 16 from 256 by the use of the symmetric LBP. The study underscores the significance of the number of directions considered in the new approach. Consequently, the results obtained emphasize the importance of the research topic.

2407.13179

Peibei Cao

Peibei Cao, Haoyu Chen, Jingzhe Ma, Yu-Chieh Yuan, Zhiyong Xie, Xin Xie, Haiqing Bai, and Kede Ma

Learned HDR Image Compression for Perceptually Optimal Storage and Display

eess.IV cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

High dynamic range (HDR) capture and display have seen significant growth in popularity driven by the advancements in technology and increasing consumer demand for superior image quality. As a result, HDR image compression is crucial to fully realize the benefits of HDR imaging without suffering from large file sizes and inefficient data handling. Conventionally, this is achieved by introducing a residual/gain map as additional metadata to bridge the gap between HDR and low dynamic range (LDR) images, making the former compatible with LDR image codecs but offering suboptimal rate-distortion performance. In this work, we initiate efforts towards end-to-end optimized HDR image compression for perceptually optimal storage and display. Specifically, we learn to compress an HDR image into two bitstreams: one for generating an LDR image to ensure compatibility with legacy LDR displays, and another as side information to aid HDR image reconstruction from the output LDR image. To measure the perceptual quality of output HDR and LDR images, we use two recently proposed image distortion metrics, both validated against human perceptual data of image quality and with reference to the uncompressed HDR image. Through end-to-end optimization for rate-distortion performance, our method dramatically improves HDR and LDR image quality at all bit rates.

2407.13180

Shahar Hod

Shahar Hod

Survival probabilities in the Sisyphus random walk model with absorbing traps

11 pages

Annals of Physics 406, 200-206 (2019)

cond-mat.stat-mech

http://creativecommons.org/licenses/by/4.0/

We analyze the dynamics of the Sisyphus random walk model, a discrete Markov chain in which the walkers may randomly return to their initial position $x_0$. In particular, we present a remarkably compact derivation of the time-dependent survival probability function $S(t;x_0)$ which characterizes the random walkers in the presence of an absorbing trap at the origin. The survival probabilities are expressed in a compact mathematical form in terms of the $x_0$-generalized Fibonacci-like numbers $G^{(x_0)}_t$. Interestingly, it is proved that, as opposed to the standard random walk model in which the survival probabilities depend linearly on the initial distance $x_0$ of the walkers from the trap and decay asymptotically as an inverse power of the time, in the Sisyphus random walk model the asymptotic survival probabilities decay exponentially in time and are characterized by a non-trivial (non-linear) dependence on the initial gap $x_0$ from the absorbing trap. We use the analytically derived results in order to analyze the underlying dynamics of the `survival-game', a highly risky investment strategy in which non-absorbed agents receive a reward $P(t)$ which gradually increases in time.

2407.13181

Xuanhua He

Xuanhua He, Lang Li, Yingying Wang, Hui Zheng, Ke Cao, Keyu Yan, Rui Li, Chengjun Xie, Jie Zhang, Man Zhou

Training-Free Large Model Priors for Multiple-in-One Image Restoration

cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Image restoration aims to reconstruct the latent clear images from their degraded versions. Despite the notable achievement, existing methods predominantly focus on handling specific degradation types and thus require specialized models, impeding real-world applications in dynamic degradation scenarios. To address this issue, we propose Large Model Driven Image Restoration framework (LMDIR), a novel multiple-in-one image restoration paradigm that leverages the generic priors from large multi-modal language models (MMLMs) and the pretrained diffusion models. In detail, LMDIR integrates three key prior knowledges: 1) global degradation knowledge from MMLMs, 2) scene-aware contextual descriptions generated by MMLMs, and 3) fine-grained high-quality reference images synthesized by diffusion models guided by MMLM descriptions. Standing on above priors, our architecture comprises a query-based prompt encoder, degradation-aware transformer block injecting global degradation knowledge, content-aware transformer block incorporating scene description, and reference-based transformer block incorporating fine-grained image priors. This design facilitates single-stage training paradigm to address various degradations while supporting both automatic and user-guided restoration. Extensive experiments demonstrate that our designed method outperforms state-of-the-art competitors on multiple evaluation benchmarks.

2407.13182

Wenwen Min

Xiaoyu Li, Fangfang Zhu and Wenwen Min

SpaDiT: Diffusion Transformer for Spatial Gene Expression Prediction using scRNA-seq

cs.LG cs.AI q-bio.GN

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

The rapid development of spatial transcriptomics (ST) technologies is revolutionizing our understanding of the spatial organization of biological tissues. Current ST methods, categorized into next-generation sequencing-based (seq-based) and fluorescence in situ hybridization-based (image-based) methods, offer innovative insights into the functional dynamics of biological tissues. However, these methods are limited by their cellular resolution and the quantity of genes they can detect. To address these limitations, we propose SpaDiT, a deep learning method that utilizes a diffusion generative model to integrate scRNA-seq and ST data for the prediction of undetected genes. By employing a Transformer-based diffusion model, SpaDiT not only accurately predicts unknown genes but also effectively generates the spatial structure of ST genes. We have demonstrated the effectiveness of SpaDiT through extensive experiments on both seq-based and image-based ST data. SpaDiT significantly contributes to ST gene prediction methods with its innovative approach. Compared to eight leading baseline methods, SpaDiT achieved state-of-the-art performance across multiple metrics, highlighting its substantial bioinformatics contribution.

2407.13183

Abhijith Beeravolu Reddy

Abhijith R. Beeravolu, Ian Brent Masters, Mirjam Jonkman, Kheng Cher Yeo, Spyridon Prountzos, Rahul J Thomas, Eva Ignatious, Sami Azam, Gabrielle B McCallum, Efthymia Alexopoulou, Anne B Chang, Friso De Boer

Methods to Measure the Broncho-Arterial Ratio and Wall Thickness in the Right Lower Lobe for Defining Radiographic Reversibility of Bronchiectasis

14 pages

eess.IV cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

The diagnosis of bronchiectasis requires measuring abnormal bronchial dilation. It is confirmed using a chest CT scan, where the key feature is an increased broncho-arterial ratio (BAR) (>0.8 in children), often with bronchial wall thickening. Image processing methods facilitate quicker interpretation and detailed evaluations by lobes and segments. Challenges like inclined nature, oblique orientation, and partial volume effect make it difficult to obtain accurate measurements in the upper and middle lobes using the same algorithms. Therefore, accurate detection and measurement of airway and artery regions for BAR and wall thickness in each lobe require different image processing/machine learning methods. We propose methods for: 1. Separating the right lower lobe (RLL) region from full-length CT scans using the tracheal bifurcation (Carina) point as a central marker; 2. Locating the inner diameter of airways and outer diameter of arteries for BAR measurement; and 3. Measuring airway wall thickness (WT) by identifying the outer and inner diameters of airway boundaries. Analysis of 13 HRCT scans with varying thicknesses (0.67mm, 1mm, 2mm) shows the tracheal bifurcation frame can be detected accurately, with a deviation of +/- 2 frames in some cases. A Windows app was developed for measuring inner airway diameter, artery diameter, BAR, and wall thickness, allowing users to draw boundaries around visible BA pairs in the RLL region. Measurements of 10 BA pairs revealed accurate results comparable to those of a human reader, with deviations of +/- 0.10-0.15mm. Additional studies and validation are needed to consolidate inter- and intra-rater variability and enhance the methods.

2407.13184

Andrey Savchenko

Andrey V. Savchenko

HSEmotion Team at the 7th ABAW Challenge: Multi-Task Learning and Compound Facial Expression Recognition

16 pages, 7 figures, 9 tables

cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

In this paper, we describe the results of the HSEmotion team in two tasks of the seventh Affective Behavior Analysis in-the-wild (ABAW) competition, namely, multi-task learning for simultaneous prediction of facial expression, valence, arousal, and detection of action units, and compound expression recognition. We propose an efficient pipeline based on frame-level facial feature extractors pre-trained in multi-task settings to estimate valence-arousal and basic facial expressions given a facial photo. We ensure the privacy-awareness of our techniques by using the lightweight architectures of neural networks, such as MT-EmotiDDAMFN, MT-EmotiEffNet, and MT-EmotiMobileFaceNet, that can run even on a mobile device without the need to send facial video to a remote server. It was demonstrated that a significant step in improving the overall accuracy is the smoothing of neural network output scores using Gaussian or box filters. It was experimentally demonstrated that such a simple post-processing of predictions from simple blending of two top visual models improves the F1-score of facial expression recognition up to 7%. At the same time, the mean Concordance Correlation Coefficient (CCC) of valence and arousal is increased by up to 1.25 times compared to each model's frame-level predictions. As a result, our final performance score on the validation set from the multi-task learning challenge is 4.5 times higher than the baseline (1.494 vs 0.32).

2407.13185

Yifan Zhan

Yifan Zhan, Zhuoxiao Li, Muyao Niu, Zhihang Zhong, Shohei Nobuhara, Ko Nishino, Yinqiang Zheng

KFD-NeRF: Rethinking Dynamic NeRF with Kalman Filter

accepted to eccv2024

cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

We introduce KFD-NeRF, a novel dynamic neural radiance field integrated with an efficient and high-quality motion reconstruction framework based on Kalman filtering. Our key idea is to model the dynamic radiance field as a dynamic system whose temporally varying states are estimated based on two sources of knowledge: observations and predictions. We introduce a novel plug-in Kalman filter guided deformation field that enables accurate deformation estimation from scene observations and predictions. We use a shallow Multi-Layer Perceptron (MLP) for observations and model the motion as locally linear to calculate predictions with motion equations. To further enhance the performance of the observation MLP, we introduce regularization in the canonical space to facilitate the network's ability to learn warping for different frames. Additionally, we employ an efficient tri-plane representation for encoding the canonical space, which has been experimentally demonstrated to converge quickly with high quality. This enables us to use a shallower observation MLP, consisting of just two layers in our implementation. We conduct experiments on synthetic and real data and compare with past dynamic NeRF methods. Our KFD-NeRF demonstrates similar or even superior rendering performance within comparable computational time and achieves state-of-the-art view synthesis performance with thorough training.

2407.13186

Takumi Komatsu

Takumi Komatsu and Motonari Kambara and Shumpei Hatanaka and Haruka Matsuo and Tsubasa Hirakawa and Takayoshi Yamashita and Hironobu Fujiyoshi and Komei Sugiura

Nearest Neighbor Future Captioning: Generating Descriptions for Possible Collisions in Object Placement Tasks

Accepted for presentation at Advanced Robotics 24

cs.RO

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Domestic service robots (DSRs) that support people in everyday environments have been widely investigated. However, their ability to predict and describe future risks resulting from their own actions remains insufficient. In this study, we focus on the linguistic explainability of DSRs. Most existing methods do not explicitly model the region of possible collisions; thus, they do not properly generate descriptions of these regions. In this paper, we propose the Nearest Neighbor Future Captioning Model that introduces the Nearest Neighbor Language Model for future captioning of possible collisions, which enhances the model output with a nearest neighbors retrieval mechanism. Furthermore, we introduce the Collision Attention Module that attends regions of possible collisions, which enables our model to generate descriptions that adequately reflect the objects associated with possible collisions. To validate our method, we constructed a new dataset containing samples of collisions that can occur when a DSR places an object in a simulation environment. The experimental results demonstrated that our method outperformed baseline methods, based on the standard metrics. In particular, on CIDEr-D, the baseline method obtained 25.09 points, whereas our method obtained 33.08 points.

