October 19, 2019

3218 words 16 mins read

Paper Group ANR 270

Partitioned Variational Inference: A unified framework encompassing federated and continual learning. BALSON: Bayesian Least Squares Optimization with Nonnegative L1-Norm Constraint. Global Search with Bernoulli Alternation Kernel for Task-oriented Grasping Informed by Simulation. The Barbados 2018 List of Open Issues in Continual Learning. Closed- …

Partitioned Variational Inference: A unified framework encompassing federated and continual learning


Title	Partitioned Variational Inference: A unified framework encompassing federated and continual learning
Authors	Thang D. Bui, Cuong V. Nguyen, Siddharth Swaroop, Richard E. Turner
Abstract	Variational inference (VI) has become the method of choice for fitting many modern probabilistic models. However, practitioners are faced with a fragmented literature that offers a bewildering array of algorithmic options. First, the variational family. Second, the granularity of the updates e.g. whether the updates are local to each data point and employ message passing or global. Third, the method of optimization (bespoke or blackbox, closed-form or stochastic updates, etc.). This paper presents a new framework, termed Partitioned Variational Inference (PVI), that explicitly acknowledges these algorithmic dimensions of VI, unifies disparate literature, and provides guidance on usage. Crucially, the proposed PVI framework allows us to identify new ways of performing VI that are ideally suited to challenging learning scenarios including federated learning (where distributed computing is leveraged to process non-centralized data) and continual learning (where new data and tasks arrive over time and must be accommodated quickly). We showcase these new capabilities by developing communication-efficient federated training of Bayesian neural networks and continual learning for Gaussian process models with private pseudo-points. The new methods significantly outperform the state-of-the-art, whilst being almost as straightforward to implement as standard VI.
Tasks	Continual Learning
Published	2018-11-27
URL	http://arxiv.org/abs/1811.11206v1
PDF	http://arxiv.org/pdf/1811.11206v1.pdf
PWC	https://paperswithcode.com/paper/partitioned-variational-inference-a-unified
Repo
Framework

BALSON: Bayesian Least Squares Optimization with Nonnegative L1-Norm Constraint


Title	BALSON: Bayesian Least Squares Optimization with Nonnegative L1-Norm Constraint
Authors	Jiyang Xie, Zhanyu Ma, Guoqiang Zhang, Jing-Hao Xue, Jen-Tzung Chien, Zhiqing Lin, Jun Guo
Abstract	A Bayesian approach termed BAyesian Least Squares Optimization with Nonnegative L1-norm constraint (BALSON) is proposed. The error distribution of data fitting is described by Gaussian likelihood. The parameter distribution is assumed to be a Dirichlet distribution. With the Bayes rule, searching for the optimal parameters is equivalent to finding the mode of the posterior distribution. In order to explicitly characterize the nonnegative L1-norm constraint of the parameters, we further approximate the true posterior distribution by a Dirichlet distribution. We estimate the statistics of the approximating Dirichlet posterior distribution by sampling methods. Four sampling methods have been introduced. With the estimated posterior distributions, the original parameters can be effectively reconstructed in polynomial fitting problems, and the BALSON framework is found to perform better than conventional methods.
Tasks
Published	2018-07-08
URL	http://arxiv.org/abs/1807.02795v1
PDF	http://arxiv.org/pdf/1807.02795v1.pdf
PWC	https://paperswithcode.com/paper/balson-bayesian-least-squares-optimization
Repo
Framework

Global Search with Bernoulli Alternation Kernel for Task-oriented Grasping Informed by Simulation


Title	Global Search with Bernoulli Alternation Kernel for Task-oriented Grasping Informed by Simulation
Authors	Rika Antonova, Mia Kokic, Johannes A. Stork, Danica Kragic
Abstract	We develop an approach that benefits from large simulated datasets and takes full advantage of the limited online data that is most relevant. We propose a variant of Bayesian optimization that alternates between using informed and uninformed kernels. With this Bernoulli Alternation Kernel we ensure that discrepancies between simulation and reality do not hinder adapting robot control policies online. The proposed approach is applied to a challenging real-world problem of task-oriented grasping with novel objects. Our further contribution is a neural network architecture and training pipeline that use experience from grasping objects in simulation to learn grasp stability scores. We learn task scores from a labeled dataset with a convolutional network, which is used to construct an informed kernel for our variant of Bayesian optimization. Experiments on an ABB Yumi robot with real sensor data demonstrate success of our approach, despite the challenge of fulfilling task requirements and high uncertainty over physical properties of objects.
Tasks
Published	2018-10-10
URL	http://arxiv.org/abs/1810.04438v1
PDF	http://arxiv.org/pdf/1810.04438v1.pdf
PWC	https://paperswithcode.com/paper/global-search-with-bernoulli-alternation
Repo
Framework

