Reducing the variance in online optimization by transporting past gradients
Sébastien M. R. Arnold
Pierre-Antoine Manzagol
Reza Babanezhad Harikandeh
Ioannis Mitliagkas
Nicolas Le Roux
Most stochastic optimization methods use gradients once before discarding them. While variance reduction methods have shown that reusing pas… (see more)t gradients can be beneficial when there is a finite number of datapoints, they do not easily extend to the online setting. One issue is the staleness due to using past gradients. We propose to correct this staleness using the idea of implicit gradient transport (IGT) which transforms gradients computed at previous iterates into gradients evaluated at the current iterate without using the Hessian explicitly. In addition to reducing the variance and bias of our updates over time, IGT can be used as a drop-in replacement for the gradient estimate in a number of well-understood methods such as heavy ball or Adam. We show experimentally that it achieves state-of-the-art results on a wide range of architectures and benchmarks. Additionally, the IGT gradient estimator yields the optimal asymptotic convergence rate for online stochastic optimization in the restricted setting where the Hessians of all component functions are equal.
arxiv.org
La science, un droit humain. Mettre en œuvre le principe d’une science participative, équitable et accessible à tous
Catherine Régis
The Termination Critic
Anna Harutyunyan
Will Dabney
Diana Borsa
Nicolas Heess
Remi Munos
Doina Precup
In this work, we consider the problem of autonomously discovering behavioral abstractions, or options, for reinforcement learning agents. We… (see more) propose an algorithm that focuses on the termination function, as opposed to - as is common - the policy. The termination function is usually trained to optimize a control objective: an option ought to terminate if another has better value. We offer a different, information-theoretic perspective, and propose that terminations should focus instead on the compressibility of the option’s encoding - arguably a key reason for using abstractions.To achieve this algorithmically, we leverage the classical options framework, and learn the option transition model as a “critic” for the termination function. Using this model, we derive gradients that optimize the desired criteria. We show that the resulting options are non-trivial, intuitively meaningful, and useful for learning.
2019-01-01
AISTATS (published)
arxiv.org
arxiv.org
Toward Requirements Specification for Machine-Learned Components
Mona Rahimi
Jin Guo
Sahar Kokaly
Marsha Chechik
In current practice, the behavior of Machine-Learned Components (MLCs) is not sufficiently specified by the predefined requirements. Instead… (see more), they "learn" existing patterns from the available training data, and make predictions for unseen data when deployed. On the surface, their ability to extract patterns and to behave accordingly is specifically useful for hard-to-specify concepts in certain safety critical domains (e.g., the definition of a pedestrian in a pedestrian detection component in a vehicle). However, the lack of requirements specifications on their behaviors makes further software engineering tasks challenging for such components. This is especially concerning for tasks such as safety assessment and assurance. In this position paper, we call for more attention from the requirements engineering community on supporting the specification of requirements for MLCs in safety critical domains. Towards that end, we propose an approach to improve the process of requirements specification in which an MLC is developed and operates by explicitly specifying domain-related concepts. Our approach extracts a universally accepted benchmark for hard-to-specify concepts (e.g., "pedestrian") and can be used to identify gaps in the associated dataset and the constructed machine-learned model.
2019-01-01
2019 IEEE 27th International Requirements Engineering Conference Workshops (REW) (published)
doi.org
Unsupervised State Representation Learning in Atari
Ankesh Anand
Evan Racah
Sherjil Ozair
Yoshua Bengio
Marc-Alexandre Côté
(Rex) Devon Hjelm
State representation learning, or the ability to capture latent generative factors of an environment, is crucial for building intelligent ag… (see more)ents that can perform a wide variety of tasks. Learning such representations without supervision from rewards is a challenging open problem. We introduce a method that learns state representations by maximizing mutual information across spatially and temporally distinct features of a neural encoder of the observations. We also introduce a new benchmark based on Atari 2600 games where we evaluate representations based on how well they capture the ground truth state variables. We believe this new framework for evaluating representation learning models will be crucial for future representation learning research. Finally, we compare our technique with other state-of-the-art generative and contrastive representation learning methods. The code associated with this work is available at this https URL
arxiv.org
Updates of Equilibrium Prop Match Gradients of Backprop Through Time in an RNN with Static Input
Maxence Ernoult
Julie Grollier
Damien Querlioz
Yoshua Bengio
Benjamin Scellier
Equilibrium Propagation (EP) is a biologically inspired learning algorithm for convergent recurrent neural networks, i.e. RNNs that are fed … (see more)by a static input x and settle to a steady state. Training convergent RNNs consists in adjusting the weights until the steady state of output neurons coincides with a target y. Convergent RNNs can also be trained with the more conventional Backpropagation Through Time (BPTT) algorithm. In its original formulation EP was described in the case of real-time neuronal dynamics, which is computationally costly. In this work, we introduce a discrete-time version of EP with simplified equations and with reduced simulation time, bringing EP closer to practical machine learning tasks. We first prove theoretically, as well as numerically that the neural and weight updates of EP, computed by forward-time dynamics, are step-by-step equal to the ones obtained by BPTT, with gradients computed backward in time. The equality is strict when the transition function of the dynamics derives from a primitive function and the steady state is maintained long enough. We then show for more standard discrete-time neural network dynamics that the same property is approximately respected and we subsequently demonstrate training with EP with equivalent performance to BPTT. In particular, we define the first convolutional architecture trained with EP achieving ~ 1% test error on MNIST, which is the lowest error reported with EP. These results can guide the development of deep neural networks trained with EP.