2407.13187

Hiroshi Noguchi

Hiroshi Noguchi and Jens Elgeti

Growing tissue sheets on substrates: buds, buckles, and pores

10 pages, 9 figures

physics.bio-ph cond-mat.soft

http://creativecommons.org/licenses/by/4.0/

Many tissues take the form of thin sheets, being only a single cell thick, but millions of cells wide. These tissue sheets can bend and buckle in the third dimension. In this work, we investigated the growth of suspended and supported tissue sheets using particle-based simulations. We combine particle-based tissue growth and meshless membrane models to simulate the growth of tissue sheets with mechanical feedback. Free suspended growing tissues exhibit wrinkling when growth is sufficiently fast. Conversely, tissues on a substrate form buds when the adhesion to the substrate is weak and/or when the friction with the substrate is strong. These buds undergo a membrane-mediated attraction and subsequently fuse. The complete detachment of tissues from the substrate and straight buckled bump formation are also obtained at very weak adhesion and/or fast growth rates. Tissue pores grow via Ostwald ripening and coalescence. The reported dynamics can also be applied in research on the detachment dynamics of different tissues with weakened adhesion.

2407.13188

Guoli Jia

Zhiyuan Ma, Guoli Jia, Biqing Qi, Bowen Zhou

Safe-SD: Safe and Traceable Stable Diffusion with Text Prompt Trigger for Invisible Generative Watermarking

cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Recently, stable diffusion (SD) models have typically flourished in the field of image synthesis and personalized editing, with a range of photorealistic and unprecedented images being successfully generated. As a result, widespread interest has been ignited to develop and use various SD-based tools for visual content creation. However, the exposure of AI-created content on public platforms could raise both legal and ethical risks. In this regard, the traditional methods of adding watermarks to the already generated images (i.e. post-processing) may face a dilemma (e.g., being erased or modified) in terms of copyright protection and content monitoring, since the powerful image inversion and text-to-image editing techniques have been widely explored in SD-based methods. In this work, we propose a Safe and high-traceable Stable Diffusion framework (namely Safe-SD) to adaptively implant the graphical watermarks (e.g., QR code) into the imperceptible structure-related pixels during the generative diffusion process for supporting text-driven invisible watermarking and detection. Different from the previous high-cost injection-then-detection training framework, we design a simple and unified architecture, which makes it possible to simultaneously train watermark injection and detection in a single network, greatly improving the efficiency and convenience of use. Moreover, to further support text-driven generative watermarking and deeply explore its robustness and high-traceability, we elaborately design lambda sampling and encryption algorithm to fine-tune a latent diffuser wrapped by a VAE for balancing high-fidelity image synthesis and high-traceable watermark detection. We present our quantitative and qualitative results on two representative datasets LSUN, COCO and FFHQ, demonstrating state-of-the-art performance of Safe-SD and showing it significantly outperforms the previous approaches.

2407.13189

George Moustakides

George V. Moustakides

Data-Driven Estimation of Conditional Expectations, Application to Optimal Stopping and Reinforcement Learning

20 pages, 6 figures

stat.ML cs.LG

http://creativecommons.org/licenses/by/4.0/

When the underlying conditional density is known, conditional expectations can be computed analytically or numerically. When, however, such knowledge is not available and instead we are given a collection of training data, the goal of this work is to propose simple and purely data-driven means for estimating directly the desired conditional expectation. Because conditional expectations appear in the description of a number of stochastic optimization problems with the corresponding optimal solution satisfying a system of nonlinear equations, we extend our data-driven method to cover such cases as well. We test our methodology by applying it to Optimal Stopping and Optimal Action Policy in Reinforcement Learning.

2407.13190

Sarathkumar N S

V. B. Kiran Kumar and N. S. Sarathkumar

A Note on Generalized Locally Toeplitz Operators

math.NA cs.NA

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Generalized Locally Toeplitz (GLT) matrix sequences arise from large linear systems that approximate Partial Differential Equations (PDEs), Fractional Differential Equations (FDEs), and Integro-Differential Equations (IDEs). GLT sequences of matrices have been developed to study the spectral/singular value behaviour of the numerical approximations to various PDEs, Fades and IDEs. These approximations can be achieved using any discretization method on appropriate grids through local techniques such as Finite Differences, Finite Elements, Finite Volumes, Isogeometric Analysis, and Discontinuous Galerkin methods. Spectral and singular value symbols are essential for analyzing the eigenvalue and singular value distributions of matrix sequences in the Weyl sense. In this article, we provide a comprehensive overview of the operator-theoretic aspect of GLT sequences. The theory of GLT sequences, along with findings on the asymptotic spectral distribution of perturbed matrix sequences, is a highly effective and successful method for calculating the spectral symbol f. Therefore, developing an automatic procedure to compute the spectral symbols of these matrix sequences would be advantageous, a task that Ahmed Ratnani, N S Sarathkumar, S. Serra-Capizzano have partially undertaken. As an application of the theory developed here, we propose an automatic procedure for computing the symbol of the underlying sequences of matrices, assuming they form a GLT sequence that meets mild conditions.

2407.13191

Dileep Mampallil

Surjit Bharatsingh, Piyush Sahu, Gaurav Salwan, Dileep Mampallil

Water drop impact on thin viscous oil layers

under review

physics.flu-dyn cond-mat.soft

http://creativecommons.org/licenses/by/4.0/

We conducted experimental investigations into the short-term and long-term impact dynamics of pure water drops on silicone oil-coated surfaces. Our observations revealed distinct phases: initially, rapid maximal spreading followed by retraction, subsequent oscillations characterized by a relatively stable contact angle, and a final phase involved the gradual spreading of the drop on the underlying solid surface due to the rupturing of the oil layer. The maximal spreading radius follows $We^{1/4}$ scaling, where $We$ is the Weber number, independent of the viscosity of the underlying oil layer. By introducing fluorescent tracer particles in the oil, we noted that the drop expanded over the oil layer during the initial spreading without displacing oil. Subsequently, as the oscillations dampen, the drop ruptures the oil layer, initiating a dewetting process and spreads over the solid surface, leaving a tiny oil droplet under the water drop. Upon completion of this dewetting process, the onset time for the drop spreading nonlinearly increased with the oil layer thickness as $\sim d_o^2$, which we attribute to the dewetting dynamics. Our findings unveil intriguing dynamics of water drops impacting on oil-coated surfaces.

2407.13192

Gyounghun Ko

Gyounghun Ko and Sung-jun Son

Global Stability of the Boltzmann Equation for a Polyatomic Gas with Initial Data Allowing Large Oscillations

29 pages, 2 figures

math.AP

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

In this paper, we consider the Boltzmann equation for a polyatomic gas. We establish that the mild solution to the Boltzmann equation on the torus is globally well-posed, provided the initial data that satisfy bounded velocity-weighted $L^{\infty}$ norm and the smallness condition on the initial relative entropy. Furthermore, we also study the asymptotic behavior of solutions, converging to the global Maxwellian with an exponential rate. A key point in the proof is to develop the pointwise estimate on the gain term of non-linear collision operator for Gr\"{o}nwall's argument.

2407.13193

Shangyu Wu

Shangyu Wu, Ying Xiong, Yufei Cui, Haolun Wu, Can Chen, Ye Yuan, Lianming Huang, Xue Liu, Tei-Wei Kuo, Nan Guan, Chun Jason Xue

Retrieval-Augmented Generation for Natural Language Processing: A Survey

cs.CL

http://creativecommons.org/licenses/by/4.0/

Large language models (LLMs) have demonstrated great success in various fields, benefiting from their huge amount of parameters that store knowledge. However, LLMs still suffer from several key issues, such as hallucination problems, knowledge update issues, and lacking domain-specific expertise. The appearance of retrieval-augmented generation (RAG), which leverages an external knowledge database to augment LLMs, makes up those drawbacks of LLMs. This paper reviews all significant techniques of RAG, especially in the retriever and the retrieval fusions. Besides, tutorial codes are provided for implementing the representative techniques in RAG. This paper further discusses the RAG training, including RAG with/without datastore update. Then, we introduce the application of RAG in representative natural language processing tasks and industrial scenarios. Finally, this paper discusses the future directions and challenges of RAG for promoting its development.

2407.13194

Hui He

Hui He, Qi Zhang, Kun Yi, Xiaojun Xue, Shoujin Wang, Liang Hu and Longbing Cao

Robust Multivariate Time Series Forecasting against Intra- and Inter-Series Transitional Shift

19 pages, 11 figures

cs.LG cs.AI

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

The non-stationary nature of real-world Multivariate Time Series (MTS) data presents forecasting models with a formidable challenge of the time-variant distribution of time series, referred to as distribution shift. Existing studies on the distribution shift mostly adhere to adaptive normalization techniques for alleviating temporal mean and covariance shifts or time-variant modeling for capturing temporal shifts. Despite improving model generalization, these normalization-based methods often assume a time-invariant transition between outputs and inputs but disregard specific intra-/inter-series correlations, while time-variant models overlook the intrinsic causes of the distribution shift. This limits model expressiveness and interpretability of tackling the distribution shift for MTS forecasting. To mitigate such a dilemma, we present a unified Probabilistic Graphical Model to Jointly capturing intra-/inter-series correlations and modeling the time-variant transitional distribution, and instantiate a neural framework called JointPGM for non-stationary MTS forecasting. Specifically, JointPGM first employs multiple Fourier basis functions to learn dynamic time factors and designs two distinct learners: intra-series and inter-series learners. The intra-series learner effectively captures temporal dynamics by utilizing temporal gates, while the inter-series learner explicitly models spatial dynamics through multi-hop propagation, incorporating Gumbel-softmax sampling. These two types of series dynamics are subsequently fused into a latent variable, which is inversely employed to infer time factors, generate final prediction, and perform reconstruction. We validate the effectiveness and efficiency of JointPGM through extensive experiments on six highly non-stationary MTS datasets, achieving state-of-the-art forecasting performance of MTS forecasting.

2407.13195

Yingru Li

Yingru Li, Jiawei Xu, Zhi-Quan Luo

Adaptive Foundation Models for Online Decisions: HyperAgent with Fast Incremental Uncertainty Estimation

41 pages

cs.LG cs.AI cs.HC cs.IT math.IT stat.ML

http://creativecommons.org/licenses/by/4.0/

Foundation models often struggle with uncertainty when faced with new situations in online decision-making, necessitating scalable and efficient exploration to resolve this uncertainty. We introduce GPT-HyperAgent, an augmentation of GPT with HyperAgent for uncertainty-aware, scalable exploration in contextual bandits, a fundamental online decision problem involving natural language input. We prove that HyperAgent achieves fast incremental uncertainty estimation with $\tilde{O}(\log T)$ per-step computational complexity over $T$ periods under the linear realizable assumption. Our analysis demonstrates that HyperAgent's regret order matches that of exact Thompson sampling in linear contextual bandits, closing a significant theoretical gap in scalable exploration. Empirical results in real-world contextual bandit tasks, such as automated content moderation with human feedback, validate the practical effectiveness of GPT-HyperAgent for safety-critical decisions. Our code is open-sourced at \url{https://github.com/szrlee/GPT-HyperAgent/}.

