arxiv_id stringlengths 12 12 | title stringlengths 19 176 | text stringlengths 487 2k | spans listlengths 1 14 |
|---|---|---|---|
2607.22972v1 | Learned Interventions in Lean 4 grind | Lean~4's \grind{} tactic combines congruence closure, \ematch{}ing, and case-splitting into a single automated solver, and like any such solver, it relies on hand-tuned heuristics to decide what to instantiate and where to case-split. These heuristics are tempting targets for learning, but there is a catch: because \gr... | [
{
"start": 797,
"end": 862,
"label": "Result",
"text": "A lookahead step, proves five theorems it otherwise times out on."
},
{
"start": 925,
"end": 1028,
"label": "Result",
"text": "across four feature-based models, statically predicting the correct case split is no better than ... |
2607.22351v1 | IQ-JEPA: A Joint-Embedding Predictive Architecture with a Hermitian Vision Transformer for Sound Speed and Attenuation Estimation from Ultrasound IQ Data | The speed of sound in tissue is a prerequisite for well-focused imaging and has diagnostic value, but recovering it from raw pulse-echo channel data is fundamentally a nonlinear inverse problem. Learned solvers are fast yet label hungry. Simulated sound-speed labels are expensive, while abundant real channel data is un... | [
{
"start": 329,
"end": 375,
"label": "Contribution",
"text": "We propose IQ-JEPA to exploit both data types."
},
{
"start": 376,
"end": 556,
"label": "Method",
"text": "An encoder is pretrained without labels to predict the latent representation of masked in-phase and quadrature ... |
2606.23057v1 | Who Owns the AI Recommendation? A Multi-Industry Empirical Map of Brand Category Ownership Across Large Language Models | Large language models now mediate how buyers discover products and services, making the competitive structure of AI-generated recommendations a strategic concern for brands. A basic question has lacked large-scale empirical answers: in a given category, which brand does a model recommend, and how concentrated is that o... | [
{
"start": 330,
"end": 569,
"label": "Method",
"text": "Across 3,750 responses spanning 50 brands, five industries, and 250 brand-free category queries on three models (GPT-5.2, Google Gemini 3 Flash, and Perplexity sonar-pro), each query repeated five times under a dice-roll stability protocol"
}... |
2607.22268v1 | General Value Functions for Remaining Useful Life and Failure-Mode Prediction | Remaining useful life (RUL) prediction and failure-mode classification are central tasks in predictive maintenance. Many data-driven pipelines use fixed-window supervised learning with complete terminal labels; such routes do not naturally encode the temporal recursion linking successive degradation-state predictions w... | [
{
"start": 385,
"end": 617,
"label": "Contribution",
"text": "We formulate prognostics as vector General Value Function (GVF) prediction on an absorbing degradation process, treating RUL and failure-mode probabilities as temporally consistent targets rather than independent window-level labels"
},... |
2607.21995v1 | QC-PHAST Search: Classical--Quantum Query Benchmarks for Finite-Pool Rare-Regime Discovery | Rare-regime discovery in parameterized dynamical systems is an active-search problem: find one verified parameter at which a scientifically defined qualitative threshold is crossed, even when acceptable candidates are rare, nonconvex, or fragmented. We introduce Quantum-Classical Phase-space and Stability-Threshold Sea... | [
{
"start": 250,
"end": 334,
"label": "Contribution",
"text": "We introduce Quantum-Classical Phase-space and Stability-Threshold Search (QC-PHAST)"
},
{
"start": 435,
"end": 545,
"label": "Method",
"text": "A candidate induces a dynamical object, simulator-derived criticality sco... |
2607.23237v1 | Context-Aware Concept Distillation for Trustworthy Flood Prediction | Effective flood risk management relies on accurate forecasting, yet the "black box" nature of stateof-the-art Deep Learning models creates a barrier to trust and accountability in high-stakes public safety decisions. While existing Explainable AI (XAI) methods offer local attributions, they fail to provide the verifiab... | [
{
"start": 446,
"end": 636,
"label": "Contribution",
"text": "we propose Context-Aware Concept Distillation (CACD), a framework developed in collaboration with domain experts to distill opaque LSTMs into interpretable, hydrology-aware surrogate models."
},
{
"start": 822,
"end": 908,
... |
2607.20933v1 | Transformer-Assisted LLM-Based Source Code Summarisation: to Enable More Secure Software Development | Neural Source Code Summarisation (NSCS) aims to generate natural language summaries of source code to improve developers' and maintainers' understanding of code. Source code summaries are vital during the maintenance phase of the Secure Software Development Lifecycle (SSDLC), as they improve maintainers' understanding ... | [
{
"start": 1335,
"end": 1494,
"label": "Contribution",
"text": "We show how combining these two methods, by using Transformer-generated summaries in prompt engineering, may enable LLMs to create better source code summaries"
},
{
"start": 1552,
"end": 1678,
"label": "Method",
"te... |
2607.22201v1 | Trajectory-Regularized Stochastic Optimal Control via KL Divergence | We introduce trajectory-regularized stochastic optimal control (TRSOC), which augments standard stochastic optimal control (SOC) with a Kullback--Leibler (KL) divergence between controlled and reference trajectory distributions. Using Girsanov's theorem, the trajectory KL reduces to a quadratic drift mismatch penalty, ... | [
{
"start": 0,
"end": 228,
"label": "Contribution",
"text": "We introduce trajectory-regularized stochastic optimal control (TRSOC), which augments standard stochastic optimal control (SOC) with a Kullback--Leibler (KL) divergence between controlled and reference trajectory distributions."
},
{
... |
2607.23278v1 | Co-Evolving Graph and Text Memory for Training-Free Multi-Hop Question Answering | Multi-hop question answering requires coordinating relational and textual evidence across reasoning steps, a combination neither a text corpus nor a knowledge graph can supply alone. Prior work often emphasizes only part of this loop: graph-augmented RAG retrieves from a pre-built or query-updated graph, KGQA systems s... | [
{
"start": 482,
"end": 588,
"label": "Contribution",
"text": "We propose Co-E, a training-free system built around synchronized bidirectional graph-text working memory."