The Barbados 2018 List of Open Issues in Continual Learning


Title	The Barbados 2018 List of Open Issues in Continual Learning
Authors	Tom Schaul, Hado van Hasselt, Joseph Modayil, Martha White, Adam White, Pierre-Luc Bacon, Jean Harb, Shibl Mourad, Marc Bellemare, Doina Precup
Abstract	We want to make progress toward artificial general intelligence, namely general-purpose agents that autonomously learn how to competently act in complex environments. The purpose of this report is to sketch a research outline, share some of the most important open issues we are facing, and stimulate further discussion in the community. The content is based on some of our discussions during a week-long workshop held in Barbados in February 2018.
Tasks	Continual Learning
Published	2018-11-16
URL	http://arxiv.org/abs/1811.07004v1
PDF	http://arxiv.org/pdf/1811.07004v1.pdf
PWC	https://paperswithcode.com/paper/the-barbados-2018-list-of-open-issues-in
Repo
Framework

Closed-Loop Memory GAN for Continual Learning


Title	Closed-Loop Memory GAN for Continual Learning
Authors	Amanda Rios, Laurent Itti
Abstract	Sequential learning of tasks using gradient descent leads to an unremitting decline in the accuracy of tasks for which training data is no longer available, termed catastrophic forgetting. Generative models have been explored as a means to approximate the distribution of old tasks and bypass storage of real data. Here we propose a cumulative closed-loop memory replay GAN (CloGAN) provided with external regularization by a small memory unit selected for maximum sample diversity. We evaluate incremental class learning using a notoriously hard paradigm, single-headed learning, in which each task is a disjoint subset of classes in the overall dataset, and performance is evaluated on all previous classes. First, we show that when constructing a dynamic memory unit to preserve sample heterogeneity, model performance asymptotically approaches training on the full dataset. We then show that using a stochastic generator to continuously output fresh new images during training increases performance significantly further meanwhile generating quality images. We compare our approach to several baselines including fine-tuning by gradient descent (FGD), Elastic Weight Consolidation (EWC), Deep Generative Replay (DGR) and Memory Replay GAN (MeRGAN). Our method has very low long-term memory cost, the memory unit, as well as negligible intermediate memory storage.
Tasks	Continual Learning
Published	2018-11-03
URL	https://arxiv.org/abs/1811.01146v2
PDF	https://arxiv.org/pdf/1811.01146v2.pdf
PWC	https://paperswithcode.com/paper/closed-loop-gan-for-continual-learning
Repo
Framework

From Gameplay to Symbolic Reasoning: Learning SAT Solver Heuristics in the Style of Alpha(Go) Zero


Title	From Gameplay to Symbolic Reasoning: Learning SAT Solver Heuristics in the Style of Alpha(Go) Zero
Authors	Fei Wang, Tiark Rompf
Abstract	Despite the recent successes of deep neural networks in various fields such as image and speech recognition, natural language processing, and reinforcement learning, we still face big challenges in bringing the power of numeric optimization to symbolic reasoning. Researchers have proposed different avenues such as neural machine translation for proof synthesis, vectorization of symbols and expressions for representing symbolic patterns, and coupling of neural back-ends for dimensionality reduction with symbolic front-ends for decision making. However, these initial explorations are still only point solutions, and bear other shortcomings such as lack of correctness guarantees. In this paper, we present our approach of casting symbolic reasoning as games, and directly harnessing the power of deep reinforcement learning in the style of Alpha(Go) Zero on symbolic problems. Using the Boolean Satisfiability (SAT) problem as showcase, we demonstrate the feasibility of our method, and the advantages of modularity, efficiency, and correctness guarantees.
Tasks	Decision Making, Dimensionality Reduction, Machine Translation, Speech Recognition
Published	2018-02-14
URL	http://arxiv.org/abs/1802.05340v1
PDF	http://arxiv.org/pdf/1802.05340v1.pdf
PWC	https://paperswithcode.com/paper/from-gameplay-to-symbolic-reasoning-learning
Repo
Framework