2407.13196

Romuald A. Janik

Dante R. Chialvo and Romuald A. Janik

Statistical thermodynamics of the human brain activity, the Hagedorn temperature and the Zipf law

5+4 pages

q-bio.NC cond-mat.stat-mech

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

It is well established that the brain spontaneously traverses through a very large number of states. Nevertheless, despite its relevance to understanding brain function, a formal description of this phenomenon is still lacking. To this end, we introduce a machine learning based method allowing for the determination of the probabilities of all possible states at a given coarse-graining, from which all the thermodynamics can be derived. This is a challenge not unique to the brain, since similar problems are at the heart of the statistical mechanics of complex systems. This paper uncovers a linear scaling of the entropies and energies of the brain states, a behaviour first conjectured by Hagedorn to be typical at the limiting temperature in which ordinary matter disintegrates into quark matter. Equivalently, this establishes the existence of a Zipf law scaling underlying the appearance of a wide range of brain states. Based on our estimation of the density of states for large scale functional magnetic resonance imaging (fMRI) human brain recordings, we observe that the brain operates asymptotically at the Hagedorn temperature. The presented approach is not only relevant to brain function but should be applicable for a wide variety of complex systems.

2407.13197

Carlo Rovelli

Carlo Rovelli and David Wolpert

Why you do not need to worry about the standard argument that you are a Boltzmann brain

5 pages, no figures

physics.hist-ph physics.data-an

http://creativecommons.org/licenses/by/4.0/

Are you, with your perceptions, memories and observational data, a Boltzmann brain, namely a fleeting statistical fluctuation out of the thermal equilibrium of the universe? Arguments are given in the literature claiming that this bizarre hypothesis needs to be considered seriously, that all of our data about the past is actually a mirage. We point to a difficulty in these arguments. They are based on the dynamical laws and on statistical arguments, but they disregard the fact that we infer the dynamical laws presupposing the reliability of our data records about the past. Hence the reasoning in favor of the Boltzmann brain hypothesis contradicts itself, relying on the reliability of our data about the past to conclude that that data is wrong. More broadly, it is based on incomplete evidence. Incomplete evidence notoriously leads to false conclusions.

2407.13198

Baihan Li

Baihan Li, Zeyu Xie, Xuenan Xu, Yiwei Guo, Ming Yan, Ji Zhang, Kai Yu, Mengyue Wu

DiveSound: LLM-Assisted Automatic Taxonomy Construction for Diverse Audio Generation

cs.SD eess.AS

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Audio generation has attracted significant attention. Despite remarkable enhancement in audio quality, existing models overlook diversity evaluation. This is partially due to the lack of a systematic sound class diversity framework and a matching dataset. To address these issues, we propose DiveSound, a novel framework for constructing multimodal datasets with in-class diversified taxonomy, assisted by large language models. As both textual and visual information can be utilized to guide diverse generation, DiveSound leverages multimodal contrastive representations in data construction. Our framework is highly autonomous and can be easily scaled up. We provide a textaudio-image aligned diversity dataset whose sound event class tags have an average of 2.42 subcategories. Text-to-audio experiments on the constructed dataset show a substantial increase of diversity with the help of the guidance of visual information.

2407.13199

Joshua Liberman

Joshua Liberman, Sebastiaan Y. Haffert, Jared R. Males, Kian Milani

Analyzing Misalignment Tolerances for Implicit Electric Field Conjugation

8 pages, 6 figures, SPIE Astronomical Telescopes + Instrumentation 2024

astro-ph.IM physics.optics

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

High contrast imaging of extrasolar planets and circumstellar disks requires extreme wavefront stability. Such stability can be achieved with active wavefront control (WFC). The next generation of ground- and space-based telescopes will require a robust form of WFC in order to image planets at small inner working angles and extreme flux ratios with respect to the host star. WFC algorithms such as implicit Electric Field Conjugation (iEFC) reduce stellar leakage by minimizing the electric field within a given region of an image, creating a dark hole. iEFC utilizes an empirical approach to sense and remove speckles in the focal plane. While iEFC is empirically calibrated and can handle optical model errors, there are still model assumptions made during the calibration. The performance of iEFC will degrade if the system changes due to slow, optomechanical drifts. In this work, we assess the iEFC performance impacts of pupil misalignments on the deformable mirror and focal plane misalignments on the detector. We base our analysis on the MagAO-X instrument, an extreme AO system installed on the Magellan-Clay telescope, to develop iEFC misalignment tolerancing requirements for both ground- and space-based missions. We present end-to-end physical optics simulations of the MagAO-X instrument, demonstrating iEFC alignment tolerance.

2407.13200

Mengke Li

Mengke Li, Da Li, Guoqing Yang, Yiu-ming Cheung, Hui Huang

Adapt PointFormer: 3D Point Cloud Analysis via Adapting 2D Visual Transformers

ECAI 2024 main conference paper

cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Pre-trained large-scale models have exhibited remarkable efficacy in computer vision, particularly for 2D image analysis. However, when it comes to 3D point clouds, the constrained accessibility of data, in contrast to the vast repositories of images, poses a challenge for the development of 3D pre-trained models. This paper therefore attempts to directly leverage pre-trained models with 2D prior knowledge to accomplish the tasks for 3D point cloud analysis. Accordingly, we propose the Adaptive PointFormer (APF), which fine-tunes pre-trained 2D models with only a modest number of parameters to directly process point clouds, obviating the need for mapping to images. Specifically, we convert raw point clouds into point embeddings for aligning dimensions with image tokens. Given the inherent disorder in point clouds, in contrast to the structured nature of images, we then sequence the point embeddings to optimize the utilization of 2D attention priors. To calibrate attention across 3D and 2D domains and reduce computational overhead, a trainable PointFormer with a limited number of parameters is subsequently concatenated to a frozen pre-trained image model. Extensive experiments on various benchmarks demonstrate the effectiveness of the proposed APF. The source code and more details are available at https://vcc.tech/research/2024/PointFormer.

2407.13201

Kun Wang

Kun Wang, Christopher M. Poskitt, Yang Sun, Jun Sun, Jingyi Wang, Peng Cheng, Jiming Chen

$\mu$Drive: User-Controlled Autonomous Driving

cs.SE

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Autonomous Vehicles (AVs) rely on sophisticated Autonomous Driving Systems (ADSs) to provide passengers a satisfying and safe journey. The individual preferences of riders plays a crucial role in shaping the perception of safety and comfort while they are in the car. Existing ADSs, however, lack mechanisms to systematically capture and integrate rider preferences into their planning modules. To bridge this gap, we propose $\mu$Drive, an event-based Domain-Specific Language (DSL) designed for specifying autonomous vehicle behaviour. $\mu$Drive enables users to express their preferences through rules triggered by contextual events, such as encountering obstacles or navigating complex traffic situations. These rules dynamically adjust the parameter settings of the ADS planning module, facilitating seamless integration of rider preferences into the driving plan. In our evaluation, we demonstrate the feasibility and efficacy of $\mu$Drive by integrating it with the Apollo ADS framework. Our findings show that users can effectively influence Apollo's planning through $\mu$Drive, assisting ADS in achieving improved compliance with traffic regulations. The response time for $\mu$Drive commands remains consistently at the second or millisecond level. This suggests that $\mu$Drive may help pave the way to more personalizsed and user-centric AV experiences.

2407.13202

Zubiao Xiao

Zubiao Xiao, Hongwei Jia and Zhengyu Yin

Scaled packing pressures on subsets for amenable group actions

30pages

math.DS

http://creativecommons.org/licenses/by/4.0/

In this paper, we study the properties of the scaled packing topological pressures for topological dynamical system $(X,G)$, where $G$ is a countable discrete infinite amenable group. We show that the scaled packing topological pressures can be determined by the scaled Bowen topological pressures. We obtain Billingsley's Theorem for the scaled packing pressures with a $G$-action. Then we get a variational principle between the scaled packing pressures and the scaled measure-theoretic upper local pressures. Finally, we give some restrictions on the scaled sequence $\mathbf{b}$, then in the case of the set $X_{\mu}$ of generic points, we prove that $$P^{P}(X_{\mu},\left\{F_{n}\right\},f,\mathbf{b})=h_{\mu}(X)+\int_{X} f \mathrm{d}\mu,$$ if $\left\{F_{n}\right\}$ is tempered and $\mu$ is a $G$-invariant ergodic Borel probability measure.

2407.13203

Qing Cui

Qing Cui

Minimal hypersurfaces in $\mathbb{S}^5$ with constant scalar curvature and zero Gauss curvature are totally geodesic

11 pages, no figure

math.DG

http://creativecommons.org/publicdomain/zero/1.0/

We show that a closed minimal hypersurface in $\mathbb{S}^5$ with constant scalar curvature and zero Gauss curvature is totally geodesic.

2407.13204

Manudeep Bhuller

Richard Audoly and Manudeep Bhuller and Tore Adam Reiremo

The Pay and Non-Pay Content of Job Ads

econ.GN q-fin.EC

http://creativecommons.org/licenses/by/4.0/

How informative are job ads about the actual pay and amenities offered by employers? Using a comprehensive database of job ads posted by Norwegian employers, we develop a methodology to systematically classify the information on both pay and non-pay job attributes advertised in vacancy texts. We link this information to measures of employer attractiveness, which we derive from a job search model estimated on observed wages and worker mobility flows. About 55 percent of job ads provide information related to pay and nearly all ads feature information on non-pay attributes. We show that publicly advertised job attributes are meaningful predictors of employer attractiveness, and non-pay attributes are about as predictive as pay-related attributes. High-pay employers mention pay-related attributes more often, while high-amenity employers are more likely to advertise flexible working hours and contract duration.

2407.13205

Guohang He

Wei Lan, Guohang He, Mingyang Liu, Qingfeng Chen, Junyue Cao, Wei Peng

Transformer-based Single-Cell Language Model: A Survey

cs.CL

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

The transformers have achieved significant accomplishments in the natural language processing as its outstanding parallel processing capabilities and highly flexible attention mechanism. In addition, increasing studies based on transformers have been proposed to model single-cell data. In this review, we attempt to systematically summarize the single-cell language models and applications based on transformers. First, we provide a detailed introduction about the structure and principles of transformers. Then, we review the single-cell language models and large language models for single-cell data analysis. Moreover, we explore the datasets and applications of single-cell language models in downstream tasks such as batch correction, cell clustering, cell type annotation, gene regulatory network inference and perturbation response. Further, we discuss the challenges of single-cell language models and provide promising research directions. We hope this review will serve as an up-to-date reference for researchers interested in the direction of single-cell language models.

2407.13206

Vygintas Gontis

Vygintas Gontis

Double stochastic opinion dynamics with fractional inflow of new opinions

9 pages, 2 figures, 1 table

physics.soc-ph

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

A recent analysis of empirical limit order flow data highlights the necessity for a more refined order flow model that integrates the power-law distribution of limit order cancellation times. These cancellation times follow a discrete probability mass function derived from the Tsallis $q$-exponential distribution, or equivalently, the second form of the Pareto distribution. By combining fractional L'{e}vy stable motion as the model for limit order inflow with the power-law distribution for cancellation times, we propose an innovative approach to modeling order imbalance in financial markets. We extend this model to a broader context, illustrating its applicability to opinion dynamics in social systems where opinions have a finite lifespan. This proposed model exemplifies a stochastic time series characterized by stationary increments and broken self-similarity. Consequently, it offers a novel framework for testing methods to evaluate long-range dependence in such time series.

