title
stringlengths 7
246
| abstract
stringlengths 3
3.31k
|
|---|---|
Iterative Semi-parametric Dynamics Model Learning For Autonomous Racing | Accurately modeling robot dynamics is crucial to safe and efficient motion control. In this paper, we develop and apply an iterative learning semi-parametric model, with a neural network, to the task of autonomous racing with a Model Predictive Controller (MPC). We present a novel non-linear semi-parametric dynamics model where we represent the known dynamics with a parametric model, and a neural network captures the unknown dynamics. We show that our model can learn more accurately than a purely parametric model and generalize better than a purely non-parametric model, making it ideal for real-world applications where collecting data from the full state space is not feasible. We present a system where the model is bootstrapped on pre-recorded data and then updated iteratively at run time. Then we apply our iterative learning approach to the simulated problem of autonomous racing and show that it can safely adapt to modified dynamics online and even achieve better performance than models trained on data from manual driving. |
What can linearized neural networks actually say about generalization? | For certain infinitely-wide neural networks, the neural tangent kernel (NTK) theory fully characterizes generalization, but for the networks used in practice, the empirical NTK only provides a rough first-order approximation. Still, a growing body of work keeps leveraging this approximation to successfully analyze important deep learning phenomena and design algorithms for new applications. In our work, we provide strong empirical evidence to determine the practical validity of such approximation by conducting a systematic comparison of the behavior of different neural networks and their linear approximations on different tasks. We show that the linear approximations can indeed rank the learning complexity of certain tasks for neural networks, even when they achieve very different performances. However, in contrast to what was previously reported, we discover that neural networks do not always perform better than their kernel approximations, and reveal that the performance gap heavily depends on architecture, dataset size and training task. We discover that networks overfit to these tasks mostly due to the evolution of their kernel during training, thus, revealing a new type of implicit bias. |
Automated Detection of Abnormalities from an EEG Recording of Epilepsy Patients With a Compact Convolutional Neural Network | Electroencephalography (EEG) is essential for the diagnosis of epilepsy, but it requires expertise and experience to identify abnormalities. It is thus crucial to develop automated models for the detection of abnormalities in EEGs related to epilepsy. This paper describes the development of a novel class of compact convolutional neural networks (CNNs) for detecting abnormal patterns and electrodes in EEGs for epilepsy. The designed model is inspired by a CNN developed for brain-computer interfacing called multichannel EEGNet (mEEGNet). Unlike the EEGNet, the proposed model, mEEGNet, has the same number of electrode inputs and outputs to detect abnormal patterns. The mEEGNet was evaluated with a clinical dataset consisting of 29 cases of juvenile and childhood absence epilepsy labeled by a clinical expert. The labels were given to paroxysmal discharges visually observed in both ictal (seizure) and interictal (nonseizure) durations. Results showed that the mEEGNet detected abnormalities with the area under the curve, F1-values, and sensitivity equivalent to or higher than those of existing CNNs. Moreover, the number of parameters is much smaller than other CNN models. To our knowledge, the dataset of absence epilepsy validated with machine learning through this research is the largest in the literature. |
PrivFair: a Library for Privacy-Preserving Fairness Auditing | Machine learning (ML) has become prominent in applications that directly affect people's quality of life, including in healthcare, justice, and finance. ML models have been found to exhibit discrimination based on sensitive attributes such as gender, race, or disability. Assessing if an ML model is free of bias remains challenging to date, and by definition has to be done with sensitive user characteristics that are subject of anti-discrimination and data protection law. Existing libraries for fairness auditing of ML models offer no mechanism to protect the privacy of the audit data. We present PrivFair, a library for privacy-preserving fairness audits of ML models. Through the use of Secure Multiparty Computation (MPC), PrivFair protects the confidentiality of the model under audit and the sensitive data used for the audit, hence it supports scenarios in which a proprietary classifier owned by a company is audited using sensitive audit data from an external investigator. We demonstrate the use of PrivFair for group fairness auditing with tabular data or image data, without requiring the investigator to disclose their data to anyone in an unencrypted manner, or the model owner to reveal their model parameters to anyone in plaintext. |
A hyperparameter-tuning approach to automated inverse planning | Radiotherapy inverse planning often requires planners to modify parameters in the treatment planning system's objective function to produce clinically acceptable plans. Due to the manual steps in this process, plan quality can vary depending on the planning time available and the planner's skills. This study investigates two hyperparameter-tuning methods for automated inverse planning. Because this framework does not train a model on previously-optimized plans, it can be readily adapted to practice pattern changes, and plan quality is not limited by that of a training cohort. We selected 10 patients who received lung SBRT using manually-generated clinical plans. We used random sampling (RS) and Bayesian optimization (BO) to tune parameters using linear-quadratic utility functions based on 11 clinical goals. Normalizing all plans to have PTV D95 equal to 48 Gy, we compared plan quality for the automatically-generated and manually-generated plans. We also investigated the impact of iteration count on the automatically-generated plans, comparing planning time and plan utility for RS and BO plans with and without stopping criteria. Without stopping criteria, the median planning time was 1.9 and 2.3 hours for RS and BO plans. The OAR doses in the RS and BO plans had a median percent difference (MPD) of 48.7% and 60.4% below clinical dose limits and an MPD of 2.8% and 3.3% below clinical plan doses. With stopping criteria, the utility decreased by an MPD of 5.3% and 3.9% for RS and BO plans, but the median planning time was reduced to 0.5 and 0.7 hours, and the OAR doses still had an MPD of 42.9% and 49.7% below clinical dose limits and an MPD of 0.3% and 1.8% below clinical plan doses. This study demonstrates that hyperparameter-tuning approaches to automated inverse planning can reduce active planning time with plan quality that is similar to or better than manually-generated plans. |
Show me the Way: Intrinsic Motivation from Demonstrations | The study of exploration in the domain of decision making has a long history but remains actively debated. From the vast literature that addressed this topic for decades under various points of view (e.g., developmental psychology, experimental design, artificial intelligence), intrinsic motivation emerged as a concept that can practically be transferred to artificial agents. Especially, in the recent field of Deep Reinforcement Learning (RL), agents implement such a concept (mainly using a novelty argument) in the shape of an exploration bonus, added to the task reward, that encourages visiting the whole environment. This approach is supported by the large amount of theory on RL for which convergence to optimality assumes exhaustive exploration. Yet, Human Beings and mammals do not exhaustively explore the world and their motivation is not only based on novelty but also on various other factors (e.g., curiosity, fun, style, pleasure, safety, competition, etc.). They optimize for life-long learning and train to learn transferable skills in playgrounds without obvious goals. They also apply innate or learned priors to save time and stay safe. For these reasons, we propose to learn an exploration bonus from demonstrations that could transfer these motivations to an artificial agent with little assumptions about their rationale. Using an inverse RL approach, we show that complex exploration behaviors, reflecting different motivations, can be learnt and efficiently used by RL agents to solve tasks for which exhaustive exploration is prohibitive. |
Guided Exploration of Data Summaries | Data summarization is the process of producing interpretable and representative subsets of an input dataset. It is usually performed following a one-shot process with the purpose of finding the best summary. A useful summary contains k individually uniform sets that are collectively diverse to be representative. Uniformity addresses interpretability and diversity addresses representativity. Finding such as summary is a difficult task when data is highly diverse and large. We examine the applicability of Exploratory Data Analysis (EDA) to data summarization and formalize Eda4Sum, the problem of guided exploration of data summaries that seeks to sequentially produce connected summaries with the goal of maximizing their cumulative utility. EdA4Sum generalizes one-shot summarization. We propose to solve it with one of two approaches: (i) Top1Sum which chooses the most useful summary at each step; (ii) RLSum which trains a policy with Deep Reinforcement Learning that rewards an agent for finding a diverse and new collection of uniform sets at each step. We compare these approaches with one-shot summarization and top-performing EDA solutions. We run extensive experiments on three large datasets. Our results demonstrate the superiority of our approaches for summarizing very large data, and the need to provide guidance to domain experts. |
An Edge Information and Mask Shrinking Based Image Inpainting Approach | In the image inpainting task, the ability to repair both high-frequency and low-frequency information in the missing regions has a substantial influence on the quality of the restored image. However, existing inpainting methods usually fail to consider both high-frequency and low-frequency information simultaneously. To solve this problem, this paper proposes edge information and mask shrinking based image inpainting approach, which consists of two models. The first model is an edge generation model used to generate complete edge information from the damaged image, and the second model is an image completion model used to fix the missing regions with the generated edge information and the valid contents of the damaged image. The mask shrinking strategy is employed in the image completion model to track the areas to be repaired. The proposed approach is evaluated qualitatively and quantitatively on the dataset Places2. The result shows our approach outperforms state-of-the-art methods. |
Exclusion and Inclusion -- A model agnostic approach to feature importance in DNNs | Deep Neural Networks in NLP have enabled systems to learn complex non-linear relationships. One of the major bottlenecks towards being able to use DNNs for real world applications is their characterization as black boxes. To solve this problem, we introduce a model agnostic algorithm which calculates phrase-wise importance of input features. We contend that our method is generalizable to a diverse set of tasks, by carrying out experiments for both Regression and Classification. We also observe that our approach is robust to outliers, implying that it only captures the essential aspects of the input. |
Re-Examining Linear Embeddings for High-Dimensional Bayesian Optimization | Bayesian optimization (BO) is a popular approach to optimize expensive-to-evaluate black-box functions. A significant challenge in BO is to scale to high-dimensional parameter spaces while retaining sample efficiency. A solution considered in existing literature is to embed the high-dimensional space in a lower-dimensional manifold, often via a random linear embedding. In this paper, we identify several crucial issues and misconceptions about the use of linear embeddings for BO. We study the properties of linear embeddings from the literature and show that some of the design choices in current approaches adversely impact their performance. We show empirically that properly addressing these issues significantly improves the efficacy of linear embeddings for BO on a range of problems, including learning a gait policy for robot locomotion. |
Predicting Future Lane Changes of Other Highway Vehicles using RNN-based Deep Models | In the event of sensor failure, autonomous vehicles need to safely execute emergency maneuvers while avoiding other vehicles on the road. To accomplish this, the sensor-failed vehicle must predict the future semantic behaviors of other drivers, such as lane changes, as well as their future trajectories given a recent window of past sensor observations. We address the first issue of semantic behavior prediction in this paper, which is a precursor to trajectory prediction, by introducing a framework that leverages the power of recurrent neural networks (RNNs) and graphical models. Our goal is to predict the future categorical driving intent, for lane changes, of neighboring vehicles up to three seconds into the future given as little as a one-second window of past LIDAR, GPS, inertial, and map data. We collect real-world data containing over 20 hours of highway driving using an autonomous Toyota vehicle. We propose a composite RNN model by adopting the methodology of Structural Recurrent Neural Networks (RNNs) to learn factor functions and take advantage of both the high-level structure of graphical models and the sequence modeling power of RNNs, which we expect to afford more transparent modeling and activity than opaque, single RNN models. To demonstrate our approach, we validate our model using authentic interstate highway driving to predict the future lane change maneuvers of other vehicles neighboring our autonomous vehicle. We find that our composite Structural RNN outperforms baselines by as much as 12% in balanced accuracy metrics. |
Value of structural health information in partially observable stochastic environments | Efficient integration of uncertain observations with decision-making optimization is key for prescribing informed intervention actions, able to preserve structural safety of deteriorating engineering systems. To this end, it is necessary that scheduling of inspection and monitoring strategies be objectively performed on the basis of their expected value-based gains that, among others, reflect quantitative metrics such as the Value of Information (VoI) and the Value of Structural Health Monitoring (VoSHM). In this work, we introduce and study the theoretical and computational foundations of the above metrics within the context of Partially Observable Markov Decision Processes (POMDPs), thus alluding to a broad class of decision-making problems of partially observable stochastic deteriorating environments that can be modeled as POMDPs. Step-wise and life-cycle VoI and VoSHM definitions are devised and their bounds are analyzed as per the properties stemming from the Bellman equation and the resulting optimal value function. It is shown that a POMDP policy inherently leverages the notion of VoI to guide observational actions in an optimal way at every decision step, and that the permanent or intermittent information provided by SHM or inspection visits, respectively, can only improve the cost of this policy in the long-term, something that is not necessarily true under locally optimal policies, typically adopted in decision-making of structures and infrastructure. POMDP solutions are derived based on point-based value iteration methods, and the various definitions are quantified in stationary and non-stationary deteriorating environments, with both infinite and finite planning horizons, featuring single- or multi-component engineering systems. |