Learning to Automatically Generate Fill-In-The-Blank Quizzes


Title	Learning to Automatically Generate Fill-In-The-Blank Quizzes
Authors	Edison Marrese-Taylor, Ai Nakajima, Yutaka Matsuo, Ono Yuichi
Abstract	In this paper we formalize the problem automatic fill-in-the-blank question generation using two standard NLP machine learning schemes, proposing concrete deep learning models for each. We present an empirical study based on data obtained from a language learning platform showing that both of our proposed settings offer promising results.
Tasks	Question Generation
Published	2018-06-12
URL	http://arxiv.org/abs/1806.04524v1
PDF	http://arxiv.org/pdf/1806.04524v1.pdf
PWC	https://paperswithcode.com/paper/learning-to-automatically-generate-fill-in
Repo
Framework

Video Summarisation by Classification with Deep Reinforcement Learning


Title	Video Summarisation by Classification with Deep Reinforcement Learning
Authors	Kaiyang Zhou, Tao Xiang, Andrea Cavallaro
Abstract	Most existing video summarisation methods are based on either supervised or unsupervised learning. In this paper, we propose a reinforcement learning-based weakly supervised method that exploits easy-to-obtain, video-level category labels and encourages summaries to contain category-related information and maintain category recognisability. Specifically, We formulate video summarisation as a sequential decision-making process and train a summarisation network with deep Q-learning (DQSN). A companion classification network is also trained to provide rewards for training the DQSN. With the classification network, we develop a global recognisability reward based on the classification result. Critically, a novel dense ranking-based reward is also proposed in order to cope with the temporally delayed and sparse reward problems for long sequence reinforcement learning. Extensive experiments on two benchmark datasets show that the proposed approach achieves state-of-the-art performance.
Tasks	Decision Making, Q-Learning
Published	2018-07-09
URL	http://arxiv.org/abs/1807.03089v3
PDF	http://arxiv.org/pdf/1807.03089v3.pdf
PWC	https://paperswithcode.com/paper/video-summarisation-by-classification-with
Repo
Framework

HFL-RC System at SemEval-2018 Task 11: Hybrid Multi-Aspects Model for Commonsense Reading Comprehension


Title	HFL-RC System at SemEval-2018 Task 11: Hybrid Multi-Aspects Model for Commonsense Reading Comprehension
Authors	Zhipeng Chen, Yiming Cui, Wentao Ma, Shijin Wang, Ting Liu, Guoping Hu
Abstract	This paper describes the system which got the state-of-the-art results at SemEval-2018 Task 11: Machine Comprehension using Commonsense Knowledge. In this paper, we present a neural network called Hybrid Multi-Aspects (HMA) model, which mimic the human’s intuitions on dealing with the multiple-choice reading comprehension. In this model, we aim to produce the predictions in multiple aspects by calculating attention among the text, question and choices, and combine these results for final predictions. Experimental results show that our HMA model could give substantial improvements over the baseline system and got the first place on the final test set leaderboard with the accuracy of 84.13%.
Tasks	Reading Comprehension
Published	2018-03-15
URL	http://arxiv.org/abs/1803.05655v1
PDF	http://arxiv.org/pdf/1803.05655v1.pdf
PWC	https://paperswithcode.com/paper/hfl-rc-system-at-semeval-2018-task-11-hybrid
Repo
Framework

Adversarial Signal Denoising with Encoder-Decoder Networks


Title	Adversarial Signal Denoising with Encoder-Decoder Networks
Authors	Leslie Casas, Attila Klimmek, Nassir Navab, Vasileios Belagiannis
Abstract	The presence of noise is common in signal processing independent of the signal type. Deep neural networks have shown good performance in removing signal noise, especially in the image domain. In this work, we consider deep neural networks as a denoising tool where our focus is on one dimensional signals. For that purpose, we introduce an encoder-decoder network architecture to denoise signals, represented by a sequence of measurements. Instead of relying only on the standard reconstruction error to train the encoder-decoder network, we treat the task of signal denoising as distribution alignment between the clean and noisy signals. Then, we propose to train the encoder-decoder with adversarial learning, where the goal is to align the clean and noisy signal latent representation. Unlike standard adversarial learning, we do not have access to the distribution of the clean signal’s latent representation in advance. For that reason, we propose a new formulation where both clean and noisy signals pass through the encoder to produce the latent representation. Afterwards, a discriminator neural network has to detect whether the latent representation comes from the clean or noisy signal. At the end of training, aligning the two signal distributions results in removing the noise. In our experiments, we study two signal types with complex noise models. First, we evaluate on electrocardiography and later on motion signal denoising. We show better performance than the related learning-based and non-learning approaches, such as autoencoders, wavenet denoiser, recurrent neural networks and wavelets, demonstrating the benefits of adversarial learning for one dimensional signal denoising.
Tasks	Denoising
Published	2018-12-20
URL	https://arxiv.org/abs/1812.08555v2
PDF	https://arxiv.org/pdf/1812.08555v2.pdf
PWC	https://paperswithcode.com/paper/adversarial-signal-denoising-with-encoder
Repo
Framework