2407.13207

Rishabh Singh Teja

Rishabh Singh Teja, Jared A. Goldberg, D. K. Sahu, G. C. Anupama, Avinash Singh, Vishwajeet Swain, and Varun Bhalerao

SN 2021wvw: A core-collapse supernova at the sub-luminous, slower, and shorter end of Type IIPs

Accepted for publication in the Astrophysical Journal (18 pages, 13 figures, 4 tables)

astro-ph.HE

http://creativecommons.org/licenses/by/4.0/

We present detailed multi-band photometric and spectroscopic observations and analysis of a rare core-collapse supernova SN 2021wvw, that includes photometric evolution up to 250 d and spectroscopic coverage up to 100 d post-explosion. A unique event that does not fit well within the general trends observed for Type II-P supernovae, SN 2021wvw shows an intermediate luminosity with a short plateau phase of just about 75 d, followed by a very sharp (~10 d) transition to the tail phase. Even in the velocity space, it lies at a lower velocity compared to a larger Type II sample. The observed peak absolute magnitude is -16.1 mag in r-band, and the nickel mass is well constrained to 0.020(6) Msol. Detailed hydrodynamical modeling using MESA+STELLA suggests a radially compact, low-metallicity, high-mass Red Supergiant progenitor (ZAMS mass=18 Msol), which exploded with ~0.2e51 erg/s leaving an ejecta mass of Mej~5 Msol. Significant late-time fallback during the shock propagation phase is also seen in progenitor+explosion models consistent with the light curve properties. As the faintest short-plateau supernova characterized to date, this event adds to the growing diversity of transitional events between the canonical ~100 d plateau Type IIP and stripped-envelope events.

2407.13208

Erika Roldan

Inga Johnson and Erika Roldan

Solution Numbers for Eight Blocks to Madness Puzzle

22 pages, 12 figures, 3 Tables

math.CO cs.DM math.HO

http://creativecommons.org/licenses/by-nc-nd/4.0/

The 30 MacMahon colored cubes have each face painted with one of six colors and every color appears on at least one face. One puzzle involving these cubes is to create a $2\times2\times2$ model with eight distinct MacMahon cubes to recreate a larger version with the external coloring of a specified target cube, also a MacMahon cube, and touching interior faces are the same color. J.H. Conway is credited with arranging the cubes in a $6\times6$ tableau that gives a solution to this puzzle. In fact, the particular set of eight cubes that solves this puzzle can be arranged in exactly \textit{two} distinct ways to solve the puzzle. We study a less restrictive puzzle without requiring interior face matching. We describe solutions to the $2\times2\times2$ puzzle and the number of distinct solutions attainable for a collection of eight cubes. Additionally, given a collection of eight MacMahon cubes, we study the number of target cubes that can be built in a $2\times2\times2$ model. We calculate the distribution of the number of cubes that can be built over all collections of eight cubes (the maximum number is five) and provide a complete characterization of the collections that can build five distinct cubes. Furthermore, we identify nine new sets of twelve cubes, called Minimum Universal sets, from which all 30 cubes can be built.

2407.13209

Zhong-Ying Fan

Hong-Ming Cui and Zhong-Ying Fan

Criticality of global monopole charges in diverse dimensions

21pages, 9figures;

hep-th gr-qc

http://creativecommons.org/licenses/by/4.0/

In this work, we construct charged AdS black holes with a global monopole charge in diverse dimensions and study the thermodynamics. We find a critical monopole charge below which the solution exhibits Van-der Waals like behaviors. In the context of holography, this could be intepreted using the boundary degrees of freedoms. As an example, we study the phase diagram in the four dimensions analytically. We further analyze the microstructures of the solution using Ruppeiner geometry. We find that repulsive interactions dominates for black holes in a wide range of temperatures. However, around the critical point, attractive interactions is dominant and the Ruppeiner scalar curvature shows universal behaviors: it has a critical exponent $2$ and coefficient $-1/8$ in diverse dimensions. Universality of the results is interpreted from the scaling behavior of free energy near the critical point for Van-der Waals like fluids.

2407.13210

Xiaoming Zhang

Chunli Li, Xiaoming Zhang, Yuan Gao, Xiaoli Yin, Le Lu, Ling Zhang, Ke Yan, Yu Shi

Improved Esophageal Varices Assessment from Non-Contrast CT Scans

Early accepted to MICCAI 2024

eess.IV cs.CV

http://creativecommons.org/licenses/by-nc-nd/4.0/

Esophageal varices (EV), a serious health concern resulting from portal hypertension, are traditionally diagnosed through invasive endoscopic procedures. Despite non-contrast computed tomography (NC-CT) imaging being a less expensive and non-invasive imaging modality, it has yet to gain full acceptance as a primary clinical diagnostic tool for EV evaluation. To overcome existing diagnostic challenges, we present the Multi-Organ-cOhesion-Network (MOON), a novel framework enhancing the analysis of critical organ features in NC-CT scans for effective assessment of EV. Drawing inspiration from the thorough assessment practices of radiologists, MOON establishes a cohesive multiorgan analysis model that unifies the imaging features of the related organs of EV, namely esophagus, liver, and spleen. This integration significantly increases the diagnostic accuracy for EV. We have compiled an extensive NC-CT dataset of 1,255 patients diagnosed with EV, spanning three grades of severity. Each case is corroborated by endoscopic diagnostic results. The efficacy of MOON has been substantiated through a validation process involving multi-fold cross-validation on 1,010 cases and an independent test on 245 cases, exhibiting superior diagnostic performance compared to methods focusing solely on the esophagus (for classifying severe grade: AUC of 0.864 versus 0.803, and for moderate to severe grades: AUC of 0.832 versus 0.793). To our knowledge, MOON is the first work to incorporate a synchronized multi-organ NC-CT analysis for EV assessment, providing a more acceptable and minimally invasive alternative for patients compared to traditional endoscopy.

2407.13211

Hao Yan

Hao Yan, Zixiang Wang, Zhengjia Xu, Zhuoyue Wang, Zhizhong Wu, Ranran Lyu

Research on Image Super-Resolution Reconstruction Mechanism based on Convolutional Neural Network

cs.CV eess.IV

http://creativecommons.org/licenses/by/4.0/

Super-resolution reconstruction techniques entail the utilization of software algorithms to transform one or more sets of low-resolution images captured from the same scene into high-resolution images. In recent years, considerable advancement has been observed in the domain of single-image super-resolution algorithms, particularly those based on deep learning techniques. Nevertheless, the extraction of image features and nonlinear mapping methods in the reconstruction process remain challenging for existing algorithms. These issues result in the network architecture being unable to effectively utilize the diverse range of information at different levels. The loss of high-frequency details is significant, and the final reconstructed image features are overly smooth, with a lack of fine texture details. This negatively impacts the subjective visual quality of the image. The objective is to recover high-quality, high-resolution images from low-resolution images. In this work, an enhanced deep convolutional neural network model is employed, comprising multiple convolutional layers, each of which is configured with specific filters and activation functions to effectively capture the diverse features of the image. Furthermore, a residual learning strategy is employed to accelerate training and enhance the convergence of the network, while sub-pixel convolutional layers are utilized to refine the high-frequency details and textures of the image. The experimental analysis demonstrates the superior performance of the proposed model on multiple public datasets when compared with the traditional bicubic interpolation method and several other learning-based super-resolution methods. Furthermore, it proves the model's efficacy in maintaining image edges and textures.

2407.13212

Liang-Liang Yu

Liang-Liang Yu

Probe the regolith characteristics of asteroids from 9-years infrared observations of WISE/NEOWISE: A case study of the Main-Belt Object (656) Beagle

10 pages, 9 figures, accepted for publication in The Astronomical Journal. arXiv admin note: text overlap with arXiv:2104.02909

astro-ph.EP

http://creativecommons.org/licenses/by/4.0/

This work presents data processing, fitting procedure, modelling and analyzing of 9-years infrared light curves provided by the WISE/NEOWISE telescope, by which the regolith characteristics of Main-Belt Object (656) Beagle is studied. We determine Beagle's effective diameter $D_{\rm eff}=57.3^{+4.5}_{-2.2}$ km, geometric albedo $p_{\rm v}=0.05^{+0.004}_{-0.007}$, mean roughness $\theta_{\rm RMS}=44\pm4^\circ$, mean grain size $b=100^{+350}_{-90}~\mu$m, mean specific heat capacity $c_{\rm p}=173\sim516\rm~JKg^{-1}K^{-1}$, mean thermal conductivity $\kappa=0.7\sim1.3\times10^{-3}\rm~Wm^{-1}K^{-1}$ and mean thermal inertia $\Gamma=14\sim32\rm~Jm^{-2}s^{-0.5}K^{-1}$. The albedo of Beagle is a little anomalous that the albedos of Beagle's neighbouring asteroids are more close to Themis, rather than Beagle itself. The W1-band near-infrared light curves don't reveal significant heterogeneous NIR features on the surface of Beagle, being inconsistent with the expectation of a family parent that has members with diverse NIR spectral types. These results add new clues of Beagle probably being an interloper or a sister, rather than the parent of its neighbouring asteroids including the first main-belt comet (MBC) 133P, hence may lead to new scenarios about the origin of famous MBC 133P. Besides, we found that asteroidal shape models from inversion of optical light curves are imperfect for modeling infrared lightcurves, thus could mislead evaluations of both the heterogeneity of regolith reflectivity at near infrared and thermophysical characteristics at thermal infrared.

2407.13213

Antonino Zanette

Ludovic Goudenege, Andrea Molent, Antonino Zanette

Leveraging Machine Learning for High-Dimensional Option Pricing within the Uncertain Volatility Model

q-fin.CP

http://creativecommons.org/licenses/by/4.0/

This paper explores the application of Machine Learning techniques for pricing high-dimensional options within the framework of the Uncertain Volatility Model (UVM). The UVM is a robust framework that accounts for the inherent unpredictability of market volatility by setting upper and lower bounds on volatility and the correlation among underlying assets. By leveraging historical data and extreme values of estimated volatilities and correlations, the model establishes a confidence interval for future volatility and correlations, thus providing a more realistic approach to option pricing. By integrating advanced Machine Learning algorithms, we aim to enhance the accuracy and efficiency of option pricing under the UVM, especially when the option price depends on a large number of variables, such as in basket or path-dependent options. Our approach evolves backward in time, dynamically selecting at each time step the most expensive volatility and correlation for each market state. Specifically, it identifies the particular values of volatility and correlation that maximize the expected option value at the next time step. This is achieved through the use of Gaussian Process regression, the computation of expectations via a single step of a multidimensional tree and the Sequential Quadratic Programming optimization algorithm. The numerical results demonstrate that the proposed approach can significantly improve the precision of option pricing and risk management strategies compared with methods already in the literature, particularly in high-dimensional contexts.

2407.13214

Lu Gan

Lu Gan and Xi Li

TXL-PBC: a freely accessible labeled peripheral blood cell dataset

cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

In a recent study, we found that publicly BCCD and BCD datasets have significant issues such as labeling errors, insufficient sample size, and poor data quality. To address these problems, we performed sample deletion, re-labeling, and integration of these two datasets. Additionally, we introduced the PBC and Raabin-WBC datasets, and ultimately created a high-quality, sample-balanced new dataset, which we named TXL-PBC. The dataset contains 1008 training sets, 288 validation sets, and 144 test sets. Firstly, The dataset underwent strict manual annotation, automatic annotation with YOLOv8n model, and manual audit steps to ensure the accuracy and consistency of annotations. Secondly, we addresses the blood cell mislabeling problem of the original datasets. The distribution of label boundary box areas and the number of labels are better than the BCCD and BCD datasets. Moreover, we used the YOLOv8n model to train these three datasets, the performance of the TXL-PBC dataset surpass the original two datasets. Finally, we employed YOLOv5n, YOLOv5s, YOLOv5l, YOLOv8s, YOLOv8m detection models as the baseline models for TXL-PBC. This study not only enhances the quality of the blood cell dataset but also supports researchers in improving models for blood cell target detection. We published our freely accessible TXL-PBC dataset at https://github.com/lugan113/TXL-PBC\_Dataset.