},
{
"start": 589,
"end": 746,
"label": "Method",
"text": "A synchronization cycle consolidates textual memory... |
2607.16619v2 | SAGE: A Socially-Aware Generative Engine for Heterogeneous Multi-Agent Navigation | Safe and socially compliant navigation in open human-robot environments requires robots to reason about heterogeneous participants with different dynamics, autonomy levels, and social roles. Existing trajectory prediction and planning methods often rely on homogeneous interaction assumptions or enforce only geometric c... | [
{
"start": 456,
"end": 558,
"label": "Contribution",
"text": "This paper proposes SAGE, a socially-aware generative engine for heterogeneous multi-agent navigation."
},
{
"start": 559,
"end": 741,
"label": "Method",
"text": "SAGE represents robots and surrounding entities as a di... |
2607.21817v1 | Longitudinal Random Forests for Sparse and Irregular Response Trajectories | Longitudinal studies often collect data at sparse, irregular, and unequally spaced time points. Such heterogeneity is often driven by subject-specific covariates, yet existing methods have been restricted to a scalar endpoint value, completely neglecting the underlying response trajectories. We propose a novel Longitud... | [
{
"start": 293,
"end": 354,
"label": "Contribution",
"text": "We propose a novel Longitudinal Random Forest (LRF) framework"
},
{
"start": 360,
"end": 466,
"label": "Method",
"text": "leverages tree-based ensemble machine learning with adaptive node-wise longitudinal trajectory e... |
2607.22287v1 | Efficient Recommendations via Graph Coarsening and Label Propagation | Graph-based recommendations are widely adopted in real-world industrial applications. However, graphs in these systems often reach a massive scale, posing notable scalability and efficiency challenges. This requires techniques that can effectively balance predictive quality with computational cost. One promising approa... | [
{
"start": 529,
"end": 580,
"label": "Contribution",
"text": "we propose a flexible two-stage diffusion framework"
},
{
"start": 595,
"end": 679,
"label": "Method",
"text": "graph coarsening with multi-step label propagation in the telecommunications domain."
},
{
"start"... |
2607.22880v1 | Do Coverage and Mutation Scores of LLM-Generated Test Suites Correlate with Their Effectiveness? (Replicability Study) | Recent advances in large language models (LLMs) have driven growing interest in using LLMs to automate test generation. Prior work commonly evaluates generated test suites using proxy metrics such as code coverage and mutation score. However, studies by Inozemtseva et al. and Papadakis et al. show that, for human-writt... | [
{
"start": 730,
"end": 866,
"label": "Contribution",
"text": "we conduct a large-scale replication study of these two prior works using a wide range of test suites generated by a diverse set of LLMs"
},
{
"start": 965,
"end": 1019,
"label": "Result",
"text": "Our findings diverge... |
2606.26079v1 | Same Evidence, Different Answer: Auditing Order Sensitivity in Multimodal Large Language Models | Standard benchmarks for multimodal large language models (MLLMs) score each item on one canonical ordering and miss whether order-irrelevant shuffling changes the answer, a baseline reliability property called for by emerging AI evaluation guidelines. We introduce Facet-Probe, a five-facet audit (option, evidence-chunk... | [
{
"start": 252,
"end": 414,
"label": "Contribution",
"text": "We introduce Facet-Probe, a five-facet audit (option, evidence-chunk, document-rank, image-set, and mixed-modality ordering) of 18 frontier and open-weight MLLMs."
},
{
"start": 415,
"end": 578,
"label": "Method",
"tex... |
2607.21120v1 | Relative Value Learning | In reinforcement learning, critics typically estimate absolute state values $V(s)$, estimating how good a particular situation is in isolation. However, it turns out that only differences in value are relevant for control. Motivated by this, we propose Relative Value Learning (RV), a framework that learns value differe... | [
{
"start": 242,
"end": 363,
"label": "Contribution",
"text": "we propose Relative Value Learning (RV), a framework that learns value differences directly via an antisymmetric function"
},
{
"start": 599,
"end": 725,
"label": "Method",
"text": "reconstruct generalized advantage es... |
2607.18200v1 | Learning Adaptive Safety Margins for Visual Navigation | Robots in cluttered indoor spaces often fail not because they cannot generate collision-free paths, but because a fixed safety margin is mis-calibrated: conservative margins cause detours and timeouts, while permissive margins lead to near-boundary shortcuts under perception bias. Diffusion-based planners propose diver... | [
{
"start": 415,
"end": 538,
"label": "Contribution",
"text": "We propose a context-conditioned safety critic that learns an adaptive clearance preference for ranking diffusion proposals"
},
{
"start": 540,
"end": 1002,
"label": "Method",
"text": "decomposed into three complementa... |
2607.22342v1 | Beyond Binary Rooftop Mapping: A Four-Class Deep Learning Framework for Green Roof Potential Assessment from Open Swiss Geospatial Data | The development of effective urban climate adaptation strategies requires comprehensive spatial information on rooftops and buildings, since such information underpins the assessment of ecosystem services provided by green infrastructure, particularly for urban heat island (UHI) mitigation. Although green roofs are wid... | [
{
"start": 506,
"end": 701,
"label": "Contribution",
"text": "This study presents a modified deep convolutional neural network rooftop classification framework based on Roofpedia, developed by the Urban Analytics Lab at the National University of Singapore."
},
{
"start": 702,
"end": 828... |
2607.23285v1 | Photonic reservoir computing with complex networks | Photonic reservoir computing has attracted increasing attention as a fast and low-cost approach for time-series prediction. Photonic reservoir computing utilizes the high speed, broad bandwidth, and spatial parallelism of light. However, the effect of the internal connection structure (network topology) on the computin... | [
{
"start": 124,
"end": 228,
"label": "Method",
"text": "Photonic reservoir computing utilizes the high speed, broad bandwidth, and spatial parallelism of light."