Optimization of Transfer Learning for Sign Language Recognition Targeting Mobile Platform


Title	Optimization of Transfer Learning for Sign Language Recognition Targeting Mobile Platform
Authors	Dhruv Rathi
Abstract	The target of this research is to experiment, iterate and recommend a system that is successful in recognition of American Sign Language (ASL). It is a challenging as well as an interesting problem that if solved will bring a leap in social and technological aspects alike. In this paper, we propose a real-time recognizer of ASL based on a mobile platform, so that it will have more accessibility and provides an ease of use. The technique implemented is Transfer Learning of new data of Hand gestures for alphabets in ASL to be modelled on various pre-trained high- end models and optimize the best model to run on a mobile platform considering the various limitations of the same during optimization. The data used consists of 27,455 images of 24 alphabets of ASL. The optimized model when ran over a memory-efficient mobile application, provides an accuracy of 95.03% of accurate recognition with an average recognition time of 2.42 seconds. This method ensures considerable discrimination in accuracy and recognition time than the previous research.
Tasks	Sign Language Recognition, Transfer Learning
Published	2018-05-17
URL	http://arxiv.org/abs/1805.06618v1
PDF	http://arxiv.org/pdf/1805.06618v1.pdf
PWC	https://paperswithcode.com/paper/optimization-of-transfer-learning-for-sign
Repo
Framework

Information Directed Sampling and Bandits with Heteroscedastic Noise


Title	Information Directed Sampling and Bandits with Heteroscedastic Noise
Authors	Johannes Kirschner, Andreas Krause
Abstract	In the stochastic bandit problem, the goal is to maximize an unknown function via a sequence of noisy evaluations. Typically, the observation noise is assumed to be independent of the evaluation point and to satisfy a tail bound uniformly on the domain; a restrictive assumption for many applications. In this work, we consider bandits with heteroscedastic noise, where we explicitly allow the noise distribution to depend on the evaluation point. We show that this leads to new trade-offs for information and regret, which are not taken into account by existing approaches like upper confidence bound algorithms (UCB) or Thompson Sampling. To address these shortcomings, we introduce a frequentist regret analysis framework, that is similar to the Bayesian framework of Russo and Van Roy (2014), and we prove a new high-probability regret bound for general, possibly randomized policies, which depends on a quantity we refer to as regret-information ratio. From this bound, we define a frequentist version of Information Directed Sampling (IDS) to minimize the regret-information ratio over all possible action sampling distributions. This further relies on concentration inequalities for online least squares regression in separable Hilbert spaces, which we generalize to the case of heteroscedastic noise. We then formulate several variants of IDS for linear and reproducing kernel Hilbert space response functions, yielding novel algorithms for Bayesian optimization. We also prove frequentist regret bounds, which in the homoscedastic case recover known bounds for UCB, but can be much better when the noise is heteroscedastic. Empirically, we demonstrate in a linear setting with heteroscedastic noise, that some of our methods can outperform UCB and Thompson Sampling, while staying competitive when the noise is homoscedastic.
Tasks
Published	2018-01-29
URL	http://arxiv.org/abs/1801.09667v2
PDF	http://arxiv.org/pdf/1801.09667v2.pdf
PWC	https://paperswithcode.com/paper/information-directed-sampling-and-bandits
Repo
Framework