Using LSTMs for climate change assessment studies on droughts and floods | Climate change affects occurrences of floods and droughts worldwide. However, predicting climate impacts over individual watersheds is difficult, primarily because accurate hydrological forecasts require models that are calibrated to past data. In this work we present a large-scale LSTM-based modeling approach that -- by training on large data sets -- learns a diversity of hydrological behaviors. Previous work shows that this model is more accurate than current state-of-the-art models, even when the LSTM-based approach operates out-of-sample and the latter in-sample. In this work, we show how this model can assess the sensitivity of the underlying systems with regard to extreme (high and low) flows in individual watersheds over the continental US. |
Transient growth of accelerated optimization algorithms | Optimization algorithms are increasingly being used in applications with limited time budgets. In many real-time and embedded scenarios, only a few iterations can be performed and traditional convergence metrics cannot be used to evaluate performance in these non-asymptotic regimes. In this paper, we examine the transient behavior of accelerated first-order optimization algorithms. For convex quadratic problems, we employ tools from linear systems theory to show that transient growth arises from the presence of non-normal dynamics. We identify the existence of modes that yield an algebraic growth in early iterations and quantify the transient excursion from the optimal solution caused by these modes. For strongly convex smooth optimization problems, we utilize the theory of integral quadratic constraints (IQCs) to establish an upper bound on the magnitude of the transient response of Nesterov's accelerated algorithm. We show that both the Euclidean distance between the optimization variable and the global minimizer and the rise time to the transient peak are proportional to the square root of the condition number of the problem. Finally, for problems with large condition numbers, we demonstrate tightness of the bounds that we derive up to constant factors. |
DeepFake Detection Based on the Discrepancy Between the Face and its Context | We propose a method for detecting face swapping and other identity manipulations in single images. Face swapping methods, such as DeepFake, manipulate the face region, aiming to adjust the face to the appearance of its context, while leaving the context unchanged. We show that this modus operandi produces discrepancies between the two regions. These discrepancies offer exploitable telltale signs of manipulation. Our approach involves two networks: (i) a face identification network that considers the face region bounded by a tight semantic segmentation, and (ii) a context recognition network that considers the face context (e.g., hair, ears, neck). We describe a method which uses the recognition signals from our two networks to detect such discrepancies, providing a complementary detection signal that improves conventional real vs. fake classifiers commonly used for detecting fake images. Our method achieves state of the art results on the FaceForensics++, Celeb-DF-v2, and DFDC benchmarks for face manipulation detection, and even generalizes to detect fakes produced by unseen methods. |
Towards a New Participatory Approach for Designing Artificial Intelligence and Data-Driven Technologies | With there being many technical and ethical issues with artificial intelligence (AI) that involve marginalized communities, there is a growing interest for design methods used with marginalized people that may be transferable to the design of AI technologies. Participatory design (PD) is a design method that is often used with marginalized communities for the design of social development, policy, IT and other matters and solutions. However, there are issues with the current PD, raising concerns when it is applied to the design of technologies, including AI technologies. This paper argues for the use of PD for the design of AI technologies, and introduces and proposes a new PD, which we call agile participatory design, that not only can could be used for the design of AI and data-driven technologies, but also overcomes issues surrounding current PD and its use in the design of such technologies. |
Reproducing BowNet: Learning Representations by Predicting Bags of Visual Words | This work aims to reproduce results from the CVPR 2020 paper by Gidaris et al. Self-supervised learning (SSL) is used to learn feature representations of an image using an unlabeled dataset. This work proposes to use bag-of-words (BoW) deep feature descriptors as a self-supervised learning target to learn robust, deep representations. BowNet is trained to reconstruct the histogram of visual words (ie. the deep BoW descriptor) of a reference image when presented a perturbed version of the image as input. Thus, this method aims to learn perturbation-invariant and context-aware image features that can be useful for few-shot tasks or supervised downstream tasks. In the paper, the author describes BowNet as a network consisting of a convolutional feature extractor $\Phi(\cdot)$ and a Dense-softmax layer $\Omega(\cdot)$ trained to predict BoW features from images. After BoW training, the features of $\Phi$ are used in downstream tasks. For this challenge we were trying to build and train a network that could reproduce the CIFAR-100 accuracy improvements reported in the original paper. However, we were unsuccessful in reproducing an accuracy improvement comparable to what the authors mentioned. This could be for a variety of factors and we believe that time constraints were the primary bottleneck. |
Semantic Video Segmentation for Intracytoplasmic Sperm Injection Procedures | We present the first deep learning model for the analysis of intracytoplasmic sperm injection (ICSI) procedures. Using a dataset of ICSI procedure videos, we train a deep neural network to segment key objects in the videos achieving a mean IoU of 0.962, and to localize the needle tip achieving a mean pixel error of 3.793 pixels at 14 FPS on a single GPU. We further analyze the variation between the dataset's human annotators and find the model's performance to be comparable to human experts. |
Stateful Offline Contextual Policy Evaluation and Learning | We study off-policy evaluation and learning from sequential data in a structured class of Markov decision processes that arise from repeated interactions with an exogenous sequence of arrivals with contexts, which generate unknown individual-level responses to agent actions. This model can be thought of as an offline generalization of contextual bandits with resource constraints. We formalize the relevant causal structure of problems such as dynamic personalized pricing and other operations management problems in the presence of potentially high-dimensional user types. The key insight is that an individual-level response is often not causally affected by the state variable and can therefore easily be generalized across timesteps and states. When this is true, we study implications for (doubly robust) off-policy evaluation and learning by instead leveraging single time-step evaluation, estimating the expectation over a single arrival via data from a population, for fitted-value iteration in a marginal MDP. We study sample complexity and analyze error amplification that leads to the persistence, rather than attenuation, of confounding error over time. In simulations of dynamic and capacitated pricing, we show improved out-of-sample policy performance in this class of relevant problems. |
"Did You Hear That?" Learning to Play Video Games from Audio Cues | Game-playing AI research has focused for a long time on learning to play video games from visual input or symbolic information. However, humans benefit from a wider array of sensors which we utilise in order to navigate the world around us. In particular, sounds and music are key to how many of us perceive the world and influence the decisions we make. In this paper, we present initial experiments on game-playing agents learning to play video games solely from audio cues. We expand the Video Game Description Language to allow for audio specification, and the General Video Game AI framework to provide new audio games and an API for learning agents to make use of audio observations. We analyse the games and the audio game design process, include initial results with simple Q~Learning agents, and encourage further research in this area. |
MetAL: Active Semi-Supervised Learning on Graphs via Meta Learning | The objective of active learning (AL) is to train classification models with less number of labeled instances by selecting only the most informative instances for labeling. The AL algorithms designed for other data types such as images and text do not perform well on graph-structured data. Although a few heuristics-based AL algorithms have been proposed for graphs, a principled approach is lacking. In this paper, we propose MetAL, an AL approach that selects unlabeled instances that directly improve the future performance of a classification model. For a semi-supervised learning problem, we formulate the AL task as a bilevel optimization problem. Based on recent work in meta-learning, we use the meta-gradients to approximate the impact of retraining the model with any unlabeled instance on the model performance. Using multiple graph datasets belonging to different domains, we demonstrate that MetAL efficiently outperforms existing state-of-the-art AL algorithms. |
Supernet Training for Federated Image Classification under System Heterogeneity | Efficient deployment of deep neural networks across many devices and resource constraints, especially on edge devices, is one of the most challenging problems in the presence of data-privacy preservation issues. Conventional approaches have evolved to either improve a single global model while keeping each local training data decentralized (i.e., data-heterogeneity) or to train a once-for-all network that supports diverse architectural settings to address heterogeneous systems equipped with different computational capabilities (i.e., model-heterogeneity). However, little research has considered both directions simultaneously. In this work, we propose a novel framework to consider both scenarios, namely Federation of Supernet Training (FedSup), where clients send and receive a supernet whereby it contains all possible architectures sampled from itself. It is inspired by how averaging parameters in the model aggregation stage of Federated Learning (FL) is similar to weight-sharing in supernet training. Specifically, in the FedSup framework, a weight-sharing approach widely used in the training single shot model is combined with the averaging of Federated Learning (FedAvg). Under our framework, we present an efficient algorithm (E-FedSup) by sending the sub-model to clients in the broadcast stage for reducing communication costs and training overhead. We demonstrate several strategies to enhance supernet training in the FL environment and conduct extensive empirical evaluations. The resulting framework is shown to pave the way for the robustness of both data- and model-heterogeneity on several standard benchmarks. |
DeepClean -- self-supervised artefact rejection for intensive care waveform data using deep generative learning | Waveform physiological data is important in the treatment of critically ill patients in the intensive care unit. Such recordings are susceptible to artefacts, which must be removed before the data can be re-used for alerting or reprocessed for other clinical or research purposes. Accurate removal of artefacts reduces bias and uncertainty in clinical assessment, as well as the false positive rate of intensive care unit alarms, and is therefore a key component in providing optimal clinical care. In this work, we present DeepClean; a prototype self-supervised artefact detection system using a convolutional variational autoencoder deep neural network that avoids costly and painstaking manual annotation, requiring only easily-obtained 'good' data for training. For a test case with invasive arterial blood pressure, we demonstrate that our algorithm can detect the presence of an artefact within a 10-second sample of data with sensitivity and specificity around 90%. Furthermore, DeepClean was able to identify regions of artefact within such samples with high accuracy and we show that it significantly outperforms a baseline principle component analysis approach in both signal reconstruction and artefact detection. DeepClean learns a generative model and therefore may also be used for imputation of missing data. |
Clustering Algorithms: A Comparative Approach | Many real-world systems can be studied in terms of pattern recognition tasks, so that proper use (and understanding) of machine learning methods in practical applications becomes essential. While a myriad of classification methods have been proposed, there is no consensus on which methods are more suitable for a given dataset. As a consequence, it is important to comprehensively compare methods in many possible scenarios. In this context, we performed a systematic comparison of 7 well-known clustering methods available in the R language. In order to account for the many possible variations of data, we considered artificial datasets with several tunable properties (number of classes, separation between classes, etc). In addition, we also evaluated the sensitivity of the clustering methods with regard to their parameters configuration. The results revealed that, when considering the default configurations of the adopted methods, the spectral approach usually outperformed the other clustering algorithms. We also found that the default configuration of the adopted implementations was not accurate. In these cases, a simple approach based on random selection of parameters values proved to be a good alternative to improve the performance. All in all, the reported approach provides subsidies guiding the choice of clustering algorithms. |
Identifying Planetary Transit Candidates in TESS Full-Frame Image Light Curves via Convolutional Neural Networks | The Transiting Exoplanet Survey Satellite (TESS) mission measured light from stars in ~75% of the sky throughout its two year primary mission, resulting in millions of TESS 30-minute cadence light curves to analyze in the search for transiting exoplanets. To search this vast data trove for transit signals, we aim to provide an approach that is both computationally efficient and produces highly performant predictions. This approach minimizes the required human search effort. We present a convolutional neural network, which we train to identify planetary transit signals and dismiss false positives. To make a prediction for a given light curve, our network requires no prior transit parameters identified using other methods. Our network performs inference on a TESS 30-minute cadence light curve in ~5ms on a single GPU, enabling large scale archival searches. We present 181 new planet candidates identified by our network, which pass subsequent human vetting designed to rule out false positives. Our neural network model is additionally provided as open-source code for public use and extension. |