},
{
"start": 412,
"end": 635,
"label": "Contribution",
"text": "we experimentally and numerically demonstrate photoni... |
2607.22083v1 | Nanbeige4.2-3B: Unlocking Agentic Capabilities in a Compact Mode | We present Nanbeige4.2-3B, a compact general agentic model with 3B non-embedding parameters. It delivers strong performance across code-agent, office-agent, and complex tool-use tasks while maintaining highly competitive reasoning capabilities in mathematics, coding, and science. Nanbeige4.2-3B is pretrained from scrat... | [
{
"start": 0,
"end": 58,
"label": "Contribution",
"text": "We present Nanbeige4.2-3B, a compact general agentic model"
},
{
"start": 93,
"end": 280,
"label": "Method",
"text": "It delivers strong performance across code-agent, office-agent, and complex tool-use tasks while mainta... |
2607.21372v1 | Mean-to-Score Discrete Diffusion: Posterior-Mean Denoisers for Score Entropy | Score Entropy Discrete Diffusion (SEDD) parameterizes discrete reverse processes with unconstrained positive score ratios. While positivity guarantees nonnegative reverse jump rates, it does not ensure Bayes realizability: ratios at a noisy state need not be jointly induced by any clean-token posterior under the forwar... | [
{
"start": 0,
"end": 122,
"label": "Method",
"text": "Score Entropy Discrete Diffusion (SEDD) parameterizes discrete reverse processes with unconstrained positive score ratios."
},
{
"start": 123,
"end": 329,
"label": "Limitation",
"text": "While positivity guarantees nonnegative... |
2607.22929v1 | Distribution-Specific Curvature Control with Finite-Sample Guarantees for Open-Weight Safety | A short fine-tuning run can undo the safety guards of an open-weight model---retraining a refusal-trained assistant to aid weapons development or produce hate speech. Preventing such harmful fine-tuning while retaining benign adaptability remains difficult: the only prior method with an explicit curvature certificate, ... | [
{
"start": 167,
"end": 256,
"label": "Limitation",
"text": "Preventing such harmful fine-tuning while retaining benign adaptability remains difficult"
},
{
"start": 441,
"end": 461,
"label": "Contribution",
"text": "We propose HarmAlign"
},
{
"start": 463,
"end": 568,... |
2607.23310v1 | Online Fair Division with Budget Constraints | We study an online variant of discrete fair division under generalized assignment budget constraints. Goods arrive one at a time and must be assigned irrevocably to a feasible agent or to charity, which holds all unallocated goods, while fairness is evaluated only against budget-feasible subsets of every recipient's bu... | [
{
"start": 0,
"end": 101,
"label": "Contribution",
"text": "We study an online variant of discrete fair division under generalized assignment budget constraints."
},
{
"start": 102,
"end": 325,
"label": "Method",
"text": "Goods arrive one at a time and must be assigned irrevocabl... |
2607.21016v1 | CultureTalk-ID: A Multi-Task Dialogue Benchmark for Cultural Commonsense in Indonesian Local Languages | Culture is lived through conversation, yet existing Indonesian cultural commonsense benchmarks evaluate LLMs on short and isolated prompts, stripping away the dialogic context in which cultural nuances actually surface. We introduce CultureTalk-ID, the first dialogue-based benchmark for cultural commonsense in Indonesi... | [
{
"start": 220,
"end": 247,
"label": "Contribution",
"text": "We introduce CultureTalk-ID"
},
{
"start": 348,
"end": 538,
"label": "Method",
"text": "comprising 4,496 culturally grounded dialogues across 11 languages and 13 culturally salient topics, curated through a multi-stage... |
2607.23198v1 | Variance-Preserving Orthogonal Selection (VPOS): Greedy Feature Selection via Orthogonal Deflation in PCA Loading Space | We propose Variance-Preserving Orthogonal Selection (VPOS), a greedy framework for unsupervised feature selection that operates in the weighted PCA loading space. After each selection, VPOS projects out the chosen feature's variance direction via null-space deflation, forcing subsequent selections to cover orthogonal p... | [
{
"start": 0,
"end": 113,
"label": "Contribution",
"text": "We propose Variance-Preserving Orthogonal Selection (VPOS), a greedy framework for unsupervised feature selection"
},
{
"start": 119,
"end": 353,
"label": "Method",
"text": "operates in the weighted PCA loading space. Af... |
2607.22039v1 | Enough is as good as a feast: A Comprehensive Analysis of How Reinforcement Learning Mitigates Task Conflicts in LLMs | Model merging plays a crucial role in consolidating multiple specialized models into a single, unified model, especially in the era of large language models (LLMs). Recent research has primarily focused on developing strategies to enhance merging performance with the trained models, while the impact of training paradig... | [
{
"start": 470,
"end": 582,
"label": "Contribution",
"text": "we systematically explore the merging behavior of RL-trained LLMs compared to those trained with traditional SFT"
},
{
"start": 584,
"end": 650,
"label": "Method",
"text": "Through comprehensive evaluations across five... |
2607.16943v1 | SinD 2.0: A Multi-City UAV Dataset with Semantic Risk Annotations for SOTIF-Oriented Safety Validation at Signalized Intersections | Safety validation at signalized intersections remains a critical bottleneck for the deployment of autonomous driving systems (ADS), as these scenarios involve dense heterogeneous traffic, contested right of way, and long-tail safety-critical interactions, posing significant challenges to the Safety of the Intended Func... | [
{
"start": 589,
"end": 610,
"label": "Method",
"text": "To address these gaps"
},
{
"start": 612,
"end": 642,
"label": "Contribution",
"text": "this paper introduces SinD 2.0"
},
{
"start": 644,
"end": 737,
"label": "Method",
"text": "a large-scale drone-based... |
2607.22068v1 | Rethinking Multi-Branch and Cross-Backbone Fusion for Vehicle Re-Identification in the Foundation-Model Era | Multi-branch architectures and CNN-Transformer fusion have long been regarded as effective ways to improve vehicle re-identification (Re-ID) by combining complementary representations. In this work, we revisit this assumption in the foundation-model era through a comprehensive empirical study. A single DINOv3-pretraine... | [
{
"start": 199,
"end": 293,
"label": "Contribution",
"text": "we revisit this assumption in the foundation-model era through a comprehensive empirical study"
},
{
"start": 295,
"end": 358,
"label": "Method",
"text": "A single DINOv3-pretrained ConvNeXt trained with a tuned recipe... |
2607.21961v1 | On Improving Faithfulness of Podcasts from Documents | Large language models (LLMs) are increasingly used to generate long-form conversational content such as podcasts from textual sources. While these systems produce fluent and engaging narratives, they often introduce ungrounded information. In this work, we present the first systematic study of faithfulness in document-... | [