2407.13215

Martin Hairer

Luca Gerolla, Martin Hairer, Xue-Mei Li

Scaling limit of the KPZ equation with non-integrable spatial correlations

math.PR math-ph math.MP

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

We study the large scale fluctuations of the KPZ equation in dimensions $d \geq 3$ driven by Gaussian noise that is white in time Gaussian but features non-integrable spatial correlation with decay rate $\kappa \in (2, d)$ and a suitable limiting profile. We show that its scaling limit is described by the corresponding additive stochastic heat equation. In contrast to the case of compactly supported covariance, the noise in the stochastic heat equation retains spatial correlation with covariance $|x|^{-\kappa}$. Surprisingly, the noise driving the limiting equation turns out to be the scaling limit of the noise driving the KPZ equation so that, under a suitable coupling, one has convergence in probability, unlike in the case of integrable correlations where fluctuations are enhanced in the limit and convergence is necessarily weak.

2407.13216

Cao Doanh Bui

Trinh T. L. Vuong, Doanh C. Bui and Jin Tae Kwak

QuIIL at T3 challenge: Towards Automation in Life-Saving Intervention Procedures from First-Person View

MICCAI-Thompson Challenge 2023

cs.CV

http://creativecommons.org/licenses/by/4.0/

In this paper, we present our solutions for a spectrum of automation tasks in life-saving intervention procedures within the Trauma THOMPSON (T3) Challenge, encompassing action recognition, action anticipation, and Visual Question Answering (VQA). For action recognition and anticipation, we propose a pre-processing strategy that samples and stitches multiple inputs into a single image and then incorporates momentum- and attention-based knowledge distillation to improve the performance of the two tasks. For training, we present an action dictionary-guided design, which consistently yields the most favorable results across our experiments. In the realm of VQA, we leverage object-level features and deploy co-attention networks to train both object and question features. Notably, we introduce a novel frame-question cross-attention mechanism at the network's core for enhanced performance. Our solutions achieve the $2^{nd}$ rank in action recognition and anticipation tasks and $1^{st}$ rank in the VQA task.

2407.13217

Xiaoming Zhang

Wei Huang, Wei Liu, Xiaoming Zhang, Xiaoli Yin, Xu Han, Chunli Li, Yuan Gao, Yu Shi, Le Lu, Ling Zhang, Lei Zhang, Ke Yan

LIDIA: Precise Liver Tumor Diagnosis on Multi-Phase Contrast-Enhanced CT via Iterative Fusion and Asymmetric Contrastive Learning

Accepted to MICCAI 2024

cs.CV

http://creativecommons.org/licenses/by-nc-nd/4.0/

The early detection and precise diagnosis of liver tumors are tasks of critical clinical value, yet they pose significant challenges due to the high heterogeneity and variability of liver tumors. In this work, a precise LIver tumor DIAgnosis network on multi-phase contrast-enhance CT, named LIDIA, is proposed for real-world scenario. To fully utilize all available phases in contrast-enhanced CT, LIDIA first employs the iterative fusion module to aggregate variable numbers of image phases, thereby capturing the features of lesions at different phases for better tumor diagnosis. To effectively mitigate the high heterogeneity problem of liver tumors, LIDIA incorporates asymmetric contrastive learning to enhance the discriminability between different classes. To evaluate our method, we constructed a large-scale dataset comprising 1,921 patients and 8,138 lesions. LIDIA has achieved an average AUC of 93.6% across eight different types of lesions, demonstrating its effectiveness. Besides, LIDIA also demonstrated strong generalizability with an average AUC of 89.3% when tested on an external cohort of 828 patients.

2407.13218

Fedor Borisyuk

Fedor Borisyuk, Qingquan Song, Mingzhou Zhou, Ganesh Parameswaran, Madhu Arun, Siva Popuri, Tugrul Bingol, Zhuotao Pei, Kuang-Hsuan Lee, Lu Zheng, Qizhan Shao, Ali Naqvi, Sen Zhou, Aman Gupta

LiNR: Model Based Neural Retrieval on GPUs at LinkedIn

cs.LG cs.AI

http://creativecommons.org/licenses/by/4.0/

This paper introduces LiNR, LinkedIn's large-scale, GPU-based retrieval system. LiNR supports a billion-sized index on GPU models. We discuss our experiences and challenges in creating scalable, differentiable search indexes using TensorFlow and PyTorch at production scale. In LiNR, both items and model weights are integrated into the model binary. Viewing index construction as a form of model training, we describe scaling our system for large indexes, incorporating full scans and efficient filtering. A key focus is on enabling attribute-based pre-filtering for exhaustive GPU searches, addressing the common challenge of post-filtering in KNN searches that often reduces system quality. We further provide multi-embedding retrieval algorithms and strategies for tackling cold start issues in retrieval. Our advancements in supporting larger indexes through quantization are also discussed. We believe LiNR represents one of the industry's first Live-updated model-based retrieval indexes. Applied to out-of-network post recommendations on LinkedIn Feed, LiNR has contributed to a 3% relative increase in professional daily active users. We envisage LiNR as a step towards integrating retrieval and ranking into a single GPU model, simplifying complex infrastructures and enabling end-to-end optimization of the entire differentiable infrastructure through gradient descent.

2407.13219

Wei Feng

Wei Feng, Xin Wang, Hong Chen, Zeyang Zhang, Wenwu Zhu

Multi-sentence Video Grounding for Long Video Generation

cs.CV

http://creativecommons.org/licenses/by-nc-nd/4.0/

Video generation has witnessed great success recently, but their application in generating long videos still remains challenging due to the difficulty in maintaining the temporal consistency of generated videos and the high memory cost during generation. To tackle the problems, in this paper, we propose a brave and new idea of Multi-sentence Video Grounding for Long Video Generation, connecting the massive video moment retrieval to the video generation task for the first time, providing a new paradigm for long video generation. The method of our work can be summarized as three steps: (i) We design sequential scene text prompts as the queries for video grounding, utilizing the massive video moment retrieval to search for video moment segments that meet the text requirements in the video database. (ii) Based on the source frames of retrieved video moment segments, we adopt video editing methods to create new video content while preserving the temporal consistency of the retrieved video. Since the editing can be conducted segment by segment, and even frame by frame, it largely reduces the memory cost. (iii) We also attempt video morphing and personalized generation methods to improve the subject consistency of long video generation, providing ablation experimental results for the subtasks of long video generation. Our approach seamlessly extends the development in image/video editing, video morphing and personalized generation, and video grounding to the long video generation, offering effective solutions for generating long videos at low memory cost.

2407.13220

Huadai Liu

Huadai Liu, Jialei Wang, Rongjie Huang, Yang Liu, Jiayang Xu, Zhou Zhao

MEDIC: Zero-shot Music Editing with Disentangled Inversion Control

eess.AS cs.SD

http://creativecommons.org/publicdomain/zero/1.0/

Text-guided diffusion models catalyze a paradigm shift in audio generation, facilitating the adaptability of source audio to conform to specific textual prompts. Recent advancements introduce inversion techniques, like DDIM inversion, to zero-shot editing, exploiting pre-trained diffusion models for audio modification. Nonetheless, our investigation exposes that DDIM inversion suffers from an accumulation of errors across each diffusion step, undermining its efficacy. And the lack of attention control hinders the fine-grained manipulations of music. To counteract these limitations, we introduce the \textit{Disentangled Inversion} technique, which is designed to disentangle the diffusion process into triple branches, thereby magnifying their individual capabilities for both precise editing and preservation. Furthermore, we propose the \textit{Harmonized Attention Control} framework, which unifies the mutual self-attention and cross-attention with an additional Harmonic Branch to achieve the desired composition and structural information in the target music. Collectively, these innovations comprise the \textit{Disentangled Inversion Control (DIC)} framework, enabling accurate music editing whilst safeguarding structural integrity. To benchmark audio editing efficacy, we introduce \textit{ZoME-Bench}, a comprehensive music editing benchmark hosting 1,100 samples spread across 10 distinct editing categories, which facilitates both zero-shot and instruction-based music editing tasks. Our method demonstrates unparalleled performance in edit fidelity and essential content preservation, outperforming contemporary state-of-the-art inversion techniques.

2407.13221

Taian Guo

Taian Guo, Taolin Zhang, Haoqian Wu, Hanjun Li, Ruizhi Qiao, Xing Sun

Multimodal Label Relevance Ranking via Reinforcement Learning

Accepted to ECCV2024

cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Conventional multi-label recognition methods often focus on label confidence, frequently overlooking the pivotal role of partial order relations consistent with human preference. To resolve these issues, we introduce a novel method for multimodal label relevance ranking, named Label Relevance Ranking with Proximal Policy Optimization (LR\textsuperscript{2}PPO), which effectively discerns partial order relations among labels. LR\textsuperscript{2}PPO first utilizes partial order pairs in the target domain to train a reward model, which aims to capture human preference intrinsic to the specific scenario. Furthermore, we meticulously design state representation and a policy loss tailored for ranking tasks, enabling LR\textsuperscript{2}PPO to boost the performance of label relevance ranking model and largely reduce the requirement of partial order annotation for transferring to new scenes. To assist in the evaluation of our approach and similar methods, we further propose a novel benchmark dataset, LRMovieNet, featuring multimodal labels and their corresponding partial order data. Extensive experiments demonstrate that our LR\textsuperscript{2}PPO algorithm achieves state-of-the-art performance, proving its effectiveness in addressing the multimodal label relevance ranking problem. Codes and the proposed LRMovieNet dataset are publicly available at \url{https://github.com/ChazzyGordon/LR2PPO}.

2407.13222

Mohammad Wassaf Ali

Mohammad Wassaf Ali, Ayushi Gupta, Mujeev Khan and Mohd Wajid

Non-Contact Breath Rate Classification Using SVM Model and mmWave Radar Sensor Data

6 Pages, 5 figures

cs.LG

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

This work presents the use of frequency modulated continuous wave (FMCW) radar technology combined with a machine learning model to differentiate between normal and abnormal breath rates. The proposed system non-contactly collects data using FMCW radar, which depends on breath rates. Various support vector machine kernels are used to classify the observed data into normal and abnormal states. Prolonged experiments show good accuracy in breath rate classification, confirming the model's efficacy. The best accuracy is 95 percent with the smallest number of support vectors in the case of the quadratic polynomial kernel.

2407.13223

Mohammad R. Garousi

Mohammad R. Garousi

$O(9,25)$ symmetry of heterotic string theory at orders $\alpha'$, $\alpha'^2$

19 pages, Latex file, no figure

hep-th

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

In a recent study, we have observed that by imposing a truncated T-duality transformation on the circular reduction of the bosonic couplings in the heterotic theory at four- and six-derivative orders, we can calculate these couplings in a particular YM gauge where the YM potential vanishes but its field strength remains non-zero. Importantly, the coupling constants are independent of the gauge choice, so these results are valid across different YM gauge choices. In this work, we explore the cosmological reduction of these couplings when the YM gauge fields belong to the Cartan subalgebra of $SO(32)$ or $E_8 \times E_8$. We demonstrate that after applying appropriate one-dimensional field redefinitions and total derivative terms, the couplings can be expressed in a proposed $O(9,25)$-invariant canonical form, which is the extension of the canonical $O(9,9)$-invariant form for just the NS-NS fields proposed by Hohm and Zwiebach. This $O(9,25)$-invariant expression is in terms of the trace of the first time derivative of the generalized metric, which encompasses both the YM field and the NS-NS fields.