FixEval: Execution-based Evaluation of Program Fixes for Competitive Programming Problems | Source code repositories consist of large codebases, often containing error-prone programs. The increasing complexity of software has led to a drastic rise in time and costs for identifying and fixing these defects. Various methods exist to automatically generate fixes for buggy code. However, due to the large combinatorial space of possible solutions for a particular bug, there are not many tools and datasets available to evaluate generated code effectively. In this work, we introduce FixEval, a benchmark comprising buggy code submissions to competitive programming problems and their respective fixes. We introduce a rich test suite to evaluate and assess the correctness of model-generated program fixes. We consider two Transformer language models pretrained on programming languages as our baselines, and compare them using match-based and execution-based evaluation metrics. Our experiments show that match-based metrics do not reflect model-generated program fixes accurately, while execution-based methods evaluate programs through all cases and scenarios specifically designed for that solution. Therefore, we believe FixEval provides a step towards real-world automatic bug fixing and model-generated code evaluation. |
CBCT-to-CT synthesis with a single neural network for head-and-neck, lung and breast cancer adaptive radiotherapy | Purpose: CBCT-based adaptive radiotherapy requires daily images for accurate dose calculations. This study investigates the feasibility of applying a single convolutional network to facilitate CBCT-to-CT synthesis for head-and-neck, lung, and breast cancer patients. Methods: Ninety-nine patients diagnosed with head-and-neck, lung or breast cancer undergoing radiotherapy with CBCT-based position verification were included in this study. CBCTs were registered to planning CTs according to clinical procedures. Three cycle-consistent generative adversarial networks (cycle-GANs) were trained in an unpaired manner on 15 patients per anatomical site generating synthetic-CTs (sCTs). Another network was trained with all the anatomical sites together. Performances of all four networks were compared and evaluated for image similarity against rescan CT (rCT). Clinical plans were recalculated on CT and sCT and analysed through voxel-based dose differences and {\gamma}-analysis. Results: A sCT was generated in 10 seconds. Image similarity was comparable between models trained on different anatomical sites and a single model for all sites. Mean dose differences < 0.5% were obtained in high-dose regions. Mean gamma (2%,2mm) pass-rates > 95% were achieved for all sites. Conclusions: Cycle-GAN reduced CBCT artefacts and increased HU similarity to CT, enabling sCT-based dose calculations. The speed of the network can facilitate on-line adaptive radiotherapy using a single network for head-and-neck, lung and breast cancer patients. |
Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations | The key idea behind the unsupervised learning of disentangled representations is that real-world data is generated by a few explanatory factors of variation which can be recovered by unsupervised learning algorithms. In this paper, we provide a sober look at recent progress in the field and challenge some common assumptions. We first theoretically show that the unsupervised learning of disentangled representations is fundamentally impossible without inductive biases on both the models and the data. Then, we train more than 12000 models covering most prominent methods and evaluation metrics in a reproducible large-scale experimental study on seven different data sets. We observe that while the different methods successfully enforce properties ``encouraged'' by the corresponding losses, well-disentangled models seemingly cannot be identified without supervision. Furthermore, increased disentanglement does not seem to lead to a decreased sample complexity of learning for downstream tasks. Our results suggest that future work on disentanglement learning should be explicit about the role of inductive biases and (implicit) supervision, investigate concrete benefits of enforcing disentanglement of the learned representations, and consider a reproducible experimental setup covering several data sets. |
Integrated Clustering and Anomaly Detection (INCAD) for Streaming Data (Revised) | Most current clustering based anomaly detection methods use scoring schema and thresholds to classify anomalies. These methods are often tailored to target specific data sets with "known" number of clusters. The paper provides a streaming clustering and anomaly detection algorithm that does not require strict arbitrary thresholds on the anomaly scores or knowledge of the number of clusters while performing probabilistic anomaly detection and clustering simultaneously. This ensures that the cluster formation is not impacted by the presence of anomalous data, thereby leading to more reliable definition of "normal vs abnormal" behavior. The motivations behind developing the INCAD model and the path that leads to the streaming model is discussed. |
Demonstration-Guided Deep Reinforcement Learning of Control Policies for Dexterous Human-Robot Interaction | In this paper, we propose a method for training control policies for human-robot interactions such as handshakes or hand claps via Deep Reinforcement Learning. The policy controls a humanoid Shadow Dexterous Hand, attached to a robot arm. We propose a parameterizable multi-objective reward function that allows learning of a variety of interactions without changing the reward structure. The parameters of the reward function are estimated directly from motion capture data of human-human interactions in order to produce policies that are perceived as being natural and human-like by observers. We evaluate our method on three significantly different hand interactions: handshake, hand clap and finger touch. We provide detailed analysis of the proposed reward function and the resulting policies and conduct a large-scale user study, indicating that our policy produces natural looking motions. |
Learning Causal State Representations of Partially Observable Environments | Intelligent agents can cope with sensory-rich environments by learning task-agnostic state abstractions. In this paper, we propose an algorithm to approximate causal states, which are the coarsest partition of the joint history of actions and observations in partially-observable Markov decision processes (POMDP). Our method learns approximate causal state representations from RNNs trained to predict subsequent observations given the history. We demonstrate that these learned state representations are useful for learning policies efficiently in reinforcement learning problems with rich observation spaces. We connect causal states with causal feature sets from the causal inference literature, and also provide theoretical guarantees on the optimality of the continuous version of this causal state representation under Lipschitz assumptions by proving equivalence to bisimulation, a relation between behaviorally equivalent systems. This allows for lower bounds on the optimal value function of the learned representation, which is tight given certain assumptions. Finally, we empirically evaluate causal state representations using multiple partially observable tasks and compare with prior methods. |
TopicNet: Semantic Graph-Guided Topic Discovery | Existing deep hierarchical topic models are able to extract semantically meaningful topics from a text corpus in an unsupervised manner and automatically organize them into a topic hierarchy. However, it is unclear how to incorporate prior beliefs such as knowledge graph to guide the learning of the topic hierarchy. To address this issue, we introduce TopicNet as a deep hierarchical topic model that can inject prior structural knowledge as an inductive bias to influence learning. TopicNet represents each topic as a Gaussian-distributed embedding vector, projects the topics of all layers into a shared embedding space, and explores both the symmetric and asymmetric similarities between Gaussian embedding vectors to incorporate prior semantic hierarchies. With an auto-encoding variational inference network, the model parameters are optimized by minimizing the evidence lower bound and a regularization term via stochastic gradient descent. Experiments on widely used benchmarks show that TopicNet outperforms related deep topic models on discovering deeper interpretable topics and mining better document~representations. |
Energy awareness in low precision neural networks | Power consumption is a major obstacle in the deployment of deep neural networks (DNNs) on end devices. Existing approaches for reducing power consumption rely on quite general principles, including avoidance of multiplication operations and aggressive quantization of weights and activations. However, these methods do not take into account the precise power consumed by each module in the network, and are therefore not optimal. In this paper we develop accurate power consumption models for all arithmetic operations in the DNN, under various working conditions. We reveal several important factors that have been overlooked to date. Based on our analysis, we present PANN (power-aware neural network), a simple approach for approximating any full-precision network by a low-power fixed-precision variant. Our method can be applied to a pre-trained network, and can also be used during training to achieve improved performance. In contrast to previous methods, PANN incurs only a minor degradation in accuracy w.r.t. the full-precision version of the network, even when working at the power-budget of a 2-bit quantized variant. In addition, our scheme enables to seamlessly traverse the power-accuracy trade-off at deployment time, which is a major advantage over existing quantization methods that are constrained to specific bit widths. |
Graph Neural Networks for Small Graph and Giant Network Representation Learning: An Overview | Graph neural networks denote a group of neural network models introduced for the representation learning tasks on graph data specifically. Graph neural networks have been demonstrated to be effective for capturing network structure information, and the learned representations can achieve the state-of-the-art performance on node and graph classification tasks. Besides the different application scenarios, the architectures of graph neural network models also depend on the studied graph types a lot. Graph data studied in research can be generally categorized into two main types, i.e., small graphs vs. giant networks, which differ from each other a lot in the size, instance number and label annotation. Several different types of graph neural network models have been introduced for learning the representations from such different types of graphs already. In this paper, for these two different types of graph data, we will introduce the graph neural networks introduced in recent years. To be more specific, the graph neural networks introduced in this paper include IsoNN, SDBN, LF&ER, GCN, GAT, DifNN, GNL, GraphSage and seGEN. Among these graph neural network models, IsoNN, SDBN and LF&ER are initially proposed for small graphs and the remaining ones are initially proposed for giant networks instead. The readers are also suggested to refer to these papers for detailed information when reading this tutorial paper. |
Fast Graph Representation Learning with PyTorch Geometric | We introduce PyTorch Geometric, a library for deep learning on irregularly structured input data such as graphs, point clouds and manifolds, built upon PyTorch. In addition to general graph data structures and processing methods, it contains a variety of recently published methods from the domains of relational learning and 3D data processing. PyTorch Geometric achieves high data throughput by leveraging sparse GPU acceleration, by providing dedicated CUDA kernels and by introducing efficient mini-batch handling for input examples of different size. In this work, we present the library in detail and perform a comprehensive comparative study of the implemented methods in homogeneous evaluation scenarios. |
Deep Learning on Real Geophysical Data: A Case Study for Distributed Acoustic Sensing Research | Deep Learning approaches for real, large, and complex scientific data sets can be very challenging to design. In this work, we present a complete search for a finely-tuned and efficiently scaled deep learning classifier to identify usable energy from seismic data acquired using Distributed Acoustic Sensing (DAS). While using only a subset of labeled images during training, we were able to identify suitable models that can be accurately generalized to unknown signal patterns. We show that by using 16 times more GPUs, we can increase the training speed by more than two orders of magnitude on a 50,000-image data set. |
Optimizing for Interpretability in Deep Neural Networks with Tree Regularization | Deep models have advanced prediction in many domains, but their lack of interpretability remains a key barrier to the adoption in many real world applications. There exists a large body of work aiming to help humans understand these black box functions to varying levels of granularity -- for example, through distillation, gradients, or adversarial examples. These methods however, all tackle interpretability as a separate process after training. In this work, we take a different approach and explicitly regularize deep models so that they are well-approximated by processes that humans can step-through in little time. Specifically, we train several families of deep neural networks to resemble compact, axis-aligned decision trees without significant compromises in accuracy. The resulting axis-aligned decision functions uniquely make tree regularized models easy for humans to interpret. Moreover, for situations in which a single, global tree is a poor estimator, we introduce a regional tree regularizer that encourages the deep model to resemble a compact, axis-aligned decision tree in predefined, human-interpretable contexts. Using intuitive toy examples as well as medical tasks for patients in critical care and with HIV, we demonstrate that this new family of tree regularizers yield models that are easier for humans to simulate than simpler L1 or L2 penalties without sacrificing predictive power. |
Learning Contextualised Cross-lingual Word Embeddings and Alignments for Extremely Low-Resource Languages Using Parallel Corpora | We propose a new approach for learning contextualised cross-lingual word embeddings based on a small parallel corpus (e.g. a few hundred sentence pairs). Our method obtains word embeddings via an LSTM encoder-decoder model that simultaneously translates and reconstructs an input sentence. Through sharing model parameters among different languages, our model jointly trains the word embeddings in a common cross-lingual space. We also propose to combine word and subword embeddings to make use of orthographic similarities across different languages. We base our experiments on real-world data from endangered languages, namely Yongning Na, Shipibo-Konibo, and Griko. Our experiments on bilingual lexicon induction and word alignment tasks show that our model outperforms existing methods by a large margin for most language pairs. These results demonstrate that, contrary to common belief, an encoder-decoder translation model is beneficial for learning cross-lingual representations even in extremely low-resource conditions. Furthermore, our model also works well on high-resource conditions, achieving state-of-the-art performance on a German-English word-alignment task. |