{
"start": 254,
"end": 347,
"label": "Contribution",
"text": "we present the first systematic study of faithfulness in document-grounded podcast generation"
},
{
"start": 453,
"end": 574,
"label": "Contribution",
"text": "We construct a dataset of over 1500 documents spanning fiv... |
2607.21426v1 | Semantic-Aware Task Clustering for Constructive and Cooperative Multi-Tasking | Cooperative multi-task semantic communication (CMT-SemCom) improves task execution performance by leveraging shared representations. However, as we demonstrated in [1], cooperative multi-tasking can be either constructive or destructive, depending on the semantic relationships among tasks. To ensure constructive cooper... | [
{
"start": 327,
"end": 392,
"label": "Contribution",
"text": "we propose a semantic-aware task clustering method for CMT-SemCom"
},
{
"start": 662,
"end": 851,
"label": "Method",
"text": "the problem decomposes into two stages: (i) a semantic clustering problem leveraging hierarc... |
2607.21570v1 | MedGame: Storytelling Gamification Empowered by Large Language Models for Medical Education | Large Language Models (LLMs) show promise for medical education, but most existing systems focus on localized interactions such as question answering or single-turn feedback, rather than organizing an entire clinical case into a decision-centered learning trajectory. We introduce \textit{MedGame}, a framework that tran... | [
{
"start": 432,
"end": 537,
"label": "Method",
"text": "a Medical Narrative Designer synthesizes case-grounded clinical storylines with states and decision nodes"
},
{
"start": 539,
"end": 675,
"label": "Method",
"text": "while a Story Director converts them into dependency-aware... |
2607.21332v1 | Phonetic forced alignment for low-resource language varieties: Model training and evaluation on Chengdu Mandarin | Phonetic forced alignment is a key technique in phonetic research, yet existing alignment systems lack specialized models for low-resource language varieties. We address this by training text-dependent and text-independent aligners for Chengdu Mandarin using a 17-hour corpus and a custom G2P dictionary. We trained a te... | [
{
"start": 159,
"end": 304,
"label": "Contribution",
"text": "We address this by training text-dependent and text-independent aligners for Chengdu Mandarin using a 17-hour corpus and a custom G2P dictionary."
},
{
"start": 305,
"end": 503,
"label": "Method",
"text": "We trained a... |
2607.21577v1 | Synthetic data generation framework for quality control automation in gravure printing | Quality control in printing, particularly in rotogravure printing, still depends on slow, costly, and subjective manual inspection. Automated surface defect detection is critical for maintaining high-quality standards in rotogravure printing. Deep learning models give prospects for automation. However, training robust ... | [
{
"start": 464,
"end": 492,
"label": "Limitation",
"text": "To overcome this limitation,"
},
{
"start": 493,
"end": 608,
"label": "Contribution",
"text": "this paper introduces a novel synthetic data generation framework tailored for rotogravure printing quality control"
},
{... |
2607.21975v1 | On the Convergence of Stochastic Low-Rank Adaptation | Low-rank adaptation (LoRA) optimizes $J(B,A)=\mathcal L(W_\mathrm{base}+sBA)$ over two adapters $B \in \mathbb{R}^{m \times r}$ and $A \in \mathbb{R}^{r \times n}$ that form a low-rank update to a frozen pretrained weight matrix $W_\mathrm{base} \in \mathbb{R}^{m \times n}$. The prior analysis shows LoRA-GD takes $\exp... | [
{
"start": 456,
"end": 479,
"label": "Contribution",
"text": "We sharpen the analysis"
},
{
"start": 586,
"end": 673,
"label": "Contribution",
"text": "We further study stochastic LoRA under unbiased gradient estimates and finite variance."
},
{
"start": 790,
"end": 9... |
2607.23029v1 | Multi-Agent Privacy Game in Federated Learning: A Unified Mean-Field View | Federated learning enables collaborative model training across distributed clients without centralising their data, yet privacy remains a persistent concern because the shared model updates can leak information about local datasets. Existing privacy-preserving methods either inject calibrated noise into client updates,... | [
{
"start": 493,
"end": 608,
"label": "Contribution",
"text": "We bridge these two lines of work by formulating privacy-preserving federated learning as a mean-field privacy game"
},
{
"start": 530,
"end": 748,
"label": "Method",
"text": "formulating privacy-preserving federated l... |
2607.21173v1 | Automated Synthesis and Adversarial Validation of Executable Causal Research Pipelines | While automated research systems promise to accelerate empirical analysis, they are prone to silent failures: instances in which analysis code executes successfully yet relies on invalid causal assumptions. We present the Artificial Intelligence (AI)-based Epidemiology Research Assistant (ARA), a framework that makes t... | [
{
"start": 207,
"end": 450,
"label": "Contribution",
"text": "We present the Artificial Intelligence (AI)-based Epidemiology Research Assistant (ARA), a framework that makes these failures visible by explicitly encoding causal design principles, study-specific assumptions, and methodological constra... |
2607.22890v1 | Meshless Domain Randomization via Explicit Parameter Perturbation of 3D Gaussian Splatting | Domain Randomization (DR) is a standard technique for closing the Sim-to-Real gap, yet traditional DR pipelines rely on classical computer graphics rendering driven by polygon meshes. For complex organic subjects, such as insect specimens, extracting and rendering textured meshes is challenging. To address this issue, ... | [
{
"start": 320,
"end": 423,
"label": "Contribution",
"text": "we propose a meshless DR framework that operates on the parameter space of 3D Gaussian Splatting (3DGS)"
},
{
"start": 425,
"end": 526,
"label": "Method",
"text": "Our method employs two independent perturbation pipeli... |
2607.21162v1 | Agree on the Model, Verify the Inference: GKR Protocols for HND-Based Transformer Inference | Outsourced Transformer inference exposes clients to model substitution and incomplete execution, while direct replay removes the computational benefit of delegation. We present GKR-HND, a registered-model protocol for verifying the polynomial backbone of Homomorphic--Nonhomomorphic Decomposition Transformers. The retai... | [
{
"start": 166,
"end": 310,
"label": "Contribution",
"text": "We present GKR-HND, a registered-model protocol for verifying the polynomial backbone of Homomorphic--Nonhomomorphic Decomposition Transformers."