2407.13224

Wael Othman

Wadha Alfalasi, Wael Othman, Tanveer Hussain, Nacir Tit

Vacancy-Induced Boron Nitride Monolayers as Multifunctional Materials for Metal Ion Batteries and Hydrogen Storage Applications

cond-mat.mtrl-sci physics.comp-ph

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

This study comprehensively examined the structural, electronic, electrochemical, and energy storage properties of boron-vacancy induced porous boron nitride monolayers (BN:VB) as multifunctional materials, anodes for MIBs and H2 storage applications. Our computational approaches, density functional theory (DFT), ab initio molecular dynamics (AIMD), and thermodynamic analysis, revealed exceptionally high energy and gravimetric densities for MIBs and H2 storage, respectively. We investigated the interactions of Li, Na, and K atoms on BN:VB, which strongly bonded with binding energies stronger than their bulk cohesive energies, which ensured structural stability and the absence of metal clustering. Electronic properties, analyzed through spin-polarized partial density of states (PDOS), band structure, and Bader charge analysis, revealed significant charge transfers from the metal atoms to BN:VB, enhancing the electronic conductivity of the latter. Theoretical specific capacities were calculated as 1821.53, 786.11, and 490.51 mA h/g for Li, Na, and K, respectively, which comfortably exceeded the conventional anodes, such as graphite. Average open-circuit voltages (OCVs) were found as 0.15, 0.25, and 0.32 V, for Li, Na, and K, respectively, indicating strong electrochemical stability. Diffusion studies showed lower barriers of 0.47, 0.08, and 0.60 eV for Li, Na, and K, respectively, with increased metal loadings, suggesting enhanced mobilities and charge/discharge rates. On the other side, the metal-functionalized BN:VB monolayers exhibited remarkably high H2 gravimetric capacities, supported by Langmuir adsorption model-based statistical thermodynamic analysis. Average adsorption energies of H2 on 4Li-, 4Na-, and 4K@BN:VB, were found in perfect range for practical storage applications.

2407.13225

M. A. Gonzalez Leon

A. Alonso-Izquierdo, D. Canillas Martinez, C. Garzon Sanchez, M.A. Gonzalez Leon, A. Wereszczynski

Stability and decay of composite kinks/$Q$-balls solutions in a deformed $O(2N+1)$ linear sigma model

24 pages, 42 figures

hep-th math-ph math.MP

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

The defect-type solutions of a deformed $O(2N+1)$ linear sigma model with a real and $N$ complex fields in $(1+1)$-dimensional Minkowski spacetime are studied. All the solutions are analytically found for the $N=2$ case. Two types of solitons have been determined: (a) Simple solutions formed by a topological kink with or without the presence of a $Q$-ball. (b) Composite solutions. They are constituted by some one-parameter families of solutions which can be understood as a non-linear combination of simple solutions. The properties of all of those solutions and the analysis of their linear stability, as well as decay channels, are discussed.

2407.13226

Martin Piecka

M. Piecka, S. Hutschenreuter, J. Alves

Towards a complete picture of the Sco-Cen outflow

12 pages, 9 figures, 3 tables, accepted for publication in Astronomy & Astrophysics

astro-ph.GA astro-ph.SR

http://creativecommons.org/licenses/by/4.0/

Previous studies have shown strong evidence that the Sun is crossing an outflow originating from the Sco-Cen OB association. Understanding this outflow's origin and structure illuminates how massive star formation shapes the interstellar medium (ISM) and helps predict future Galactic conditions affecting our Solar System. We analysed H I emission and optical ISM absorption lines towards 47 early-type stars around the Upper Sco region to refine the map of the Sco-Cen outflow. Combined with data for nearby stars, we find that the outflow has at least two components: a faster, low-density component traced by Ca II, and a slower, possibly lower-density component traced by Mg II and Fe II in the UV that is passing through the Earth. A constant flow model successfully describes both components with $(l,b,|\vec{v}|) = (335.4^{\circ}, -6.8^{\circ}, 14.0 \,\,\textrm{km}\,\textrm{s}^{-1})$ and $(305.5^{\circ}, +17.6^{\circ}, 21.2\,\,\textrm{km}\,\textrm{s}^{-1})$, respectively. The origin of the faster component points towards the Sco-Cen 15 Myr population, which is consistent with the origin of the slower component within 2 $\sigma$. A simple model comparison indicates that a constant flow is favoured over a spherical flow geometry, implying an extended distribution of feedback sources within Sco-Cen. We also found that a poorly studied 25 pc long H I cloud at a distance of 107 pc belongs to the established Sco-Cen flow.

2407.13227

Jie Gao

Fei Yan, Jie Gao, Tao Feng, Jianxing Liu

Solving the Model Unavailable MARE using Q-Learning Algorithm

eess.SY cs.SY

http://creativecommons.org/licenses/by/4.0/

In this paper, the discrete-time modified algebraic Riccati equation (MARE) is solved when the system model is completely unavailable. To achieve this, firstly a brand new iterative method based on the standard discrete-time algebraic Riccati equation (DARE) and its input weighting matrix is proposed to solve the MARE. For the single-input case, the iteration can be initialized by an arbitrary positive input weighting if and only if the MARE has a stabilizing solution; nevertheless a pre-given input weighting matrix of a sufficiently large magnitude is used to perform the iteration for the multi-input case when the characteristic parameter belongs to a specified subset. Benefit from the developed specific iteration structure, the Q-learning (QL) algorithm can be employed to subtly solve the MARE where only the system input/output data is used thus the system model is not required. Finally, a numerical simulation example is given to verify the effectiveness of the theoretical results and the algorithm.

2407.13228

Peter Washington

Junwei Sun, Siqi Ma, Yiran Fan, Peter Washington

Evaluating Large Language Models for Anxiety and Depression Classification using Counseling and Psychotherapy Transcripts

cs.CL cs.CY cs.ET cs.LG

http://creativecommons.org/licenses/by/4.0/

We aim to evaluate the efficacy of traditional machine learning and large language models (LLMs) in classifying anxiety and depression from long conversational transcripts. We fine-tune both established transformer models (BERT, RoBERTa, Longformer) and more recent large models (Mistral-7B), trained a Support Vector Machine with feature engineering, and assessed GPT models through prompting. We observe that state-of-the-art models fail to enhance classification outcomes compared to traditional machine learning methods.

2407.13229

Jindou Jia

Jindou Jia, Yuhang Liu, Kexin Guo, Xiang Yu, Lihua Xie, and Lei Guo

Disturbance Observer for Estimating Coupled Disturbances

8 pages, 3 figures

cs.RO cs.SY eess.SY

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

High-precision control for nonlinear systems is impeded by the low-fidelity dynamical model and external disturbance. Especially, the intricate coupling between internal uncertainty and external disturbance is usually difficult to be modeled explicitly. Here we show an effective and convergent algorithm enabling accurate estimation of the coupled disturbance via combining control and learning philosophies. Specifically, by resorting to Chebyshev series expansion, the coupled disturbance is firstly decomposed into an unknown parameter matrix and two known structures depending on system state and external disturbance respectively. A Regularized Least Squares (RLS) algorithm is subsequently formalized to learn the parameter matrix by using historical time-series data. Finally, a higher-order disturbance observer (HODO) is developed to achieve a high-precision estimation of the coupled disturbance by utilizing the learned portion. The efficiency of the proposed algorithm is evaluated through extensive simulations. We believe this work can offer a new option to merge learning schemes into the control framework for addressing existing intractable control problems.

2407.13230

Ziqi Liu

Ziqi Liu, Chengfeng Zhu, Yuchen Gao, Zuxin Chen, Pingfan Gu, and Yu Ye

Probing spin textures in atomically thin CrSBr through tunneling magnetoresistance

8 pages, 4 figures

cond-mat.mes-hall

http://creativecommons.org/publicdomain/zero/1.0/

The exploration of spin configurations and magnetoresistance in van der Waals magnetic semiconductors, particularly in the realm of thin-layer structures, holds paramount significance for the development of two-dimensional spintronic nanodevices. In this Letter, we conducted comprehensive magnetotransport measurements on a few-layer CrSBr using a vertical tunneling device configuration. Notably, our investigation revealed that tunneling magnetoresistance possesses a distinctive capability to discern spin configurations that would otherwise remain indistinguishable through alternative techniques such as photoluminescence. We observed the existence of energy-degenerate states exhibiting identical net magnetization and comparable spin configurations, which could be differentiated based on their rectification properties, reminiscent of a diode-like behavior at positive and negative bias voltages. Specifically, in devices comprising 5-layer CrSBr, we observed an intriguing positive magnetoresistive state when subjected to an in-plane magnetic field along the $b$-axis. To gain a deeper understanding of the underlying mechanisms, we developed a one-dimensional linear chain model that successfully computed the magnetic state, thereby elucidating the underlying spin configurations responsible for the observed transport phenomena. These findings not only provide novel perspectives into the intricate spin textures of two-dimensional CrSBr but also underscore the sensitivity of tunneling as a probing technique for investigating the magnetic order in van der Waals materials.

2407.13231

Ngoc-Thanh Nguyen

Keila Lima, Ngoc-Thanh Nguyen, Rogardt Heldal, Lars Michael Kristensen, Tosin Daniel Oyetoyan, Patrizio Pelliccione, Eric Knauss

A data-flow oriented software architecture for heterogeneous marine data streams

The paper is accepted at the 21st IEEE International Conference on Software Architecture (ICSA 2024)

cs.SE

http://creativecommons.org/licenses/by/4.0/

Marine in-situ data is collected by sensors mounted on fixed or mobile systems deployed into the ocean. This type of data is crucial both for the ocean industries and public authorities, e.g., for monitoring and forecasting the state of marine ecosystems and/or climate changes. Various public organizations have collected, managed, and openly shared in-situ marine data in the past decade. Recently, initiatives like the Ocean Decade Corporate Data Group have incentivized the sharing of marine data of public interest from private companies aiding in ocean management. However, there is no clear understanding of the impact of data quality in the engineering of systems, as well as on how to manage and exploit the collected data. In this paper, we propose main architectural decisions and a data flow-oriented component and connector view for marine in-situ data streams. Our results are based on a longitudinal empirical software engineering process, and driven by knowledge extracted from the experts in the marine domain from public and private organizations, and challenges identified in the literature. The proposed software architecture is instantiated and exemplified in a prototype implementation.

2407.13232

Yuval Boneh

Yuval Boneh

Autonomous Money Supply Strategy Utilizing Control Theory

10.2139/ssrn.4844212

q-fin.RM q-fin.CP

http://creativecommons.org/licenses/by/4.0/

Decentralized Finance (DeFi) has reshaped the possibilities of reserve banking in the form of the Collateralized Debt Position (CDP). Key to the safety of CDPs is the money supply architecture that enables issued debt to maintain its value. In traditional markets, and with respect to the United States Dollar system, interest rates are set by the Federal Reserve in an attempt to influence the effects of excessive inflation. DeFi enables a more transparent approach that typically relies on interest rates or other debt recovery mechanisms being directly informed by asset price. This research investigates contemporary DeFi money supply and debt management strategies and their limitations. Furthermore, this paper introduces a time-weighted approach to interest rate management that implements a Proportional-Integral-Derivative control system to constantly adapt to market activities and protect the value of issued currency, while addressing observed limitations.

2407.13233

Tajron Juri\'c Dr. sc.