Invariant Representation Driven Neural Classifier for Anti-QCD Jet Tagging | We leverage representation learning and the inductive bias in neural-net-based Standard Model jet classification tasks, to detect non-QCD signal jets. In establishing the framework for classification-based anomaly detection in jet physics, we demonstrate that with a \emph{well-calibrated} and \emph{powerful enough feature extractor}, a well-trained \emph{mass-decorrelated} supervised Standard Model neural jet classifier can serve as a strong generic anti-QCD jet tagger for effectively reducing the QCD background. Imposing \emph{data-augmented} mass-invariance (decoupling the dominant factor) not only facilitates background estimation, but also induces more substructure-aware representation learning. We are able to reach excellent tagging efficiencies for all the test signals considered. In the best case, we reach a background rejection rate of 51 and a significance improvement factor of 3.6 at 50 \% signal acceptance, with jet mass decorrelated. This study indicates that supervised Standard Model jet classifiers have great potential in general new physics searches. |
Combining Spatial Clustering with LSTM Speech Models for Multichannel Speech Enhancement | Recurrent neural networks using the LSTM architecture can achieve significant single-channel noise reduction. It is not obvious, however, how to apply them to multi-channel inputs in a way that can generalize to new microphone configurations. In contrast, spatial clustering techniques can achieve such generalization, but lack a strong signal model. This paper combines the two approaches to attain both the spatial separation performance and generality of multichannel spatial clustering and the signal modeling performance of multiple parallel single-channel LSTM speech enhancers. The system is compared to several baselines on the CHiME3 dataset in terms of speech quality predicted by the PESQ algorithm and word error rate of a recognizer trained on mis-matched conditions, in order to focus on generalization. Our experiments show that by combining the LSTM models with the spatial clustering, we reduce word error rate by 4.6\% absolute (17.2\% relative) on the development set and 11.2\% absolute (25.5\% relative) on test set compared with spatial clustering system, and reduce by 10.75\% (32.72\% relative) on development set and 6.12\% absolute (15.76\% relative) on test data compared with LSTM model. |
Understanding and Mitigating Accuracy Disparity in Regression | With the widespread deployment of large-scale prediction systems in high-stakes domains, e.g., face recognition, criminal justice, etc., disparity in prediction accuracy between different demographic subgroups has called for fundamental understanding on the source of such disparity and algorithmic intervention to mitigate it. In this paper, we study the accuracy disparity problem in regression. To begin with, we first propose an error decomposition theorem, which decomposes the accuracy disparity into the distance between marginal label distributions and the distance between conditional representations, to help explain why such accuracy disparity appears in practice. Motivated by this error decomposition and the general idea of distribution alignment with statistical distances, we then propose an algorithm to reduce this disparity, and analyze its game-theoretic optima of the proposed objective functions. To corroborate our theoretical findings, we also conduct experiments on five benchmark datasets. The experimental results suggest that our proposed algorithms can effectively mitigate accuracy disparity while maintaining the predictive power of the regression models. |
Optimizing the optimizer for data driven deep neural networks and physics informed neural networks | We investigate the role of the optimizer in determining the quality of the model fit for neural networks with a small to medium number of parameters. We study the performance of Adam, an algorithm for first-order gradient-based optimization that uses adaptive momentum, the Levenberg and Marquardt (LM) algorithm a second order method, Broyden,Fletcher,Goldfarb and Shanno algorithm (BFGS) a second order method and LBFGS, a low memory version of BFGS. Using these optimizers we fit the function y = sinc(10x) using a neural network with a few parameters. This function has a variable amplitude and a constant frequency. We observe that the higher amplitude components of the function are fitted first and the Adam, BFGS and LBFGS struggle to fit the lower amplitude components of the function. We also solve the Burgers equation using a physics informed neural network(PINN) with the BFGS and LM optimizers. For our example problems with a small to medium number of weights, we find that the LM algorithm is able to rapidly converge to machine precision offering significant benefits over other optimizers. We further investigated the Adam optimizer with a range of models and found that Adam optimiser requires much deeper models with large numbers of hidden units containing up to 26x more parameters, in order to achieve a model fit close that achieved by the LM optimizer. The LM optimizer results illustrate that it may be possible build models with far fewer parameters. We have implemented all our methods in Keras and TensorFlow 2. |
Indeterminacy in Latent Variable Models: Characterization and Strong Identifiability | Most modern latent variable and probabilistic generative models, such as the variational autoencoder (VAE), have certain indeterminacies that are unresolvable even with an infinite amount of data. Recent applications of such models have indicated the need for strongly identifiable models, in which an observation corresponds to a unique latent code. Progress has been made towards reducing model indeterminacies while maintaining flexibility, most notably by the iVAE (arXiv:1907.04809 [stat.ML]), which excludes many -- but not all -- indeterminacies. We construct a full theoretical framework for analyzing the indeterminacies of latent variable models, and characterize them precisely in terms of properties of the generator functions and the latent variable prior distributions. To illustrate, we apply the framework to better understand the structure of recent identifiability results. We then investigate how we might specify strongly identifiable latent variable models, and construct two such classes of models. One is a straightforward modification of iVAE; the other uses ideas from optimal transport and leads to novel models and connections to recent work. |
A Comparative Study of Neural Network Compression | There has recently been an increasing desire to evaluate neural networks locally on computationally-limited devices in order to exploit their recent effectiveness for several applications; such effectiveness has nevertheless come together with a considerable increase in the size of modern neural networks, which constitute a major downside in several of the aforementioned computationally-limited settings. There has thus been a demand of compression techniques for neural networks. Several proposal in this direction have been made, which famously include hashing-based methods and pruning-based ones. However, the evaluation of the efficacy of these techniques has so far been heterogeneous, with no clear evidence in favor of any of them over the others. The goal of this work is to address this latter issue by providing a comparative study. While most previous studies test the capability of a technique in reducing the number of parameters of state-of-the-art networks , we follow [CWT + 15] in evaluating their performance on basic ar-chitectures on the MNIST dataset and variants of it, which allows for a clearer analysis of some aspects of their behavior. To the best of our knowledge, we are the first to directly compare famous approaches such as HashedNet, Optimal Brain Damage (OBD), and magnitude-based pruning with L1 and L2 regularization among them and against equivalent-size feed-forward neural networks with simple (fully-connected) and structural (convolutional) neural networks. Rather surprisingly, our experiments show that (iterative) pruning-based methods are substantially better than the HashedNet architecture, whose compression doesn't appear advantageous to a carefully chosen convolutional network. We also show that, when the compression level is high, the famous OBD pruning heuristics deteriorates to the point of being less efficient than simple magnitude-based techniques. |
Affinity-Aware Graph Networks | Graph Neural Networks (GNNs) have emerged as a powerful technique for learning on relational data. Owing to the relatively limited number of message passing steps they perform -- and hence a smaller receptive field -- there has been significant interest in improving their expressivity by incorporating structural aspects of the underlying graph. In this paper, we explore the use of affinity measures as features in graph neural networks, in particular measures arising from random walks, including effective resistance, hitting and commute times. We propose message passing networks based on these features and evaluate their performance on a variety of node and graph property prediction tasks. Our architecture has lower computational complexity, while our features are invariant to the permutations of the underlying graph. The measures we compute allow the network to exploit the connectivity properties of the graph, thereby allowing us to outperform relevant benchmarks for a wide variety of tasks, often with significantly fewer message passing steps. On one of the largest publicly available graph regression datasets, OGB-LSC-PCQM4Mv1, we obtain the best known single-model validation MAE at the time of writing. |
Concept Drift Adaptation by Exploiting Historical Knowledge | Incremental learning with concept drift has often been tackled by ensemble methods, where models built in the past can be re-trained to attain new models for the current data. Two design questions need to be addressed in developing ensemble methods for incremental learning with concept drift, i.e., which historical (i.e., previously trained) models should be preserved and how to utilize them. A novel ensemble learning method, namely Diversity and Transfer based Ensemble Learning (DTEL), is proposed in this paper. Given newly arrived data, DTEL uses each preserved historical model as an initial model and further trains it with the new data via transfer learning. Furthermore, DTEL preserves a diverse set of historical models, rather than a set of historical models that are merely accurate in terms of classification accuracy. Empirical studies on 15 synthetic data streams and 4 real-world data streams (all with concept drifts) demonstrate that DTEL can handle concept drift more effectively than 4 other state-of-the-art methods. |
On the Approximation Properties of Random ReLU Features | We study the approximation properties of random ReLU features through their reproducing kernel Hilbert space (RKHS). We first prove a universality theorem for the RKHS induced by random features whose feature maps are of the form of nodes in neural networks. The universality result implies that the random ReLU features method is a universally consistent learning algorithm. We prove that despite the universality of the RKHS induced by the random ReLU features, composition of functions in it generates substantially more complicated functions that are harder to approximate than those functions simply in the RKHS. We also prove that such composite functions can be efficiently approximated by multi-layer ReLU networks with bounded weights. This depth separation result shows that the random ReLU features models suffer from the same weakness as that of shallow models. We show in experiments that the performance of random ReLU features is comparable to that of random Fourier features and, in general, has a lower computational cost. We also demonstrate that when the target function is the composite function as described in the depth separation theorem, 3-layer neural networks indeed outperform both random ReLU features and 2-layer neural networks. |
A Scalable Hybrid Research Paper Recommender System for Microsoft Academic | We present the design and methodology for the large scale hybrid paper recommender system used by Microsoft Academic. The system provides recommendations for approximately 160 million English research papers and patents. Our approach handles incomplete citation information while also alleviating the cold-start problem that often affects other recommender systems. We use the Microsoft Academic Graph (MAG), titles, and available abstracts of research papers to build a recommendation list for all documents, thereby combining co-citation and content based approaches. Tuning system parameters also allows for blending and prioritization of each approach which, in turn, allows us to balance paper novelty versus authority in recommendation results. We evaluate the generated recommendations via a user study of 40 participants, with over 2400 recommendation pairs graded and discuss the quality of the results using P@10 and nDCG scores. We see that there is a strong correlation between participant scores and the similarity rankings produced by our system but that additional focus needs to be put towards improving recommender precision, particularly for content based recommendations. The results of the user survey and associated analysis scripts are made available via GitHub and the recommendations produced by our system are available as part of the MAG on Azure to facilitate further research and light up novel research paper recommendation applications. |
Third-Person Imitation Learning | Reinforcement learning (RL) makes it possible to train agents capable of achieving sophisticated goals in complex and uncertain environments. A key difficulty in reinforcement learning is specifying a reward function for the agent to optimize. Traditionally, imitation learning in RL has been used to overcome this problem. Unfortunately, hitherto imitation learning methods tend to require that demonstrations are supplied in the first-person: the agent is provided with a sequence of states and a specification of the actions that it should have taken. While powerful, this kind of imitation learning is limited by the relatively hard problem of collecting first-person demonstrations. Humans address this problem by learning from third-person demonstrations: they observe other humans perform tasks, infer the task, and accomplish the same task themselves. In this paper, we present a method for unsupervised third-person imitation learning. Here third-person refers to training an agent to correctly achieve a simple goal in a simple environment when it is provided a demonstration of a teacher achieving the same goal but from a different viewpoint; and unsupervised refers to the fact that the agent receives only these third-person demonstrations, and is not provided a correspondence between teacher states and student states. Our methods primary insight is that recent advances from domain confusion can be utilized to yield domain agnostic features which are crucial during the training process. To validate our approach, we report successful experiments on learning from third-person demonstrations in a pointmass domain, a reacher domain, and inverted pendulum. |