},
{
"start": 311,
"end": 468,
"label": "Method",
"text": "The retained ... |
2607.23333v1 | Training with (Swap) Regret Loss in a Single-Layer Self-Attention Model: A Case Study on the Probability Simplex | We revisit the regret loss framework introduced in Park et al. (2025), which uses decision-theoretic regret as a direct loss function for training models to make better decisions, through the lens of probability-simplex policies. Our first result shows that a single-layer self-attention model trained with regret loss a... | [
{
"start": 0,
"end": 69,
"label": "Contribution",
"text": "We revisit the regret loss framework introduced in Park et al. (2025)"
},
{
"start": 180,
"end": 228,
"label": "Method",
"text": "through the lens of probability-simplex policies"
},
{
"start": 230,
"end": 344... |
2607.03833v1 | Beyond Static Rules: Automated Discovery of Latent Vulnerabilities in Text-to-SQL | While Large Language Models (LLMs) have achieved remarkable success in Text-to-SQL tasks, their deployment in real-world environments is hindered by latent reliability issues. Identifying these latent weaknesses is critical for building trustworthy database interfaces, yet current diagnostic approaches rely heavily on ... | [
{
"start": 438,
"end": 607,
"label": "Contribution",
"text": "we propose SAGE (Systematic Automated Guided Exploration), a novel framework designed to autonomously uncover latent failure patterns in LLM-based Text-to-SQL generation"
},
{
"start": 623,
"end": 837,
"label": "Method",
... |
2607.21671v1 | Neural Feature Governance: Extending Atom Prevalence | Neural network compression and interpretability remain open challenges in modern deep learn- ing, where billion-parameter architectures deliver impressive accuracy at the cost of trans- parency, computational efficiency, and reliable uncertainty quantification. This paper introduces Neural Atom Prevalence (NAP), a prin... | [
{
"start": 262,
"end": 419,
"label": "Contribution",
"text": "This paper introduces Neural Atom Prevalence (NAP), a principled Bayesian framework for structured node-level model selection in feedforward neural networks."
},
{
"start": 420,
"end": 851,
"label": "Method",
"text": "... |
2607.13413v1 | Is the Statistical Advantage Worth the Cost? An Empirical Comparison of KANs and MLPs for Structured Data Classification | This study presents an empirical benchmarking comparison between Kolmogorov-Arnold Networks (KANs) and Multi-Layer Perceptrons (MLPs) on structured tabular classification tasks. Motivated by the growing interest in KANs as an alternative function-approximating architecture, we evaluate their out-of-the-box performance ... | [
{
"start": 0,
"end": 133,
"label": "Contribution",
"text": "This study presents an empirical benchmarking comparison between Kolmogorov-Arnold Networks (KANs) and Multi-Layer Perceptrons (MLPs)"
},
{
"start": 275,
"end": 419,
"label": "Method",
"text": "we evaluate their out-of-t... |
2607.21341v1 | Grasp, Handover, Rotate: Bimanual Object Reorientation via Compositional Diffusion and Energy-Based Optimization | Bimanual object reorientation - picking an object, handing it over between two arms, and placing it in a desired target pose - is valuable when direct placement from the initial grasp is infeasible due to collisions, kinematic constraints, or poor final orientation. However, achieving this under multiple competing obje... | [
{
"start": 348,
"end": 533,
"label": "Contribution",
"text": "We introduce BiCompoDiff, a compositional diffusion and energy-based framework that jointly optimizes grasp selection, handover, regrasp, and motion planning under multiple constraints."
},
{
"start": 534,
"end": 831,
"lab... |
2607.21273v1 | The Dark Room in the Reward Channel: Dense Prediction Rewards Collapse GRPO-Trained LLM Agents -- and What Actually Works | Dense per-step supervision is an appealing remedy for sparse-reward, long-horizon LLM agents: reward the agent for predicting its next observation, and memory should follow. We show that under group-normalized RL (GRPO), this recipe does not merely fail -- it destroys the policy. Across Qwen3-1.7B/4B/8B on ALFWorld, a ... | [
{
"start": 174,
"end": 280,
"label": "Contribution",
"text": "We show that under group-normalized RL (GRPO), this recipe does not merely fail -- it destroys the policy."
},
{
"start": 556,
"end": 708,
"label": "Result",
"text": "A single-factor ablation localizes the cause -- rem... |
2607.22954v1 | Toward Automated Detection of Documentation Inconsistencies in Electronic Health Records | Objective: To characterize the kinds of internal documentation inconsistencies a general-domain large language model (LLM) can surface from real-world discharge summaries, and to identify recurring failure modes that limit reliability at scale. Materials and Methods: We applied a two-stage LLM pipeline---open-ended can... | [
{
"start": 11,
"end": 170,
"label": "Contribution",
"text": "To characterize the kinds of internal documentation inconsistencies a general-domain large language model (LLM) can surface from real-world discharge summaries"
},
{
"start": 268,
"end": 419,
"label": "Method",
"text": ... |
2607.21847v1 | Distributional Determinantal Point Process for Repulsive Clustering of Distributions | We introduce the distributional determinantal point process (dDPP) as a novel repulsive point process whose atoms are probability distributions rather than points in a real space. The dDPP is constructed via an L-ensemble with a sliced Wasserstein (SW) kernel between distributions. We show its validity as a well-define... | [
{
"start": 0,
"end": 179,
"label": "Contribution",
"text": "We introduce the distributional determinantal point process (dDPP) as a novel repulsive point process whose atoms are probability distributions rather than points in a real space."