Axel Hrelja, Tajron Juri\'c and Filip Po\v{z}ar

Entropy of black holes, charged probes and noncommutative generalization

19 pages

hep-th gr-qc

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

The brick wall model is a semi-classical approach to understanding the microscopic origin of black hole entropy. We outline the formalism for the brick wall model in arbitrary number of dimensions and generalize it to include both charged spacetimes and charged probes in order to systematically show how to calculate the entropy for any black hole, in higher orders of the WKB approximation. We calculate the entropy for the Reissner-Nordstr\"{o}m and charged BTZ black holes, and by looking at the chargeless limits we recover the entropy of the Schwarzschild and neutral BTZ black holes. We also study noncommutative corrections to the black hole entropy by using a Drinfeld twist to deform spacetime symmetries. Using the noncommutative action for a charged scalar field, we derive the noncommutative Klein-Gordon equation and the radial equation in arbitrary dimension on which we generalize the brick wall method. We study case by case the noncommutative Reissner-Nordstr\"{o}m and noncommutative charged BTZ black holes, the entropy of which we calculate using the generalized brick wall method. Corrections due to higher order in the WKB approximation arise, and they are given by the logarithms of the black hole area. The corrections in the lowest order of WKB due to noncommutativity are also given by the logarithms of the black hole area.

2407.13234

Bruno F. Louren\c{c}o

Tianxiang Liu and Bruno F. Louren\c{c}o

Concrete convergence rates for common fixed point problems under Karamata regularity

44 pages, comments welcome

math.OC cs.NA math.FA math.MG math.NA

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

We introduce the notion of Karamata regular operators, which is a notion of regularity that is suitable for obtaining concrete convergence rates for common fixed point problems. This provides a broad framework that includes, but goes beyond, H\"olderian error bounds and H\"older regular operators. By concrete, we mean that the rates we obtain are explicitly expressed in terms of a function of the iteration number $k$ instead, of say, a function of the iterate $x^k$. While it is well-known that under H\"olderian-like assumptions many algorithms converge linearly/sublinearly (depending on the exponent), little it is known when the underlying problem data does not satisfy H\"olderian assumptions, which may happen if a problem involves exponentials and logarithms. Our main innovation is the usage of the theory of regularly varying functions which we showcase by obtaining concrete convergence rates for quasi-cylic algorithms in non-H\"olderian settings. This includes certain rates that are neither sublinear nor linear but sit somewhere in-between, including a case where the rate is expressed via the Lambert W function. Finally, we connect our discussion to o-minimal geometry and show that definable operators in any o-minimal structure are always Karamata regular.

2407.13235

Pasquale Bosso

Pasquale Bosso

Minimal-length quantum field theory: a first-principle approach

24 pages, 2 figures, 3 appendices

gr-qc hep-th

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Phenomenological models of quantum gravity often consider the existence of some form of minimal length. This feature is commonly described in the context of quantum mechanics and using the corresponding formalism and techniques. Although few attempts at a quantum field-theoretical description of a minimal length has been proposed, they are rather the exception and there is no general agreement on the correct one. Here, using the quantum-mechanical model as a guidance, we propose a first-principle definition of a quantum field theory including a minimal length. Specifically, we propose a two-step procedure, by first describing the quantum-mechanical models as a classical field theory and subsequently quantizing it. We are thus able to provide a foundation for further exploration of the implications of a minimal length in quantum field theory.

2407.13236

Dorian Martino

Dorian Martino

Regularity of unconstrained $p$-harmonic maps from curved domain and application to critical $n$-Laplace systems

23 pages

math.AP

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Given $p\geq 2$ and a map $g : B^n(0,1)\to S_n^{++}$, where $S_n^{++}$ is the group of positively definite matrices, we study critical points of the following functional: $$ v\in W^{1,p}\left(B^n(0,1);\mathbb{R}^N \right) \mapsto \int_{B^n(0,1)} |\nabla v|^p_g\, d\mathrm{vol}_g = \int_{B^n(0,1)} \left( g^{\alpha\beta}(x) \left\langle \partial_\alpha v(x), \partial_\beta v(x) \right\rangle \right)^{\frac{p}{2}}\, \sqrt{\det g(x)}\, dx. $$ We show that if $g$ is uniformly close to a constant matrix, then $v$ is locally H\"older-continuous. If $g$ is H\"older-continuous, we show that $\nabla v$ is locally H\"older-continuous. As an application, we specialize to the case $p=n$ and we prove that any continuous solution to $|\Delta_{g,n}u|\lesssim |\nabla u|^n_g$ satisfies additional regularity properties depending on the regularity of $g$.

2407.13237

Yun Qu

Boyuan Wang, Yun Qu, Yuhang Jiang, Jianzhun Shao, Chang Liu, Wenming Yang, Xiangyang Ji

LLM-Empowered State Representation for Reinforcement Learning

cs.AI

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Conventional state representations in reinforcement learning often omit critical task-related details, presenting a significant challenge for value networks in establishing accurate mappings from states to task rewards. Traditional methods typically depend on extensive sample learning to enrich state representations with task-specific information, which leads to low sample efficiency and high time costs. Recently, surging knowledgeable large language models (LLM) have provided promising substitutes for prior injection with minimal human intervention. Motivated by this, we propose LLM-Empowered State Representation (LESR), a novel approach that utilizes LLM to autonomously generate task-related state representation codes which help to enhance the continuity of network mappings and facilitate efficient training. Experimental results demonstrate LESR exhibits high sample efficiency and outperforms state-of-the-art baselines by an average of 29% in accumulated reward in Mujoco tasks and 30% in success rates in Gym-Robotics tasks.

2407.13238

Andreas Voskou

Andreas Voskou, Charalambos Christoforou, Sotirios Chatzis

Transformers with Stochastic Competition for Tabular Data Modelling

cs.LG

http://creativecommons.org/licenses/by/4.0/

Despite the prevalence and significance of tabular data across numerous industries and fields, it has been relatively underexplored in the realm of deep learning. Even today, neural networks are often overshadowed by techniques such as gradient boosted decision trees (GBDT). However, recent models are beginning to close this gap, outperforming GBDT in various setups and garnering increased attention in the field. Inspired by this development, we introduce a novel stochastic deep learning model specifically designed for tabular data. The foundation of this model is a Transformer-based architecture, carefully adapted to cater to the unique properties of tabular data through strategic architectural modifications and leveraging two forms of stochastic competition. First, we employ stochastic "Local Winner Takes All" units to promote generalization capacity through stochasticity and sparsity. Second, we introduce a novel embedding layer that selects among alternative linear embedding layers through a mechanism of stochastic competition. The effectiveness of the model is validated on a variety of widely-used, publicly available datasets. We demonstrate that, through the incorporation of these elements, our model yields high performance and marks a significant advancement in the application of deep learning to tabular data.

2407.13239

Cunliang Ma

Cunliang Ma, Weiguang Zhou and Zhoujian Cao

Gravitational Wave Mixture Separation for Future Gravitational Wave Observatories Utilizing Deep Learning

15 pages, 8 figures

astro-ph.IM

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Future GW observatories, such as the Einstein Telescope (ET), are expected to detect gravitational wave signals, some of which are likely to overlap with each other. This overlap may lead to misidentification as a single GW event, potentially biasing the estimated parameters of mixture GWs. In this paper, we adapt the concept of speech separation to address this issue by applying it to signal separation of overlapping GWs. We show that deep learning models can effectively separate overlapping GW signals. The proposed method may aid in eliminating biases in parameter estimation for such signals.

2407.13240

Steve Schneider

Marie-Laure Zollinger, Peter B. R{\o}nne, Steve Schneider, Peter Y. A. Ryan, Wojtek Jamroga

Intelligo ut Confido: Understanding, Trust and User Experience in Verifiable Receipt-Free E-Voting (long version)

cs.CR cs.HC

http://creativecommons.org/licenses/by/4.0/

Voting protocols seek to provide integrity and vote privacy in elections. To achieve integrity, procedures have been proposed allowing voters to verify their vote - however this impacts both the user experience and privacy. Especially, vote verification can lead to vote-buying or coercion, if an attacker can obtain documentation, i.e. a receipt, of the cast vote. Thus, some voting protocols go further and provide mechanisms to prevent such receipts. To be effective, this so-called receipt-freeness depends on voters being able to understand and use these mechanisms. In this paper, we present a study with 300 participants which aims to evaluate the voters' experience of the receipt-freeness procedures in the e-voting protocol Selene in the context of vote-buying. This actually constitutes the first user study dealing with vote-buying in e-voting. While the usability and trust factors were rated low in the experiments, we found a positive correlation between trust and understanding.

2407.13241

Hao Bai

Hao Bai, Yi Hong

NODER: Image Sequence Regression Based on Neural Ordinary Differential Equations

MICCAI2024

cs.CV cs.AI

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Regression on medical image sequences can capture temporal image pattern changes and predict images at missing or future time points. However, existing geodesic regression methods limit their regression performance by a strong underlying assumption of linear dynamics, while diffusion-based methods have high computational costs and lack constraints to preserve image topology. In this paper, we propose an optimization-based new framework called NODER, which leverages neural ordinary differential equations to capture complex underlying dynamics and reduces its high computational cost of handling high-dimensional image volumes by introducing the latent space. We compare our NODER with two recent regression methods, and the experimental results on ADNI and ACDC datasets demonstrate that our method achieves the state-of-the-art performance in 3D image regression. Our model needs only a couple of images in a sequence for prediction, which is practical, especially for clinical situations where extremely limited image time series are available for analysis. Our source code is available at https://github.com/ZedKing12138/NODER-pytorch.

2407.13242

Kyungyun Lee

Kyung Yun Lee, Nils Meyer-Kahlen, Georg G\"otz, U. Peter Svensson, Sebastian J. Schlecht, Vesa V\"alim\"aki

Fade-in Reverberation in Multi-room Environments Using the Common-Slope Model

2024 AES 5th International Conference on Audio for Virtual and Augmented Reality

eess.AS cs.SD

http://creativecommons.org/licenses/by/4.0/

In multi-room environments, modelling the sound propagation is complex due to the coupling of rooms and diverse source-receiver positions. A common scenario is when the source and the receiver are in different rooms without a clear line of sight. For such source-receiver configurations, an initial increase in energy is observed, referred to as the "fade-in" of reverberation. Based on recent work of representing inhomogeneous and anisotropic reverberation with common decay times, this work proposes an extended parametric model that enables the modelling of the fade-in phenomenon. The method performs fitting on the envelopes, instead of energy decay functions, and allows negative amplitudes of decaying exponentials. We evaluate the method on simulated and measured multi-room environments, where we show that the proposed approach can now model the fade-ins that were unrealisable with the previous method.

2407.13243

Ian G. Moss

Ian G Moss

Dark matter droplets

9 pages, no figures

astro-ph.CO cond-mat.quant-gas hep-th

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

A new model for dark matter is put forward which consists of uniform droplets of Bose Einstein condensate. In this model, structure forms rapidly, shortly after the hot big bang plasma de-ionises. The model also produces modifications to the expansion rate before droplet formation that affect the measurement of cosmological parameters from Cosmic Microwave Background data. The model could contribute to explaining why observations at high redshift see anomalously high structure formation and predict low values for the Hubble constant.