Assessing the Impact of a User-Item Collaborative Attack on Class of Users | Collaborative Filtering (CF) models lie at the core of most recommendation systems due to their state-of-the-art accuracy. They are commonly adopted in e-commerce and online services for their impact on sales volume and/or diversity, and their impact on companies' outcome. However, CF models are only as good as the interaction data they work with. As these models rely on outside sources of information, counterfeit data such as user ratings or reviews can be injected by attackers to manipulate the underlying data and alter the impact of resulting recommendations, thus implementing a so-called shilling attack. While previous works have focused on evaluating shilling attack strategies from a global perspective paying particular attention to the effect of the size of attacks and attacker's knowledge, in this work we explore the effectiveness of shilling attacks under novel aspects. First, we investigate the effect of attack strategies crafted on a target user in order to push the recommendation of a low-ranking item to a higher position, referred to as user-item attack. Second, we evaluate the effectiveness of attacks in altering the impact of different CF models by contemplating the class of the target user, from the perspective of the richness of her profile (i.e., cold v.s. warm user). Finally, similar to previous work we contemplate the size of attack (i.e., the amount of fake profiles injected) in examining their success. The results of experiments on two widely used datasets in business and movie domains, namely Yelp and MovieLens, suggest that warm and cold users exhibit contrasting behaviors in datasets with different characteristics. |
Siamese Capsule Networks | Capsule Networks have shown encouraging results on \textit{defacto} benchmark computer vision datasets such as MNIST, CIFAR and smallNORB. Although, they are yet to be tested on tasks where (1) the entities detected inherently have more complex internal representations and (2) there are very few instances per class to learn from and (3) where point-wise classification is not suitable. Hence, this paper carries out experiments on face verification in both controlled and uncontrolled settings that together address these points. In doing so we introduce \textit{Siamese Capsule Networks}, a new variant that can be used for pairwise learning tasks. The model is trained using contrastive loss with $\ell_2$-normalized capsule encoded pose features. We find that \textit{Siamese Capsule Networks} perform well against strong baselines on both pairwise learning datasets, yielding best results in the few-shot learning setting where image pairs in the test set contain unseen subjects. |
Improving the Performance of Fine-Grain Image Classifiers via Generative Data Augmentation | Recent advances in machine learning (ML) and computer vision tools have enabled applications in a wide variety of arenas such as financial analytics, medical diagnostics, and even within the Department of Defense. However, their widespread implementation in real-world use cases poses several challenges: (1) many applications are highly specialized, and hence operate in a \emph{sparse data} domain; (2) ML tools are sensitive to their training sets and typically require cumbersome, labor-intensive data collection and data labelling processes; and (3) ML tools can be extremely "black box," offering users little to no insight into the decision-making process or how new data might affect prediction performance. To address these challenges, we have designed and developed Data Augmentation from Proficient Pre-Training of Robust Generative Adversarial Networks (DAPPER GAN), an ML analytics support tool that automatically generates novel views of training images in order to improve downstream classifier performance. DAPPER GAN leverages high-fidelity embeddings generated by a StyleGAN2 model (trained on the LSUN cars dataset) to create novel imagery for previously unseen classes. We experimentally evaluate this technique on the Stanford Cars dataset, demonstrating improved vehicle make and model classification accuracy and reduced requirements for real data using our GAN based data augmentation framework. The method's validity was supported through an analysis of classifier performance on both augmented and non-augmented datasets, achieving comparable or better accuracy with up to 30\% less real data across visually similar classes. To support this method, we developed a novel augmentation method that can manipulate semantically meaningful dimensions (e.g., orientation) of the target object in the embedding space. |
Physical Action Categorization using Signal Analysis and Machine Learning | Daily life of thousands of individuals around the globe suffers due to physical or mental disability related to limb movement. The quality of life for such individuals can be made better by use of assistive applications and systems. In such scenario, mapping of physical actions from movement to a computer aided application can lead the way for solution. Surface Electromyography (sEMG) presents a non-invasive mechanism through which we can translate the physical movement to signals for classification and use in applications. In this paper, we propose a machine learning based framework for classification of 4 physical actions. The framework looks into the various features from different modalities which contribution from time domain, frequency domain, higher order statistics and inter channel statistics. Next, we conducted a comparative analysis of k-NN, SVM and ELM classifier using the feature set. Effect of different combinations of feature set has also been recorded. Finally, the classifier accuracy with SVM and 1-NN based classifier for a subset of features gives an accuracy of 95.21 and 95.83 respectively. Additionally, we have also proposed that dimensionality reduction by use of PCA leads to only a minor drop of less than 5.55% in accuracy while using only 9.22% of the original feature set. These finding are useful for algorithm designer to choose the best approach keeping in mind the resources available for execution of algorithm. |
Wearable Sensor Data Based Human Activity Recognition using Machine Learning: A new approach | Recent years have witnessed the rapid development of human activity recognition (HAR) based on wearable sensor data. One can find many practical applications in this area, especially in the field of health care. Many machine learning algorithms such as Decision Trees, Support Vector Machine, Naive Bayes, K-Nearest Neighbor, and Multilayer Perceptron are successfully used in HAR. Although these methods are fast and easy for implementation, they still have some limitations due to poor performance in a number of situations. In this paper, we propose a novel method based on the ensemble learning to boost the performance of these machine learning methods for HAR. |
Predicting Scores of Medical Imaging Segmentation Methods with Meta-Learning | Deep learning has led to state-of-the-art results for many medical imaging tasks, such as segmentation of different anatomical structures. With the increased numbers of deep learning publications and openly available code, the approach to choosing a model for a new task becomes more complicated, while time and (computational) resources are limited. A possible solution to choosing a model efficiently is meta-learning, a learning method in which prior performance of a model is used to predict the performance for new tasks. We investigate meta-learning for segmentation across ten datasets of different organs and modalities. We propose four ways to represent each dataset by meta-features: one based on statistical features of the images and three are based on deep learning features. We use support vector regression and deep neural networks to learn the relationship between the meta-features and prior model performance. On three external test datasets these methods give Dice scores within 0.10 of the true performance. These results demonstrate the potential of meta-learning in medical imaging. |
End-to-End Neural Speech Coding for Real-Time Communications | Deep-learning based methods have shown their advantages in audio coding over traditional ones but limited attention has been paid on real-time communications (RTC). This paper proposes the TFNet, an end-to-end neural speech codec with low latency for RTC. It takes an encoder-temporal filtering-decoder paradigm that has seldom been investigated in audio coding. An interleaved structure is proposed for temporal filtering to capture both short-term and long-term temporal dependencies. Furthermore, with end-to-end optimization, the TFNet is jointly optimized with speech enhancement and packet loss concealment, yielding a one-for-all network for three tasks. Both subjective and objective results demonstrate the efficiency of the proposed TFNet. |
Analyzing the Cross-Sensor Portability of Neural Network Architectures for LiDAR-based Semantic Labeling | State-of-the-art approaches for the semantic labeling of LiDAR point clouds heavily rely on the use of deep Convolutional Neural Networks (CNNs). However, transferring network architectures across different LiDAR sensor types represents a significant challenge, especially due to sensor specific design choices with regard to network architecture as well as data representation. In this paper we propose a new CNN architecture for the point-wise semantic labeling of LiDAR data which achieves state-of-the-art results while increasing portability across sensor types. This represents a significant advantage given the fast-paced development of LiDAR hardware technology. We perform a thorough quantitative cross-sensor analysis of semantic labeling performance in comparison to a state-of-the-art reference method. Our evaluation shows that the proposed architecture is indeed highly portable, yielding an improvement of 10 percentage points in the Intersection-over-Union (IoU) score when compared to the reference approach. Further, the results indicate that the proposed network architecture can provide an efficient way for the automated generation of large-scale training data for novel LiDAR sensor types without the need for extensive manual annotation or multi-modal label transfer. |
Machine learning for classification and quantification of monoclonal antibody preparations for cancer therapy | Monoclonal antibodies constitute one of the most important strategies to treat patients suffering from cancers such as hematological malignancies and solid tumors. In order to guarantee the quality of those preparations prepared at hospital, quality control has to be developed. The aim of this study was to explore a noninvasive, nondestructive, and rapid analytical method to ensure the quality of the final preparation without causing any delay in the process. We analyzed four mAbs (Inlfiximab, Bevacizumab, Ramucirumab and Rituximab) diluted at therapeutic concentration in chloride sodium 0.9% using Raman spectroscopy. To reduce the prediction errors obtained with traditional chemometric data analysis, we explored a data-driven approach using statistical machine learning methods where preprocessing and predictive models are jointly optimized. We prepared a data analytics workflow and submitted the problem to a collaborative data challenge platform called Rapid Analytics and Model Prototyping (RAMP). This allowed to use solutions from about 300 data scientists during five days of collaborative work. The prediction of the four mAbs samples was considerably improved with a misclassification rate and the mean error rate of 0.8% and 4%, respectively. |
Multilayer Perceptron Based Stress Evolution Analysis under DC Current Stressing for Multi-segment Wires | Electromigration (EM) is one of the major concerns in the reliability analysis of very large scale integration (VLSI) systems due to the continuous technology scaling. Accurately predicting the time-to-failure of integrated circuits (IC) becomes increasingly important for modern IC design. However, traditional methods are often not sufficiently accurate, leading to undesirable over-design especially in advanced technology nodes. In this paper, we propose an approach using multilayer perceptrons (MLP) to compute stress evolution in the interconnect trees during the void nucleation phase. The availability of a customized trial function for neural network training holds the promise of finding dynamic mesh-free stress evolution on complex interconnect trees under time-varying temperatures. Specifically, we formulate a new objective function considering the EM-induced coupled partial differential equations (PDEs), boundary conditions (BCs), and initial conditions to enforce the physics-based constraints in the spatial-temporal domain. The proposed model avoids meshing and reduces temporal iterations compared with conventional numerical approaches like FEM. Numerical results confirm its advantages on accuracy and computational performance. |
QuantumNAS: Noise-Adaptive Search for Robust Quantum Circuits | Quantum noise is the key challenge in Noisy Intermediate-Scale Quantum (NISQ) computers. Previous work for mitigating noise has primarily focused on gate-level or pulse-level noise-adaptive compilation. However, limited research efforts have explored a higher level of optimization by making the quantum circuits themselves resilient to noise. We propose QuantumNAS, a comprehensive framework for noise-adaptive co-search of the variational circuit and qubit mapping. Variational quantum circuits are a promising approach for constructing QML and quantum simulation. However, finding the best variational circuit and its optimal parameters is challenging due to the large design space and parameter training cost. We propose to decouple the circuit search and parameter training by introducing a novel SuperCircuit. The SuperCircuit is constructed with multiple layers of pre-defined parameterized gates and trained by iteratively sampling and updating the parameter subsets (SubCircuits) of it. It provides an accurate estimation of SubCircuits performance trained from scratch. Then we perform an evolutionary co-search of SubCircuit and its qubit mapping. The SubCircuit performance is estimated with parameters inherited from SuperCircuit and simulated with real device noise models. Finally, we perform iterative gate pruning and finetuning to remove redundant gates. Extensively evaluated with 12 QML and VQE benchmarks on 14 quantum computers, QuantumNAS significantly outperforms baselines. For QML, QuantumNAS is the first to demonstrate over 95% 2-class, 85% 4-class, and 32% 10-class classification accuracy on real QC. It also achieves the lowest eigenvalue for VQE tasks on H2, H2O, LiH, CH4, BeH2 compared with UCCSD. We also open-source TorchQuantum (https://github.com/mit-han-lab/torchquantum) for fast training of parameterized quantum circuits to facilitate future research. |