},
{
"start": 180,
"end": 282,
"label": "Me... |
2607.23121v1 | SMART: LLM-Augmented Hybrid Retrieval for Dynamic Product Ads | Dynamic Product Ads (DPA) require retrieving relevant items from multi-million product catalogs, balancing two competing objectives: retargeting (re-surfacing known interests) and prospecting (discovering new categories). While Large Language Models (LLMs) capture semantic intent better than traditional embedding model... | [
{
"start": 415,
"end": 658,
"label": "Method",
"text": "Through controlled experiments on millions of users, we demonstrate a critical retrieval decomposition: rule-generated queries excel at retargeting on a lexical BM25 index, while LLM-generated queries excel at prospecting on a dense ANN index."... |
2607.16806v1 | Token-Wise Latent Streaming from Slow Reasoners to Fast Planners for Dynamic Vision Language Navigation | Vision-Language Navigation in dynamic, human-centric environments exposes a fundamental tension: linguistic reasoning is slow and deliberative, whereas safe, socially compliant planning should be instant and reactive. The resulting observation staleness is safety-critical: a maneuver chosen during inference can already... | [
{
"start": 485,
"end": 553,
"label": "Contribution",
"text": "We propose SPARK-VLN, a dual-system framework for dynamic social VLN"
},
{
"start": 559,
"end": 724,
"label": "Method",
"text": "streams the slow VLM reasoner's knowledge to a fast flow-matching expert planner througho... |
2607.22413v1 | Reflector: Arrangement-Aware Harmonic Retrieval for Sample-Based Composition | Sample retrieval tools can help composers find harmonically compatible material, but querying from a fixed reference sample becomes less informative as arrangements evolve and the harmonic context shifts with each musical decision. We present Reflector, an interactive audio workstation that tracks harmonic combinations... | [
{
"start": 232,
"end": 286,
"label": "Contribution",
"text": "We present Reflector, an interactive audio workstation"
},
{
"start": 417,
"end": 573,
"label": "Method",
"text": "The system is organized around a fixed interval-class oracle: a hand-designed table of weights that sco... |
2607.23295v1 | FILLER: Feature Imputation via Latent Location Exploration and Retrieval | In real-world machine learning applications, incomplete observations create a fundamental challenge. Researchers have come up with several ideas to address this crucial problem. However, current models still face challenges in balancing scalability and structural consistency. This study proposes a feature imputation me... | [
{
"start": 277,
"end": 339,
"label": "Contribution",
"text": "This study proposes a feature imputation method, called FILLER"
},
{
"start": 346,
"end": 486,
"label": "Method",
"text": "deliberately searches the two-dimensional latent space produced by a generative model and fills... |
2607.23024v1 | When Less Is More: A Controlled Benchmark of Lightweight CNNs for Satellite Land-Cover Segmentation on DeepGlobe | High-resolution satellite imagery is the backbone of good land-cover classification, and without that, environmental monitoring, urban planning, and sustainable resource management all fall short. Deep learning architectures perform well in semantic segmentation, but the efficiency-accuracy trade-off across classical c... | [
{
"start": 408,
"end": 665,
"label": "Contribution",
"text": "This study compares five architectures VGG16, MobileNetV2, InceptionV3, AlexNet, and CNN on the DeepGlobe Land Cover Classification dataset using three progressively optimized iterations to isolate regularisation, transfer learning, and a... |
2607.21820v1 | Probing Speaker Identity Sensitivity in Audio Deepfake Detectors | Audio deepfake detectors are trained to distinguish genuine speech from synthetic speech and often perform well on standard benchmarks. Yet the same detector that achieves less than 1% error on one dataset can see its error rate increase twentyfold when evaluated on a different dataset. We argue that one contributing f... | [
{
"start": 288,
"end": 542,
"label": "Contribution",
"text": "We argue that one contributing factor is speaker-identity reliance: standard training corpora correlate speaker identity with the genuine/synthetic label, allowing detectors to partially rely on speaker-related cues rather than synthesis ... |
2607.16472v1 | Spectral-Morphological Attention U-Net: An Efficient Network for Active Wildfire Detection | Over the past decades, the frequency of global wildfires has been increasing steadily. Therefore, if the fire can be detected and precisely located at an early stage, the potential hazards caused by it can be minimized to the greatest extent. The machine learning methods based on satellite images, due to their ability ... | [
{
"start": 484,
"end": 722,
"label": "Contribution",
"text": "we proposed a new model named spectral-morphological attention U-Net(SMA-UNet), which includes a spectral attention module, a residual attention UNet backbone, a channel-spatial modulator, and a pair of differentiable morphological gates.... |
2607.21263v1 | Filter Learning for Subgraphs: Algebras and Performance Risk Bounds | Graph signal processing tasks that leverage spectral information typically assume access to the complete graph topology, which is often unavailable in practice. We propose a systematic framework for subgraph filter learning (SFL), where subgraph-supported operators approximate ambient graph filters under partial observ... | [
{
"start": 161,
"end": 229,
"label": "Contribution",
"text": "We propose a systematic framework for subgraph filter learning (SFL)"
},
{
"start": 231,
"end": 326,
"label": "Method",
"text": "where subgraph-supported operators approximate ambient graph filters under partial observ... |
2607.22910v1 | Beyond Directed Acyclic Graphs: Causal Zeros and Causal Differential Equations | Pearl's structural causal model (SCM) framework, built on directed acyclic graphs (DAGs) and the do-calculus, is the dominant formal language for causal reasoning. Yet it carries two structural restrictions: every relationship must be pre-specified as a directed causal edge, and feedback cycles are forbidden. This pape... | [
{
"start": 0,
"end": 108,
"label": "Method",
"text": "Pearl's structural causal model (SCM) framework, built on directed acyclic graphs (DAGs) and the do-calculus"
},
{
"start": 164,
"end": 310,
"label": "Result",
"text": "Yet it carries two structural restrictions: every relatio... |
2607.22199v1 | From Score Approximation to Distribution Approximation in Score-Based Diffusion Models | Score-based diffusion models have achieved remarkable empirical success in generative modeling, yet their approximation-theoretic foundations remain incomplete. In particular, although classical universal approximation theorems guarantee that neural networks can approximate score functions, it remains unclear whether s... | [
{
"start": 468,
"end": 541,
"label": "Contribution",
"text": "we establish a rigorous quantitative connection between these two notions"
},
{
"start": 557,
"end": 969,
"label": "Contribution",
"text": "we prove that if a neural network approximates the true score function suffici... |
2607.21324v1 | GRADRAG: Cross-Component Prompt Adaptation for Coordinated Multi-Agent RAG | Retrieval-Augmented Generation (RAG) systems increasingly employ multiple LLM agents. Yet, most prior work optimizes components in isolation rather than coordinating improvements across the pipeline. We introduce GRADRAG, a framework for cross-component prompt adaptation that models the RAG pipeline as a computational ... | [
{
"start": 200,
"end": 271,
"label": "Contribution",
"text": "We introduce GRADRAG, a framework for cross-component prompt adaptation"
},
{
"start": 277,
"end": 398,
"label": "Method",
"text": "models the RAG pipeline as a computational graph and propagates structured evaluation ... |
2607.21279v1 | A Unified Moral-Value Dataset for Instruction Tuning | Large language models (LLMs) have developed rapidly and become valuable tools in everyday life. However, how to align LLMs to a particular set of human values is still an open problem. Recent studies show that instruction tuning has strong potential for zero-shot tasks and may serve as an effective approach to addressi... | [
{
"start": 488,
"end": 580,
"label": "Contribution",
"text": "We construct a unified moral-value dataset that can be directly used for instruction tuning."