2407.13244

Alessandro Berti Mr

Alessandro Berti, Humam Kourani, Wil M.P. van der Aalst

PM-LLM-Benchmark: Evaluating Large Language Models on Process Mining Tasks

cs.CL cs.DB

http://creativecommons.org/licenses/by/4.0/

Large Language Models (LLMs) have the potential to semi-automate some process mining (PM) analyses. While commercial models are already adequate for many analytics tasks, the competitive level of open-source LLMs in PM tasks is unknown. In this paper, we propose PM-LLM-Benchmark, the first comprehensive benchmark for PM focusing on domain knowledge (process-mining-specific and process-specific) and on different implementation strategies. We focus also on the challenges in creating such a benchmark, related to the public availability of the data and on evaluation biases by the LLMs. Overall, we observe that most of the considered LLMs can perform some process mining tasks at a satisfactory level, but tiny models that would run on edge devices are still inadequate. We also conclude that while the proposed benchmark is useful for identifying LLMs that are adequate for process mining tasks, further research is needed to overcome the evaluation biases and perform a more thorough ranking of the competitive LLMs.

2407.13245

Jian Chen

Jian Chen, Liping Tang, Xinmin Yang

Descent Methods for Vector Optimization Problems: A Majorization-Minimization Perspective

math.OC

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

In this paper, we develop a unified framework and convergence analysis of descent methods for vector optimization problems (VOPs) from a majorization-minimization perspective. By choosing different surrogate functions, the generic method reduces to some existing descent methods with and without line search, respectively. The unified convergence analysis shows that the slow convergence of the steepest descent method is mainly due to the large gap between surrogate and objective functions. As a result, the performance of descent methods can be improved by narrowing the surrogate gap. Interestingly, we observe that selecting a tighter surrogate function is equivalent to using an appropriate base of the dual cone in the direction-finding subproblem. Furthermore, we use Barzilai-Borwein method to narrow the surrogate gap and devise a Barzilai-Borwein descent method for VOPs with polyhedral cone. By reformulating the subproblem, we provide a new insight into the Barzilai-Borwein descent method and bridge it to the steepest descent method. Finally, several numerical experiments confirm the efficiency of the proposed method.

2407.13246

Hongyuan Zhang

Yaqi Wang, Yifan Zhang, Xiaodiao Chen, Shuai Wang, Dahong Qian, Fan Ye, Feng Xu, Hongyuan Zhang, Qianni Zhang, Chengyu Wu, Yunxiang Li, Weiwei Cui, Shan Luo, Chengkai Wang, Tianhao Li, Yi Liu, Xiang Feng, Huiyu Zhou, Dongyun Liu, Qixuan Wang, Zhouhao Lin, Wei Song, Yuanlin Li, Bing Wang, Chunshi Wang, Qiupu Chen, Mingqian Li

STS MICCAI 2023 Challenge: Grand challenge on 2D and 3D semi-supervised tooth segmentation

cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Computer-aided design (CAD) tools are increasingly popular in modern dental practice, particularly for treatment planning or comprehensive prognosis evaluation. In particular, the 2D panoramic X-ray image efficiently detects invisible caries, impacted teeth and supernumerary teeth in children, while the 3D dental cone beam computed tomography (CBCT) is widely used in orthodontics and endodontics due to its low radiation dose. However, there is no open-access 2D public dataset for children's teeth and no open 3D dental CBCT dataset, which limits the development of automatic algorithms for segmenting teeth and analyzing diseases. The Semi-supervised Teeth Segmentation (STS) Challenge, a pioneering event in tooth segmentation, was held as a part of the MICCAI 2023 ToothFairy Workshop on the Alibaba Tianchi platform. This challenge aims to investigate effective semi-supervised tooth segmentation algorithms to advance the field of dentistry. In this challenge, we provide two modalities including the 2D panoramic X-ray images and the 3D CBCT tooth volumes. In Task 1, the goal was to segment tooth regions in panoramic X-ray images of both adult and pediatric teeth. Task 2 involved segmenting tooth sections using CBCT volumes. Limited labelled images with mostly unlabelled ones were provided in this challenge prompt using semi-supervised algorithms for training. In the preliminary round, the challenge received registration and result submission by 434 teams, with 64 advancing to the final round. This paper summarizes the diverse methods employed by the top-ranking teams in the STS MICCAI 2023 Challenge.

2407.13247

Jurjen De Jong

J.M.G.H.J. de Jong, R.J. van Weeren, F. Sweijen, J.B.R. Oonk, T.W. Shimwell, A.R. Offringa, L.K. Morabito, H.J.A. R\"ottgering, R. Kondapally, E.L. Escott, P.N. Best, M. Bondi, H. Ye, J.W. Petley

Into the depths: Unveiling ELAIS-N1 with LOFAR's deepest sub-arcsecond wide-field images

Images and catalogues will be available on https://lofar-surveys.org/hd-en1.html

astro-ph.IM astro-ph.CO

http://creativecommons.org/licenses/by/4.0/

We present the deepest wide-field 115-166 MHz image at sub-arcsecond resolution spanning an area of 2.5 by 2.5 degrees centred at the ELAIS-N1 deep field. To achieve this, we improved the calibration for the International LOFAR Telescope. This enhancement enabled us to efficiently process 32 hrs of data from four different 8-hr observations using the high-band antennas (HBAs) of all 52 stations, covering baselines up to approximately 2,000 km across Europe. The DI calibration was improved by using an accurate sky model and refining the series of calibration steps on the in-field calibrator, while the DD calibration was improved by adopting a more automated approach for selecting the DD calibrators and inspecting the self-calibration on these sources. We also added an additional round of self-calibration for the Dutch core and remote stations in order to refine the solutions for shorter baselines. To complement our highest resolution at 0.3", we also made intermediate resolution wide-field images at 0.6" and 1.2". Our resulting wide-field images achieve a central noise level of 14 muJy/beam at 0.3", doubling the depth and uncovering four times more objects than the Lockman Hole deep field image at comparable resolution but with only 8 hrs of data. Compared to LOFAR imaging without the international stations, we note that due to the increased collecting area and the absence of confusion noise, we reached a point-source sensitivity comparable to a 500-hr ELAIS-N1 6" image with 16 times less observing time. Importantly, we have found that the computing costs for the same amount of data are almost halved (to about 139,000 CPU hrs per 8 hrs of data) compared to previous efforts, though they remain high. Our work underscores the value and feasibility of exploiting all Dutch and international LOFAR stations to make deep wide-field images at sub-arcsecond resolution.

2407.13248

Yufei Tian

Yufei Tian, Tenghao Huang, Miri Liu, Derek Jiang, Alexander Spangher, Muhao Chen, Jonathan May, Nanyun Peng

Are Large Language Models Capable of Generating Human-Level Narratives?

cs.CL

http://creativecommons.org/licenses/by/4.0/

This paper investigates the capability of LLMs in storytelling, focusing on narrative development and plot progression. We introduce a novel computational framework to analyze narratives through three discourse-level aspects: i) story arcs, ii) turning points, and iii) affective dimensions, including arousal and valence. By leveraging expert and automatic annotations, we uncover significant discrepancies between the LLM- and human- written stories. While human-written stories are suspenseful, arousing, and diverse in narrative structures, LLM stories are homogeneously positive and lack tension. Next, we measure narrative reasoning skills as a precursor to generative capacities, concluding that most LLMs fall short of human abilities in discourse understanding. Finally, we show that explicit integration of aforementioned discourse features can enhance storytelling, as is demonstrated by over 40% improvement in neural storytelling in terms of diversity, suspense, and arousal.

2407.13249

Richard Maximilian Milbradt

Richard M. Milbradt, Qunsheng Huang, Christian B. Mendl

PyTreeNet: A Python Library for easy Utilisation of Tree Tensor Networks

(53 pages, 17 figures, code available at https://github.com/Drachier/PyTreeNet)

quant-ph cond-mat.str-el

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

In recent years, tree tensor network methods have proven capable of simulating quantum many-body and other high-dimensional systems. This work is a user guide to our Python library PyTreeNet. It includes code examples and exercises to introduce the library's functions and familiarise the reader with the concepts and methods surrounding tree tensor networks. PyTreeNet implements all the tools required to implement general tree tensor network methods, such as tensor decompositions and arbitrary tree structures. The main focus is on the time evolution of quantum systems. This includes an introduction to tree tensor network states and operators and the time-evolving block decimation and time-dependent variational principle. The library's capabilities are showcased with the example of a modified transverse field Ising model on tree structures that go far beyond the ability of common state vector methods.

2407.13250

Glen Wheeler

Piotr Rybka and Glen Wheeler

Complete classification of solitons for the surface diffusion flow of entire graphs

11 pages

math.DG math.AP

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

In this article we completely classify solitons (equilibria, self-similar solutions and travelling waves) for the surface diffusion flow of entire graphs of function over the real line.

2407.13251

Chunjing Xiao

Chunjing Xiao, Shikang Pang, Wenxin Tai, Yanlong Huang, Goce Trajcevski, Fan Zhou

Motif-Consistent Counterfactuals with Adversarial Refinement for Graph-Level Anomaly Detection

Accepted by KDD 2024

10.1145/3637528.3672050

cs.LG cs.SI

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Graph-level anomaly detection is significant in diverse domains. To improve detection performance, counterfactual graphs have been exploited to benefit the generalization capacity by learning causal relations. Most existing studies directly introduce perturbations (e.g., flipping edges) to generate counterfactual graphs, which are prone to alter the semantics of generated examples and make them off the data manifold, resulting in sub-optimal performance. To address these issues, we propose a novel approach, Motif-consistent Counterfactuals with Adversarial Refinement (MotifCAR), for graph-level anomaly detection. The model combines the motif of one graph, the core subgraph containing the identification (category) information, and the contextual subgraph (non-motif) of another graph to produce a raw counterfactual graph. However, the produced raw graph might be distorted and cannot satisfy the important counterfactual properties: Realism, Validity, Proximity and Sparsity. Towards that, we present a Generative Adversarial Network (GAN)-based graph optimizer to refine the raw counterfactual graphs. It adopts the discriminator to guide the generator to generate graphs close to realistic data, i.e., meet the property Realism. Further, we design the motif consistency to force the motif of the generated graphs to be consistent with the realistic graphs, meeting the property Validity. Also, we devise the contextual loss and connection loss to control the contextual subgraph and the newly added links to meet the properties Proximity and Sparsity. As a result, the model can generate high-quality counterfactual graphs. Experiments demonstrate the superiority of MotifCAR.

2407.13252

Qiao Li

Qiao Li, Xiaomeng Fu, Xi Wang, Jin Liu, Xingyu Gao, Jiao Dai, Jizhong Han

Unveiling Structural Memorization: Structural Membership Inference Attack for Text-to-Image Diffusion Models

cs.CV

http://arxiv.org/licenses/nonexclusive-distrib/1.0/

With the rapid advancements of large-scale text-to-image diffusion models, various practical applications have emerged, bringing significant convenience to society. However, model developers may misuse the unauthorized data to train diffusion models. These data are at risk of being memorized by the models, thus potentially violating citizens' privacy rights. Therefore, in order to judge whether a specific image is utilized as a member of a model's training set, Membership Inference Attack (MIA) is proposed to serve as a tool for privacy protection. Current MIA methods predominantly utilize pixel-wise comparisons as distinguishing clues, considering the pixel-level memorization characteristic of diffusion models. However, it is practically impossible for text-to-image models to memorize all the pixel-level information in massive training sets. Therefore, we move to the more advanced structure-level memorization. Observations on the diffusion process show that the structures of members are better preserved compared to those of nonmembers, indicating that diffusion models possess the capability to remember the structures of member images from training sets. Drawing on these insights, we propose a simple yet effective MIA method tailored for text-to-image diffusion models. Extensive experimental results validate the efficacy of our approach. Compared to current pixel-level baselines, our approach not only achieves state-of-the-art performance but also demonstrates remarkable robustness against various distortions.