Formalizing Preferences Over Runtime Distributions | When trying to solve a computational problem we are often faced with a choice among algorithms that are all guaranteed to return the right answer but that differ in their runtime distributions (e.g., SAT solvers, sorting algorithms). This paper aims to lay theoretical foundations for such choices by formalizing preferences over runtime distributions. It might seem that we should simply prefer the algorithm that minimizes expected runtime. However, such preferences would be driven by exactly how slow our algorithm is on bad inputs, whereas in practice we are typically willing to cut off occasional, sufficiently long runs before they finish. We propose a principled alternative, taking a utility-theoretic approach to characterize the scoring functions that describe preferences over algorithms. These functions depend on the way our value for solving our problem decreases with time and on the distribution from which captimes are drawn. We describe examples of realistic utility functions and show how to leverage a maximum-entropy approach for modeling underspecified captime distributions. Finally, we show how to efficiently estimate an algorithm's expected utility from runtime samples. |
Anomaly Generation using Generative Adversarial Networks in Host Based Intrusion Detection | Generative adversarial networks have been able to generate striking results in various domains. This generation capability can be general while the networks gain deep understanding regarding the data distribution. In many domains, this data distribution consists of anomalies and normal data, with the anomalies commonly occurring relatively less, creating datasets that are imbalanced. The capabilities that generative adversarial networks offer can be leveraged to examine these anomalies and help alleviate the challenge that imbalanced datasets propose via creating synthetic anomalies. This anomaly generation can be specifically beneficial in domains that have costly data creation processes as well as inherently imbalanced datasets. One of the domains that fits this description is the host-based intrusion detection domain. In this work, ADFA-LD dataset is chosen as the dataset of interest containing system calls of small foot-print next generation attacks. The data is first converted into images, and then a Cycle-GAN is used to create images of anomalous data from images of normal data. The generated data is combined with the original dataset and is used to train a model to detect anomalies. By doing so, it is shown that the classification results are improved, with the AUC rising from 0.55 to 0.71, and the anomaly detection rate rising from 17.07% to 80.49%. The results are also compared to SMOTE, showing the potential presented by generative adversarial networks in anomaly generation. |
Single Image Texture Translation for Data Augmentation | Recent advances in image synthesis enables one to translate images by learning the mapping between a source domain and a target domain. Existing methods tend to learn the distributions by training a model on a variety of datasets, with results evaluated largely in a subjective manner. Relatively few works in this area, however, study the potential use of semantic image translation methods for image recognition tasks. In this paper, we explore the use of Single Image Texture Translation (SITT) for data augmentation. We first propose a lightweight model for translating texture to images based on a single input of source texture, allowing for fast training and testing. Based on SITT, we then explore the use of augmented data in long-tailed and few-shot image classification tasks. We find the proposed method is capable of translating input data into a target domain, leading to consistent improved image recognition performance. Finally, we examine how SITT and related image translation methods can provide a basis for a data-efficient, augmentation engineering approach to model training. |
Gaussian Variational State Estimation for Nonlinear State-Space Models | In this paper, the problem of state estimation, in the context of both filtering and smoothing, for nonlinear state-space models is considered. Due to the nonlinear nature of the models, the state estimation problem is generally intractable as it involves integrals of general nonlinear functions and the filtered and smoothed state distributions lack closed-form solutions. As such, it is common to approximate the state estimation problem. In this paper, we develop an assumed Gaussian solution based on variational inference, which offers the key advantage of a flexible, but principled, mechanism for approximating the required distributions. Our main contribution lies in a new formulation of the state estimation problem as an optimisation problem, which can then be solved using standard optimisation routines that employ exact first- and second-order derivatives. The resulting state estimation approach involves a minimal number of assumptions and applies directly to nonlinear systems with both Gaussian and non-Gaussian probabilistic models. The performance of our approach is demonstrated on several examples; a challenging scalar system, a model of a simple robotic system, and a target tracking problem using a von Mises-Fisher distribution and outperforms alternative assumed Gaussian approaches to state estimation. |
Contextualizing Multiple Tasks via Learning to Decompose | One single instance could possess multiple portraits and reveal diverse relationships with others according to different contexts. Those ambiguities increase the difficulty of learning a generalizable model when there exists one concept or mixed concepts in a task. We propose a general approach Learning to Decompose Network (LeadNet) for both two cases, which contextualizes a model through meta-learning multiple maps for concepts discovery -- the representations of instances are decomposed and adapted conditioned on the contexts. Through taking a holistic view over multiple latent components over instances in a sampled pseudo task, LeadNet learns to automatically select the right concept via incorporating those rich semantics inside and between objects. LeadNet demonstrates its superiority in various applications, including exploring multiple views of confusing tasks, out-of-distribution recognition, and few-shot image classification. |
Explaining Preferences with Shapley Values | While preference modelling is becoming one of the pillars of machine learning, the problem of preference explanation remains challenging and underexplored. In this paper, we propose \textsc{Pref-SHAP}, a Shapley value-based model explanation framework for pairwise comparison data. We derive the appropriate value functions for preference models and further extend the framework to model and explain \emph{context specific} information, such as the surface type in a tennis game. To demonstrate the utility of \textsc{Pref-SHAP}, we apply our method to a variety of synthetic and real-world datasets and show that richer and more insightful explanations can be obtained over the baseline. |
Learning to Interpret Satellite Images in Global Scale Using Wikipedia | Despite recent progress in computer vision, finegrained interpretation of satellite images remains challenging because of a lack of labeled training data. To overcome this limitation, we construct a novel dataset called WikiSatNet by pairing georeferenced Wikipedia articles with satellite imagery of their corresponding locations. We then propose two strategies to learn representations of satellite images by predicting properties of the corresponding articles from the images. Leveraging this new multi-modal dataset, we can drastically reduce the quantity of human-annotated labels and time required for downstream tasks. On the recently released fMoW dataset, our pre-training strategies can boost the performance of a model pre-trained on ImageNet by up to 4:5% in F1 score. |
Coordinate-wise Armijo's condition: General case | Let $z=(x,y)$ be coordinates for the product space $\mathbb{R}^{m_1}\times \mathbb{R}^{m_2}$. Let $f:\mathbb{R}^{m_1}\times \mathbb{R}^{m_2}\rightarrow \mathbb{R}$ be a $C^1$ function, and $\nabla f=(\partial _xf,\partial _yf)$ its gradient. Fix $0<\alpha <1$. For a point $(x,y) \in \mathbb{R}^{m_1}\times \mathbb{R}^{m_2}$, a number $\delta >0$ satisfies Armijo's condition at $(x,y)$ if the following inequality holds: \begin{eqnarray*} f(x-\delta \partial _xf,y-\delta \partial _yf)-f(x,y)\leq -\alpha \delta (||\partial _xf||^2+||\partial _yf||^2). \end{eqnarray*} In one previous paper, we proposed the following {\bf coordinate-wise} Armijo's condition. Fix again $0<\alpha <1$. A pair of positive numbers $\delta _1,\delta _2>0$ satisfies the coordinate-wise variant of Armijo's condition at $(x,y)$ if the following inequality holds: \begin{eqnarray*} [f(x-\delta _1\partial _xf(x,y), y-\delta _2\partial _y f(x,y))]-[f(x,y)]\leq -\alpha (\delta _1||\partial _xf(x,y)||^2+\delta _2||\partial _yf(x,y)||^2). \end{eqnarray*} Previously we applied this condition for functions of the form $f(x,y)=f(x)+g(y)$, and proved various convergent results for them. For a general function, it is crucial - for being able to do real computations - to have a systematic algorithm for obtaining $\delta _1$ and $\delta _2$ satisfying the coordinate-wise version of Armijo's condition, much like Backtracking for the usual Armijo's condition. In this paper we propose such an algorithm, and prove according convergent results. We then analyse and present experimental results for some functions such as $f(x,y)=a|x|+y$ (given by Asl and Overton in connection to Wolfe's method), $f(x,y)=x^3 sin (1/x) + y^3 sin(1/y)$ and Rosenbrock's function. |
EEG-Based Driver Drowsiness Estimation Using Convolutional Neural Networks | Deep learning, including convolutional neural networks (CNNs), has started finding applications in brain-computer interfaces (BCIs). However, so far most such approaches focused on BCI classification problems. This paper extends EEGNet, a 3-layer CNN model for BCI classification, to BCI regression, and also utilizes a novel spectral meta-learner for regression (SMLR) approach to aggregate multiple EEGNets for improved performance. Our model uses the power spectral density (PSD) of EEG signals as the input. Compared with raw EEG inputs, the PSD inputs can reduce the computational cost significantly, yet achieve much better regression performance. Experiments on driver drowsiness estimation from EEG signals demonstrate the outstanding performance of our approach. |
Cloud Detection through Wavelet Transforms in Machine Learning and Deep Learning | Cloud detection is a specialized application of image recognition and object detection using remotely sensed data. The task presents a number of challenges, including analyzing images obtained in visible, infrared and multi-spectral frequencies, usually without ground truth data for comparison. Moreover, machine learning and deep learning (MLDL) algorithms applied to this task are required to be computationally efficient, as they are typically deployed in low-power devices and called to operate in real-time. This paper explains Wavelet Transform (WT) theory, comparing it to more widely used image and signal processing transforms, and explores the use of WT as a powerful signal compressor and feature extractor for MLDL classifiers. |
App-Aware Response Synthesis for User Reviews | Responding to user reviews promptly and satisfactorily improves application ratings, which is key to application popularity and success. The proliferation of such reviews makes it virtually impossible for developers to keep up with responding manually. To address this challenge, recent work has shown the possibility of automatic response generation. However, because the training review-response pairs are aggregated from many different apps, it remains challenging for such models to generate app-specific responses, which, on the other hand, are often desirable as apps have different features and concerns. Solving the challenge by simply building a model per app (i.e., training with review-response pairs of a single app) may be insufficient because individual apps have limited review-response pairs, and such pairs typically lack the relevant information needed to respond to a new review. To enable app-specific response generation, this work proposes AARSynth: an app-aware response synthesis system. The key idea behind AARSynth is to augment the seq2seq model with information specific to a given app. Given a new user review, it first retrieves the top-K most relevant app reviews and the most relevant snippet from the app description. The retrieved information and the new user review are then fed into a fused machine learning model that integrates the seq2seq model with a machine reading comprehension model. The latter helps digest the retrieved reviews and app description. Finally, the fused model generates a response that is customized to the given app. We evaluated AARSynth using a large corpus of reviews and responses from Google Play. The results show that AARSynth outperforms the state-of-the-art system by 22.2% on BLEU-4 score. Furthermore, our human study shows that AARSynth produces a statistically significant improvement in response quality compared to the state-of-the-art system. |
Tensorial Mixture Models | Casting neural networks in generative frameworks is a highly sought-after endeavor these days. Contemporary methods, such as Generative Adversarial Networks, capture some of the generative capabilities, but not all. In particular, they lack the ability of tractable marginalization, and thus are not suitable for many tasks. Other methods, based on arithmetic circuits and sum-product networks, do allow tractable marginalization, but their performance is challenged by the need to learn the structure of a circuit. Building on the tractability of arithmetic circuits, we leverage concepts from tensor analysis, and derive a family of generative models we call Tensorial Mixture Models (TMMs). TMMs assume a simple convolutional network structure, and in addition, lend themselves to theoretical analyses that allow comprehensive understanding of the relation between their structure and their expressive properties. We thus obtain a generative model that is tractable on one hand, and on the other hand, allows effective representation of rich distributions in an easily controlled manner. These two capabilities are brought together in the task of classification under missing data, where TMMs deliver state of the art accuracies with seamless implementation and design. |
Deep Learning Reconstruction for 9-View Dual Energy CT Baggage Scanner | For homeland and transportation security applications, 2D X-ray explosive detection system (EDS) have been widely used, but they have limitations in recognizing 3D shape of the hidden objects. Among various types of 3D computed tomography (CT) systems to address this issue, this paper is interested in a stationary CT using fixed X-ray sources and detectors. However, due to the limited number of projection views, analytic reconstruction algorithms produce severe streaking artifacts. Inspired by recent success of deep learning approach for sparse view CT reconstruction, here we propose a novel image and sinogram domain deep learning architecture for 3D reconstruction from very sparse view measurement. The algorithm has been tested with the real data from a prototype 9-view dual energy stationary CT EDS carry-on baggage scanner developed by GEMSS Medical Systems, Korea, which confirms the superior reconstruction performance over the existing approaches. |