},
{
"start": 581,
"end": 732,
"label": "Method",
"text": "This dataset is built upon existing moral-value datasets by mergi... |
2607.22859v1 | PatiGonit22K: A Comprehensive Dataset for Solving Complex Bengali MWPs | Mathematical Word Problems (MWPs) are an important benchmark for evaluating natural language understanding and quantitative reasoning. Despite recent progress in high resource languages, Bengali remains underexplored due to the limited availability of large scale annotated datasets. In this work, we introduce PatiGonit... | [
{
"start": 298,
"end": 323,
"label": "Contribution",
"text": "we introduce PatiGonit22K"
},
{
"start": 385,
"end": 511,
"label": "Method",
"text": "developed by extending the original PatiGonit dataset with a substantially larger collection of complex mathematical problems."
},... |
2607.22999v1 | WCM: World-Cognition Model for Generalizable Human-Robot Interaction | Language agents can now interact fluently with users in software, but robots still struggle to bring comparable interaction to physical tasks. Current robot-control paradigms, including vision-language-action policies and world-model-based planners, are mainly optimized for instruction execution, leaving users with lit... | [
{
"start": 468,
"end": 574,
"label": "Contribution",
"text": "we present the World-Cognition Model (WCM), a human-centered embodied agent built on the SLAK architecture"
},
{
"start": 644,
"end": 787,
"label": "Method",
"text": "SLAK separates perception, reasoning, control, and ... |
2607.21063v1 | QuantiBias: Benchmarking Quantization-Induced Bias in LLMs | Almost every large language model that reaches a broad audience is quantized: trained in full precision, then compressed for efficiency. This step is assumed harmless and its safety is rarely re-checked. We find its principal side effect is increased bias that standard safety evaluation misses. Holding the model, its t... | [
{
"start": 204,
"end": 295,
"label": "Contribution",
"text": "We find its principal side effect is increased bias that standard safety evaluation misses."
},
{
"start": 500,
"end": 798,
"label": "Result",
"text": "Yet asked an open-ended question, the same model volunteers stereo... |
2607.21574v1 | Surprisal Theory is Tautological (without Rational Grounding) | Surprisal theory holds that the human processing difficulty of a linguistic unit in context is an affine function of its surprisal under some language model. I argue this claim is a tautology without further constraint: for any non-negative difficulty measure over units in context, there exists a language model whose s... | [
{
"start": 158,
"end": 218,
"label": "Contribution",
"text": "I argue this claim is a tautology without further constraint"
},
{
"start": 283,
"end": 389,
"label": "Contribution",
"text": "there exists a language model whose surprisal is an affine function of it under mild techni... |
2607.23258v1 | CAPT: A Multi-task Continuous Autoregressive Transformer enabling Cross-dataset and Cross-species Transfer for Calcium Population Dynamics | Large-scale calcium imaging has created an opportunity to build foundation-style models for neural population dynamics, but a central question remains unresolved: \textbf{whether a model pretrained on one collection of recordings can generalize to new datasets, experimental paradigms, and even species.} Existing approa... | [
{
"start": 665,
"end": 869,
"label": "Method",
"text": "CAPT models continuous calcium traces directly through a continuous patch tokenization strategy and is trained autoregressively, enabling end-to-end pretraining and adaptation to diverse downstream tasks."
},
{
"start": 1115,
"end":... |
2607.22513v1 | Opaque Epistemic Mediation: How LLM Deployment Configurations Shape the Validation of Pseudo-Science | Commercial large language models are increasingly used as knowledge references, yet their stance on contested scientific claims is neither stable nor transparent. We tested how four major LLM families (Claude, Grok, GPT, Gemini) evaluate ethnonationalist pseudo-science derived from Frank Salter's biosocial framework ac... | [
{
"start": 163,
"end": 348,
"label": "Contribution",
"text": "We tested how four major LLM families (Claude, Grok, GPT, Gemini) evaluate ethnonationalist pseudo-science derived from Frank Salter's biosocial framework across four temporal snapshots"
},
{
"start": 379,
"end": 410,
"lab... |
2607.21424v1 | An Evaluation Framework for Structured Audio Captions Validated by Controlled Perturbations | Recent advancements in automated audio captioning (AAC) have shifted from monolithic sentence generation toward structured formats that explicitly disentangle distinct acoustic and semantic properties. However, evaluating this heterogeneous data remains a significant challenge. Existing caption metrics focus on flat te... | [
{
"start": 405,
"end": 492,
"label": "Contribution",
"text": "we propose a multi-axis evaluation framework tailored for structured audio descriptions"
},
{
"start": 494,
"end": 666,
"label": "Method",
"text": "Building on the AudioCards dataset, we evaluate outputs across five or... |
2607.22884v1 | CHiPS: Character Histograms and Positional Signals for Lightweight Authorship Attribution in Romanian Texts | We propose CHiPS, a lightweight character-level authorship attribution method for Romanian texts. All reported experiments are closed-set: the true author is one of the candidate authors in the training data. CHiPS studies two complementary fingerprints of writing style: CH-SVM, a character-histogram classifier based o... | [
{
"start": 0,
"end": 97,
"label": "Contribution",
"text": "We propose CHiPS, a lightweight character-level authorship attribution method for Romanian texts."