PALM: An Incremental Construction of Hyperplanes for Data Stream Regression


Title	PALM: An Incremental Construction of Hyperplanes for Data Stream Regression
Authors	Md Meftahul Ferdaus, Mahardhika Pratama, Sreenatha G. Anavatti, Matthew A. Garratt
Abstract	Data stream has been the underlying challenge in the age of big data because it calls for real-time data processing with the absence of a retraining process and/or an iterative learning approach. In realm of fuzzy system community, data stream is handled by algorithmic development of self-adaptive neurofuzzy systems (SANFS) characterized by the single-pass learning mode and the open structure property which enables effective handling of fast and rapidly changing natures of data streams. The underlying bottleneck of SANFSs lies in its design principle which involves a high number of free parameters (rule premise and rule consequent) to be adapted in the training process. This figure can even double in the case of type-2 fuzzy system. In this work, a novel SANFS, namely parsimonious learning machine (PALM), is proposed. PALM features utilization of a new type of fuzzy rule based on the concept of hyperplane clustering which significantly reduces the number of network parameters because it has no rule premise parameters. PALM is proposed in both type-1 and type-2 fuzzy systems where all of which characterize a fully dynamic rule-based system. That is, it is capable of automatically generating, merging and tuning the hyperplane-based fuzzy rule in the single pass manner. Moreover, an extension of PALM, namely recurrent PALM (rPALM), is proposed and adopts the concept of teacher-forcing mechanism in the deep learning literature. The efficacy of PALM has been evaluated through numerical study with six real-world and synthetic data streams from public database and our own real-world project of autonomous vehicles. The proposed model showcases significant improvements in terms of computational complexity and number of required parameters against several renowned SANFSs, while attaining comparable and often better predictive accuracy.
Tasks	Autonomous Vehicles
Published	2018-05-11
URL	http://arxiv.org/abs/1805.04258v2
PDF	http://arxiv.org/pdf/1805.04258v2.pdf
PWC	https://paperswithcode.com/paper/palm-an-incremental-construction-of
Repo
Framework

New Vistas to study Bhartrhari: Cognitive NLP


Title	New Vistas to study Bhartrhari: Cognitive NLP
Authors	Jayashree Gajjam, Diptesh Kanojia, Malhar Kulkarni
Abstract	The Sanskrit grammatical tradition which has commenced with Panini’s Astadhyayi mostly as a Padasastra has culminated as a Vakyasastra, at the hands of Bhartrhari. The grammarian-philosopher Bhartrhari and his authoritative work ‘Vakyapadiya’ have been a matter of study for modern scholars, at least for more than 50 years, since Ashok Aklujkar submitted his Ph.D. dissertation at Harvard University. The notions of a sentence and a word as a meaningful linguistic unit in the language have been a subject matter for the discussion in many works that followed later on. While some scholars have applied philological techniques to critically establish the text of the works of Bhartrhari, some others have devoted themselves to exploring philosophical insights from them. Some others have studied his works from the point of view of modern linguistics, and psychology. Few others have tried to justify the views by logical discussions. In this paper, we present a fresh view to study Bhartrhari, and his works, especially the ‘Vakyapadiya’. This view is from the field of Natural Language Processing (NLP), more specifically, what is called as Cognitive NLP. We have studied the definitions of a sentence given by Bhartrhari at the beginning of the second chapter of ‘Vakyapadiya’. We have researched one of these definitions by conducting an experiment and following the methodology of silent-reading of Sanskrit paragraphs. We collect the Gaze-behavior data of participants and analyze it to understand the underlying comprehension procedure in the human mind and present our results. We evaluate the statistical significance of our results using T-test, and discuss the caveats of our work. We also present some general remarks on this experiment and usefulness of this method for gaining more insights in the work of Bhartrhari.
Tasks
Published	2018-10-10
URL	http://arxiv.org/abs/1810.04440v1
PDF	http://arxiv.org/pdf/1810.04440v1.pdf
PWC	https://paperswithcode.com/paper/new-vistas-to-study-bhartrhari-cognitive-nlp
Repo
Framework

Quantum cognition goes beyond-quantum: modeling the collective participant in psychological measurements


Title	Quantum cognition goes beyond-quantum: modeling the collective participant in psychological measurements
Authors	Diederik Aerts, Massimiliano Sassoli de Bianchi, Sandro Sozzo, Tomas Veloz
Abstract	In psychological measurements, two levels should be distinguished: the ‘individual level’, relative to the different participants in a given cognitive situation, and the ‘collective level’, relative to the overall statistics of their outcomes, which we propose to associate with a notion of ‘collective participant’. When the distinction between these two levels is properly formalized, it reveals why the modeling of the collective participant generally requires beyond-quantum - non-Bornian - probabilistic models, when sequential measurements at the individual level are considered, and this though a pure quantum description remains valid for single measurement situations.
Tasks
Published	2018-02-24
URL	http://arxiv.org/abs/1802.10448v1
PDF	http://arxiv.org/pdf/1802.10448v1.pdf
PWC	https://paperswithcode.com/paper/quantum-cognition-goes-beyond-quantum
Repo
Framework