Rapid Near-Neighbor Interaction of High-dimensional Data via Hierarchical Clustering | Calculation of near-neighbor interactions among high dimensional, irregularly distributed data points is a fundamental task to many graph-based or kernel-based machine learning algorithms and applications. Such calculations, involving large, sparse interaction matrices, expose the limitation of conventional data-and-computation reordering techniques for improving space and time locality on modern computer memory hierarchies. We introduce a novel method for obtaining a matrix permutation that renders a desirable sparsity profile. The method is distinguished by the guiding principle to obtain a profile that is block-sparse with dense blocks. Our profile model and measure capture the essential properties affecting space and time locality, and permit variation in sparsity profile without imposing a restriction to a fixed pattern. The second distinction lies in an efficient algorithm for obtaining a desirable profile, via exploring and exploiting multi-scale cluster structure hidden in but intrinsic to the data. The algorithm accomplishes its task with key components for lower-dimensional embedding with data-specific principal feature axes, hierarchical data clustering, multi-level matrix compression storage, and multi-level interaction computations. We provide experimental results from case studies with two important data analysis algorithms. The resulting performance is remarkably comparable to the BLAS performance for the best-case interaction governed by a regularly banded matrix with the same sparsity. |
Integration of Pre-trained Networks with Continuous Token Interface for End-to-End Spoken Language Understanding | Most End-to-End (E2E) SLU networks leverage the pre-trained ASR networks but still lack the capability to understand the semantics of utterances, crucial for the SLU task. To solve this, recently proposed studies use pre-trained NLU networks. However, it is not trivial to fully utilize both pre-trained networks; many solutions were proposed, such as Knowledge Distillation, cross-modal shared embedding, and network integration with Interface. We propose a simple and robust integration method for the E2E SLU network with novel Interface, Continuous Token Interface (CTI), the junctional representation of the ASR and NLU networks when both networks are pre-trained with the same vocabulary. Because the only difference is the noise level, we directly feed the ASR network's output to the NLU network. Thus, we can train our SLU network in an E2E manner without additional modules, such as Gumbel-Softmax. We evaluate our model using SLURP, a challenging SLU dataset and achieve state-of-the-art scores on both intent classification and slot filling tasks. We also verify the NLU network, pre-trained with Masked Language Model, can utilize a noisy textual representation of CTI. Moreover, we show our model can be trained with multi-task learning from heterogeneous data even after integration with CTI. |
Dynamic Privacy Budget Allocation Improves Data Efficiency of Differentially Private Gradient Descent | Protecting privacy in learning while maintaining the model performance has become increasingly critical in many applications that involve sensitive data. A popular private learning framework is differentially private learning composed of many privatized gradient iterations by noising and clipping. Under the privacy constraint, it has been shown that the dynamic policies could improve the final iterate loss, namely the quality of published models. In this talk, we will introduce these dynamic techniques for learning rate, batch size, noise magnitude and gradient clipping. Also, we discuss how the dynamic policy could change the convergence bounds which further provides insight of the impact of dynamic methods. |
Unsupervised Domain Adaptation with Random Walks on Target Labelings | Unsupervised Domain Adaptation (DA) is used to automatize the task of labeling data: an unlabeled dataset (target) is annotated using a labeled dataset (source) from a related domain. We cast domain adaptation as the problem of finding stable labels for target examples. A new definition of label stability is proposed, motivated by a generalization error bound for large margin linear classifiers: a target labeling is stable when, with high probability, a classifier trained on a random subsample of the target with that labeling yields the same labeling. We find stable labelings using a random walk on a directed graph with transition probabilities based on labeling stability. The majority vote of those labelings visited by the walk yields a stable label for each target example. The resulting domain adaptation algorithm is strikingly easy to implement and apply: It does not rely on data transformations, which are in general computational prohibitive in the presence of many input features, and does not need to access the source data, which is advantageous when data sharing is restricted. By acting on the original feature space, our method is able to take full advantage of deep features from external pre-trained neural networks, as demonstrated by the results of our experiments. |
Unpriortized Autoencoder For Image Generation | In this paper, we treat the image generation task using an autoencoder, a representative latent model. Unlike many studies regularizing the latent variable's distribution by assuming a manually specified prior, we approach the image generation task using an autoencoder by directly estimating the latent distribution. To this end, we introduce 'latent density estimator' which captures latent distribution explicitly and propose its structure. Through experiments, we show that our generative model generates images with the improved visual quality compared to previous autoencoder-based generative models. |
Learning Graph-Level Representation for Drug Discovery | Predicating macroscopic influences of drugs on human body, like efficacy and toxicity, is a central problem of small-molecule based drug discovery. Molecules can be represented as an undirected graph, and we can utilize graph convolution networks to predication molecular properties. However, graph convolutional networks and other graph neural networks all focus on learning node-level representation rather than graph-level representation. Previous works simply sum all feature vectors for all nodes in the graph to obtain the graph feature vector for drug predication. In this paper, we introduce a dummy super node that is connected with all nodes in the graph by a directed edge as the representation of the graph and modify the graph operation to help the dummy super node learn graph-level feature. Thus, we can handle graph-level classification and regression in the same way as node-level classification and regression. In addition, we apply focal loss to address class imbalance in drug datasets. The experiments on MoleculeNet show that our method can effectively improve the performance of molecular properties predication. |
Fast and Robust Shortest Paths on Manifolds Learned from Data | We propose a fast, simple and robust algorithm for computing shortest paths and distances on Riemannian manifolds learned from data. This amounts to solving a system of ordinary differential equations (ODEs) subject to boundary conditions. Here standard solvers perform poorly because they require well-behaved Jacobians of the ODE, and usually, manifolds learned from data imply unstable and ill-conditioned Jacobians. Instead, we propose a fixed-point iteration scheme for solving the ODE that avoids Jacobians. This enhances the stability of the solver, while reduces the computational cost. In experiments involving both Riemannian metric learning and deep generative models we demonstrate significant improvements in speed and stability over both general-purpose state-of-the-art solvers as well as over specialized solvers. |
Error-Correcting Factorization | Error Correcting Output Codes (ECOC) is a successful technique in multi-class classification, which is a core problem in Pattern Recognition and Machine Learning. A major advantage of ECOC over other methods is that the multi- class problem is decoupled into a set of binary problems that are solved independently. However, literature defines a general error-correcting capability for ECOCs without analyzing how it distributes among classes, hindering a deeper analysis of pair-wise error-correction. To address these limitations this paper proposes an Error-Correcting Factorization (ECF) method, our contribution is three fold: (I) We propose a novel representation of the error-correction capability, called the design matrix, that enables us to build an ECOC on the basis of allocating correction to pairs of classes. (II) We derive the optimal code length of an ECOC using rank properties of the design matrix. (III) ECF is formulated as a discrete optimization problem, and a relaxed solution is found using an efficient constrained block coordinate descent approach. (IV) Enabled by the flexibility introduced with the design matrix we propose to allocate the error-correction on classes that are prone to confusion. Experimental results in several databases show that when allocating the error-correction to confusable classes ECF outperforms state-of-the-art approaches. |
Hello Edge: Keyword Spotting on Microcontrollers | Keyword spotting (KWS) is a critical component for enabling speech based user interactions on smart devices. It requires real-time response and high accuracy for good user experience. Recently, neural networks have become an attractive choice for KWS architecture because of their superior accuracy compared to traditional speech processing algorithms. Due to its always-on nature, KWS application has highly constrained power budget and typically runs on tiny microcontrollers with limited memory and compute capability. The design of neural network architecture for KWS must consider these constraints. In this work, we perform neural network architecture evaluation and exploration for running KWS on resource-constrained microcontrollers. We train various neural network architectures for keyword spotting published in literature to compare their accuracy and memory/compute requirements. We show that it is possible to optimize these neural network architectures to fit within the memory and compute constraints of microcontrollers without sacrificing accuracy. We further explore the depthwise separable convolutional neural network (DS-CNN) and compare it against other neural network architectures. DS-CNN achieves an accuracy of 95.4%, which is ~10% higher than the DNN model with similar number of parameters. |
Prescriptive Process Monitoring: Quo Vadis? | Prescriptive process monitoring methods seek to optimize a business process by recommending interventions at runtime to prevent negative outcomes or poorly performing cases. In recent years, various prescriptive process monitoring methods have been proposed. This paper studies existing methods in this field via a Systematic Literature Review (SLR). In order to structure the field, the paper proposes a framework for characterizing prescriptive process monitoring methods according to their performance objective, performance metrics, intervention types, modeling techniques, data inputs, and intervention policies. The SLR provides insights into challenges and areas for future research that could enhance the usefulness and applicability of prescriptive process monitoring methods. The paper highlights the need to validate existing and new methods in real-world settings, to extend the types of interventions beyond those related to the temporal and cost perspectives, and to design policies that take into account causality and second-order effects. |
SVTS: Scalable Video-to-Speech Synthesis | Video-to-speech synthesis (also known as lip-to-speech) refers to the translation of silent lip movements into the corresponding audio. This task has received an increasing amount of attention due to its self-supervised nature (i.e., can be trained without manual labelling) combined with the ever-growing collection of audio-visual data available online. Despite these strong motivations, contemporary video-to-speech works focus mainly on small- to medium-sized corpora with substantial constraints in both vocabulary and setting. In this work, we introduce a scalable video-to-speech framework consisting of two components: a video-to-spectrogram predictor and a pre-trained neural vocoder, which converts the mel-frequency spectrograms into waveform audio. We achieve state-of-the art results for GRID and considerably outperform previous approaches on LRW. More importantly, by focusing on spectrogram prediction using a simple feedforward model, we can efficiently and effectively scale our method to very large and unconstrained datasets: To the best of our knowledge, we are the first to show intelligible results on the challenging LRS3 dataset. |
Geometry of Linear Convolutional Networks | We study the family of functions that are represented by a linear convolutional neural network (LCN). These functions form a semi-algebraic subset of the set of linear maps from input space to output space. In contrast, the families of functions represented by fully-connected linear networks form algebraic sets. We observe that the functions represented by LCNs can be identified with polynomials that admit certain factorizations, and we use this perspective to describe the impact of the network's architecture on the geometry of the resulting function space. We further study the optimization of an objective function over an LCN, analyzing critical points in function space and in parameter space, and describing dynamical invariants for gradient descent. Overall, our theory predicts that the optimized parameters of an LCN will often correspond to repeated filters across layers, or filters that can be decomposed as repeated filters. We also conduct numerical and symbolic experiments that illustrate our results and present an in-depth analysis of the landscape for small architectures. |
An Inverse Power Method for Nonlinear Eigenproblems with Applications in 1-Spectral Clustering and Sparse PCA | Many problems in machine learning and statistics can be formulated as (generalized) eigenproblems. In terms of the associated optimization problem, computing linear eigenvectors amounts to finding critical points of a quadratic function subject to quadratic constraints. In this paper we show that a certain class of constrained optimization problems with nonquadratic objective and constraints can be understood as nonlinear eigenproblems. We derive a generalization of the inverse power method which is guaranteed to converge to a nonlinear eigenvector. We apply the inverse power method to 1-spectral clustering and sparse PCA which can naturally be formulated as nonlinear eigenproblems. In both applications we achieve state-of-the-art results in terms of solution quality and runtime. Moving beyond the standard eigenproblem should be useful also in many other applications and our inverse power method can be easily adapted to new problems. |
Learning to Optimize Tensor Programs | We introduce a learning-based framework to optimize tensor programs for deep learning workloads. Efficient implementations of tensor operators, such as matrix multiplication and high dimensional convolution, are key enablers of effective deep learning systems. However, existing systems rely on manually optimized libraries such as cuDNN where only a narrow range of server class GPUs are well-supported. The reliance on hardware-specific operator libraries limits the applicability of high-level graph optimizations and incurs significant engineering costs when deploying to new hardware targets. We use learning to remove this engineering burden. We learn domain-specific statistical cost models to guide the search of tensor operator implementations over billions of possible program variants. We further accelerate the search by effective model transfer across workloads. Experimental results show that our framework delivers performance competitive with state-of-the-art hand-tuned libraries for low-power CPU, mobile GPU, and server-class GPU. |