},
{
"start": 272,
"end": 358,
"label": "Method",
"text": "CH-SVM, a character-histogram classifier based on one-character... |
2607.11881v1 | Metacognition in LLMs: Foundations, Progress, and Opportunities | Metacognition is a foundational component of intelligence critical to effective learning, problem solving, decision-making, communication, and more. In recent years, it has become increasingly recognized as a cornerstone of capable, transparent AI systems. Yet while LLMs have made significant progress across diverse re... | [
{
"start": 580,
"end": 715,
"label": "Contribution",
"text": "This paper bridges this gap by presenting the first comprehensive overview of the current state of knowledge on metacognition for LLMs."
},
{
"start": 716,
"end": 974,
"label": "Method",
"text": "We analyze and taxonom... |
2607.22067v1 | Benchmarking Fine-tuning and Retrieval Strategies for a Multimodal Language Model on the NRC Reactor Operator Licensing Examination | The integration of large language models (LLMs) into the nuclear power industry requires outputs grounded in domain-specific knowledge. This study evaluates a 31-billion-parameter open-weight multimodal model (Gemma 4 31B-IT) on its capacity to apply nuclear knowledge by benchmarking eight model-retrieval configuration... | [
{
"start": 136,
"end": 268,
"label": "Contribution",
"text": "This study evaluates a 31-billion-parameter open-weight multimodal model (Gemma 4 31B-IT) on its capacity to apply nuclear knowledge"
},
{
"start": 415,
"end": 609,
"label": "Method",
"text": "We evaluate 14 Generic Fu... |
2607.20872v1 | LegalCiteTrust: Benchmarking Citation Trustworthiness in Chinese Long-Form Legal Research Reports | Long-form legal research reports increasingly rely on LLMs and agentic research systems, but their reliability depends not only on answering the task, but also on whether cited legal authorities are trustworthy. A citation can be risky even when it points to a real source: the report may omit limiting conditions, misde... | [
{
"start": 404,
"end": 529,
"label": "Contribution",
"text": "We introduce LegalCiteTrust, a benchmark for evaluating citation trustworthiness in Chinese long-form legal research reports."
},
{
"start": 530,
"end": 791,
"label": "Method",
"text": "It contains 72 densely annotated... |
2607.23009v1 | Recycling computational processes of dynamic programming for combinatorial optimization problems: a reservoir computing approach | Reusing previously computed results is a long-standing principle for reducing computational cost, but such reuse has largely been confined to a single problem's computation. Sharing computational processes across multiple simultaneously solved problems remains possible in principle, yet designing algorithms that exploi... | [
{
"start": 393,
"end": 458,
"label": "Contribution",
"text": "we use machine learning to discover such algorithms automatically"
},
{
"start": 474,
"end": 502,
"label": "Method",
"text": "based on reservoir computing"
},
{
"start": 504,
"end": 732,
"label": "Contr... |
2607.21043v1 | A Real-Time Generalized Nash Equilibrium Framework for Interaction-Aware Autonomous Driving in Mixed Traffic | Safe and efficient navigation in mixed-traffic environments remains a critical challenge for Autonomous Vehicles (AVs), primarily due to the complex interdependence between the AV's decisions and the unpredictable reactions of human drivers. This paper introduces a comprehensive decision-making framework that formulate... | [
{
"start": 242,
"end": 395,
"label": "Contribution",
"text": "This paper introduces a comprehensive decision-making framework that formulates the driving interaction as a Generalized Nash Equilibrium Problem (GNEP)."
},
{
"start": 396,
"end": 611,
"label": "Method",
"text": "Unli... |
End of preview. Expand in Data Studio
Semantic Highlighting: Rhetorical Roles in arXiv Abstracts
Rhetorical-role span annotations for ML/NLP arXiv abstracts, built for a
semantic highlighting accessibility tool (dyslexia/ADHD reading support) --
see the project repo
for the full research writeup (docs/findings.md).
Labels
Each span is tagged with one of 7 roles: Contribution, Method,
Result, Evidence, Limitation, FutureWork, Safety. Full
definitions, cue phrases, and boundary rules in
docs/annotation_guideline.md.
Schema
{
"arxiv_id": str,
"title": str,
"text": str, # the abstract
"spans": [
{"start": int, "end": int, "label": str, "text": str},
...
]
}
start/end are character offsets into text. Spans within a record do
not overlap.
Splits
- train: silver-labeled abstracts (LLM weak-labeled, see Methodology),
excluding every abstract that appears in
test. - validation: a small held-out slice of the same silver pool, used only to watch for overfitting during training -- also silver-labeled, so treat it as noisy, not as a second gold set.
- test: hand-corrected gold examples. The only split with human review.
Methodology
- Abstracts fetched from the arXiv API (
cat:cs.LG,cat:cs.CL, pluscat:cs.RO AND abs:safetyandabs:"future work"-biased queries added specifically because Safety/FutureWork are rare under random sampling). - Weak-labeled by a local LLM (Ollama,
ministral-3:latest) prompted with the annotation guideline, constrained to valid JSON matching the label schema. Every proposed span is checked against the source text character-for-character; anything the model didn't quote verbatim is dropped rather than guessed at (~9% drop rate). - A subset hand-corrected against the guideline to build
test. Systematic error patterns found during correction (Evidence tagged without any number, FutureWork tagged on plain findings with no forward-looking language, overlapping/duplicate spans) were then propagated across the entire silver corpus via automated rules inscripts/clean_silver.py-- not just the hand-corrected subset.
Full detail, including exact error rates and what didn't work, in the project's findings log.
Known limitations
- Silver labels are LLM-generated and imperfect even after automated
cleaning; only
testhas human review. - Skewed label distribution: Method/Result/Contribution/Evidence are common, Safety/FutureWork/Limitation are rare even after deliberately biasing collection toward them.
testis small (21 abstracts) -- any model evaluated against it should be treated as a directional signal, not a precise number.
Licensing
Abstract text is sourced from arXiv under arXiv's API Terms of Use; individual papers retain their own licenses (arXiv itself does not claim copyright over submitted abstracts). The span annotations (labels, offsets) added by this project are original work.
- Downloads last month
- 18