Visualizing the PHATE of Neural Networks | Understanding why and how certain neural networks outperform others is key to guiding future development of network architectures and optimization methods. To this end, we introduce a novel visualization algorithm that reveals the internal geometry of such networks: Multislice PHATE (M-PHATE), the first method designed explicitly to visualize how a neural network's hidden representations of data evolve throughout the course of training. We demonstrate that our visualization provides intuitive, detailed summaries of the learning dynamics beyond simple global measures (i.e., validation loss and accuracy), without the need to access validation data. Furthermore, M-PHATE better captures both the dynamics and community structure of the hidden units as compared to visualization based on standard dimensionality reduction methods (e.g., ISOMAP, t-SNE). We demonstrate M-PHATE with two vignettes: continual learning and generalization. In the former, the M-PHATE visualizations display the mechanism of "catastrophic forgetting" which is a major challenge for learning in task-switching contexts. In the latter, our visualizations reveal how increased heterogeneity among hidden units correlates with improved generalization performance. An implementation of M-PHATE, along with scripts to reproduce the figures in this paper, is available at https://github.com/scottgigante/M-PHATE. |
A short note on the decision tree based neural turing machine | Turing machine and decision tree have developed independently for a long time. With the recent development of differentiable models, there is an intersection between them. Neural turing machine(NTM) opens door for the memory network. It use differentiable attention mechanism to read/write external memory bank. Differentiable forest brings differentiable properties to classical decision tree. In this short note, we show the deep connection between these two models. That is: differentiable forest is a special case of NTM. Differentiable forest is actually decision tree based neural turing machine. Based on this deep connection, we propose a response augmented differential forest (RaDF). The controller of RaDF is differentiable forest, the external memory of RaDF are response vectors which would be read/write by leaf nodes. |
Diversity-Inducing Policy Gradient: Using Maximum Mean Discrepancy to Find a Set of Diverse Policies | Standard reinforcement learning methods aim to master one way of solving a task whereas there may exist multiple near-optimal policies. Being able to identify this collection of near-optimal policies can allow a domain expert to efficiently explore the space of reasonable solutions. Unfortunately, existing approaches that quantify uncertainty over policies are not ultimately relevant to finding policies with qualitatively distinct behaviors. In this work, we formalize the difference between policies as a difference between the distribution of trajectories induced by each policy, which encourages diversity with respect to both state visitation and action choices. We derive a gradient-based optimization technique that can be combined with existing policy gradient methods to now identify diverse collections of well-performing policies. We demonstrate our approach on benchmarks and a healthcare task. |
Beyond IID: data-driven decision-making in heterogeneous environments | In this work, we study data-driven decision-making and depart from the classical identically and independently distributed (i.i.d.) assumption. We present a new framework in which historical samples are generated from unknown and different distributions, which we dub heterogeneous environments. These distributions are assumed to lie in a heterogeneity ball with known radius and centered around the (also) unknown future (out-of-sample) distribution on which the performance of a decision will be evaluated. We quantify the asymptotic worst-case regret that is achievable by central data-driven policies such as Sample Average Approximation, but also by rate-optimal ones, as a function of the radius of the heterogeneity ball. Our work shows that the type of achievable performance varies considerably across different combinations of problem classes and notions of heterogeneity. We demonstrate the versatility of our framework by comparing achievable guarantees for the heterogeneous version of widely studied data-driven problems such as pricing, ski-rental, and newsvendor. En route, we establish a new connection between data-driven decision-making and distributionally robust optimization. |
Gradient Backpropagation Through Combinatorial Algorithms: Identity with Projection Works | Embedding discrete solvers as differentiable layers has given modern deep learning architectures combinatorial expressivity and discrete reasoning capabilities. The derivative of these solvers is zero or undefined, therefore a meaningful replacement is crucial for effective gradient-based learning. Prior works rely on smoothing the solver with input perturbations, relaxing the solver to continuous problems, or interpolating the loss landscape with techniques that typically require additional solver calls, introduce extra hyper-parameters or compromise performance. We propose a principled approach to exploit the geometry of the discrete solution space to treat the solver as a negative identity on the backward pass and further provide a theoretical justification. Our experiments demonstrate that such a straightforward hyper-parameter-free approach is on-par with or outperforms previous more complex methods on numerous experiments such as Traveling Salesman Problem, Shortest Path, Deep Graph Matching, and backpropagating through discrete samplers. Furthermore, we substitute the previously proposed problem-specific and label-dependent margin by a generic regularization procedure that prevents cost collapse and increases robustness. |
DeepSpark: A Spark-Based Distributed Deep Learning Framework for Commodity Clusters | The increasing complexity of deep neural networks (DNNs) has made it challenging to exploit existing large-scale data processing pipelines for handling massive data and parameters involved in DNN training. Distributed computing platforms and GPGPU-based acceleration provide a mainstream solution to this computational challenge. In this paper, we propose DeepSpark, a distributed and parallel deep learning framework that exploits Apache Spark on commodity clusters. To support parallel operations, DeepSpark automatically distributes workloads and parameters to Caffe/Tensorflow-running nodes using Spark, and iteratively aggregates training results by a novel lock-free asynchronous variant of the popular elastic averaging stochastic gradient descent based update scheme, effectively complementing the synchronized processing capabilities of Spark. DeepSpark is an on-going project, and the current release is available at http://deepspark.snu.ac.kr. |
Finding Patient Zero: Learning Contagion Source with Graph Neural Networks | Locating the source of an epidemic, or patient zero (P0), can provide critical insights into the infection's transmission course and allow efficient resource allocation. Existing methods use graph-theoretic centrality measures and expensive message-passing algorithms, requiring knowledge of the underlying dynamics and its parameters. In this paper, we revisit this problem using graph neural networks (GNNs) to learn P0. We establish a theoretical limit for the identification of P0 in a class of epidemic models. We evaluate our method against different epidemic models on both synthetic and a real-world contact network considering a disease with history and characteristics of COVID-19. % We observe that GNNs can identify P0 close to the theoretical bound on accuracy, without explicit input of dynamics or its parameters. In addition, GNN is over 100 times faster than classic methods for inference on arbitrary graph topologies. Our theoretical bound also shows that the epidemic is like a ticking clock, emphasizing the importance of early contact-tracing. We find a maximum time after which accurate recovery of the source becomes impossible, regardless of the algorithm used. |
Neural Architecture Search for Speech Emotion Recognition | Deep neural networks have brought significant advancements to speech emotion recognition (SER). However, the architecture design in SER is mainly based on expert knowledge and empirical (trial-and-error) evaluations, which is time-consuming and resource intensive. In this paper, we propose to apply neural architecture search (NAS) techniques to automatically configure the SER models. To accelerate the candidate architecture optimization, we propose a uniform path dropout strategy to encourage all candidate architecture operations to be equally optimized. Experimental results of two different neural structures on IEMOCAP show that NAS can improve SER performance (54.89\% to 56.28\%) while maintaining model parameter sizes. The proposed dropout strategy also shows superiority over the previous approaches. |
Influence-Based Multi-Agent Exploration | Intrinsically motivated reinforcement learning aims to address the exploration challenge for sparse-reward tasks. However, the study of exploration methods in transition-dependent multi-agent settings is largely absent from the literature. We aim to take a step towards solving this problem. We present two exploration methods: exploration via information-theoretic influence (EITI) and exploration via decision-theoretic influence (EDTI), by exploiting the role of interaction in coordinated behaviors of agents. EITI uses mutual information to capture influence transition dynamics. EDTI uses a novel intrinsic reward, called Value of Interaction (VoI), to characterize and quantify the influence of one agent's behavior on expected returns of other agents. By optimizing EITI or EDTI objective as a regularizer, agents are encouraged to coordinate their exploration and learn policies to optimize team performance. We show how to optimize these regularizers so that they can be easily integrated with policy gradient reinforcement learning. The resulting update rule draws a connection between coordinated exploration and intrinsic reward distribution. Finally, we empirically demonstrate the significant strength of our method in a variety of multi-agent scenarios. |
Identifying Hidden Buyers in Darknet Markets via Dirichlet Hawkes Process | The darknet markets are notorious black markets in cyberspace, which involve selling or brokering drugs, weapons, stolen credit cards, and other illicit goods. To combat illicit transactions in the cyberspace, it is important to analyze the behaviors of participants in darknet markets. Currently, many studies focus on studying the behavior of vendors. However, there is no much work on analyzing buyers. The key challenge is that the buyers are anonymized in darknet markets. For most of the darknet markets, We only observe the first and last digits of a buyer's ID, such as ``a**b''. To tackle this challenge, we propose a hidden buyer identification model, called UNMIX, which can group the transactions from one hidden buyer into one cluster given a transaction sequence from an anonymized ID. UNMIX is able to model the temporal dynamics information as well as the product, comment, and vendor information associated with each transaction. As a result, the transactions with similar patterns in terms of time and content group together as the subsequence from one hidden buyer. Experiments on the data collected from three real-world darknet markets demonstrate the effectiveness of our approach measured by various clustering metrics. Case studies on real transaction sequences explicitly show that our approach can group transactions with similar patterns into the same clusters. |
Deep Representation Learning for Clustering of Health Tweets | Twitter has been a prominent social media platform for mining population-level health data and accurate clustering of health-related tweets into topics is important for extracting relevant health insights. In this work, we propose deep convolutional autoencoders for learning compact representations of health-related tweets, further to be employed in clustering. We compare our method to several conventional tweet representation methods including bag-of-words, term frequency-inverse document frequency, Latent Dirichlet Allocation and Non-negative Matrix Factorization with 3 different clustering algorithms. Our results show that the clustering performance using proposed representation learning scheme significantly outperforms that of conventional methods for all experiments of different number of clusters. In addition, we propose a constraint on the learned representations during the neural network training in order to further enhance the clustering performance. All in all, this study introduces utilization of deep neural network-based architectures, i.e., deep convolutional autoencoders, for learning informative representations of health-related tweets. |
Universal MMSE Filtering With Logarithmic Adaptive Regret | We consider the problem of online estimation of a real-valued signal corrupted by oblivious zero-mean noise using linear estimators. The estimator is required to iteratively predict the underlying signal based on the current and several last noisy observations, and its performance is measured by the mean-square-error. We describe and analyze an algorithm for this task which: 1. Achieves logarithmic adaptive regret against the best linear filter in hindsight. This bound is assyptotically tight, and resolves the question of Moon and Weissman [1]. 2. Runs in linear time in terms of the number of filter coefficients. Previous constructions required at least quadratic time. |
Crowd Counting via Segmentation Guided Attention Networks and Curriculum Loss | Automatic crowd behaviour analysis is an important task for intelligent transportation systems to enable effective flow control and dynamic route planning for varying road participants. Crowd counting is one of the keys to automatic crowd behaviour analysis. Crowd counting using deep convolutional neural networks (CNN) has achieved encouraging progress in recent years. Researchers have devoted much effort to the design of variant CNN architectures and most of them are based on the pre-trained VGG16 model. Due to the insufficient expressive capacity, the backbone network of VGG16 is usually followed by another cumbersome network specially designed for good counting performance. Although VGG models have been outperformed by Inception models in image classification tasks, the existing crowd counting networks built with Inception modules still only have a small number of layers with basic types of Inception modules. To fill in this gap, in this paper, we firstly benchmark the baseline Inception-v3 model on commonly used crowd counting datasets and achieve surprisingly good performance comparable with or better than most existing crowd counting models. Subsequently, we push the boundary of this disruptive work further by proposing a Segmentation Guided Attention Network (SGANet) with Inception-v3 as the backbone and a novel curriculum loss for crowd counting. We conduct thorough experiments to compare the performance of our SGANet with prior arts and the proposed model can achieve state-of-the-art performance with MAE of 57.6, 6.3 and 87.6 on ShanghaiTechA, ShanghaiTechB and UCF\_QNRF, respectively. |