arxiv.org
Variational Temporal Abstraction
Taesup Kim
Sungjin Ahn
Yoshua Bengio
We introduce a variational approach to learning and inference of temporally hierarchical structure and representation for sequential data. W… (see more)e propose the Variational Temporal Abstraction (VTA), a hierarchical recurrent state space model that can infer the latent temporal structure and thus perform the stochastic state transition hierarchically. We also propose to apply this model to implement the jumpy imagination ability in imagination-augmented agent-learning in order to improve the efficiency of the imagination. In experiments, we demonstrate that our proposed method can model 2D and 3D visual sequence datasets with interpretable temporal structure discovery and that its application to jumpy imagination enables more efficient agent-learning in a 3D navigation task.
arxiv.org
Wasserstein Dependency Measure for Representation Learning
Sherjil Ozair
Corey Lynch
Yoshua Bengio
Aäron van den Oord
Sergey Levine
Pierre Sermanet
Mutual information maximization has emerged as a powerful learning objective for unsupervised representation learning obtaining state-of-the… (see more)-art performance in applications such as object recognition, speech recognition, and reinforcement learning. However, such approaches are fundamentally limited since a tight lower bound of mutual information requires sample size exponential in the mutual information. This limits the applicability of these approaches for prediction tasks with high mutual information, such as in video understanding or reinforcement learning. In these settings, such techniques are prone to overfit, both in theory and in practice, and capture only a few of the relevant factors of variation. This leads to incomplete representations that are not optimal for downstream tasks. In this work, we empirically demonstrate that mutual information-based representation learning approaches do fail to learn complete representations on a number of designed and real-world tasks. To mitigate these problems we introduce the Wasserstein dependency measure, which learns more complete representations by using the Wasserstein distance instead of the KL divergence in the mutual information estimator. We show that a practical approximation to this theoretically motivated solution, constructed using Lipschitz constraint techniques from the GAN literature, achieves substantially improved results on tasks where incomplete representations are a major challenge.
arxiv.org
Building a Neural Semantic Parser from a Domain Ontology
Jianpeng Cheng
Siva Reddy
Mirella Lapata
Semantic parsing is the task of converting natural language utterances into machine interpretable meaning representations which can be execu… (see more)ted against a real-world environment such as a database. Scaling semantic parsing to arbitrary domains faces two interrelated challenges: obtaining broad coverage training data effectively and cheaply; and developing a model that generalizes to compositional utterances and complex intentions. We address these challenges with a framework which allows to elicit training data from a domain ontology and bootstrap a neural parser which recursively builds derivations of logical forms. In our framework meaning representations are described by sequences of natural language templates, where each template corresponds to a decomposed fragment of the underlying meaning representation. Although artificial, templates can be understood and paraphrased by humans to create natural utterances, resulting in parallel triples of utterances, meaning representations, and their decompositions. These allow us to train a neural semantic parser which learns to compose rules in deriving meaning representations. We crowdsource training data on six domains, covering both single-turn utterances which exhibit rich compositionality, and sequential utterances where a complex task is procedurally performed in steps. We then develop neural semantic parsers which perform such compositional tasks. In general, our approach allows to deploy neural semantic parsers quickly and cheaply from a given domain ontology.
2018-12-25
ArXiv (preprint)
arxiv.org
arxiv.org
Clustering-Oriented Representation Learning with Attractive-Repulsive Loss
Kian Kenyon-Dean
Andre Cianflone
Lucas Caccia
Guillaume Rabusseau
Jackie Cheung
Doina Precup
The standard loss function used to train neural network classifiers, categorical cross-entropy (CCE), seeks to maximize accuracy on the trai… (see more)ning data; building useful representations is not a necessary byproduct of this objective. In this work, we propose clustering-oriented representation learning (COREL) as an alternative to CCE in the context of a generalized attractive-repulsive loss framework. COREL has the consequence of building latent representations that collectively exhibit the quality of natural clustering within the latent space of the final hidden layer, according to a predefined similarity function. Despite being simple to implement, COREL variants outperform or perform equivalently to CCE in a variety of scenarios, including image and news article classification using both feed-forward and convolutional neural networks. Analysis of the latent spaces created with different similarity functions facilitates insights on the different use cases COREL variants can satisfy, where the Cosine-COREL variant makes a consistently clusterable latent space, while Gaussian-COREL consistently obtains better classification accuracy than CCE.
2018-12-18
ArXiv (preprint)
arxiv.org
arxiv.org
Learning Typed Entailment Graphs with Global Soft Constraints
Mohammad Javad Hosseini
Nathanael Chambers
Siva Reddy
Xavier R. Holt
Shay B. Cohen
Mark Johnson
Mark Steedman
This paper presents a new method for learning typed entailment graphs from text. We extract predicate-argument structures from multiple-sour… (see more)ce news corpora, and compute local distributional similarity scores to learn entailments between predicates with typed arguments (e.g., person contracted disease). Previous work has used transitivity constraints to improve local decisions, but these constraints are intractable on large graphs. We instead propose a scalable method that learns globally consistent similarity scores based on new soft constraints that consider both the structures across typed entailment graphs and inside each graph. Learning takes only a few hours to run over 100K predicates and our results show large improvements over local similarity scores on two entailment data sets. We further show improvements over paraphrases and entailments from the Paraphrase Database, and prior state-of-the-art entailment graphs. We show that the entailment graphs improve performance in a downstream task.
2018-12-01
Transactions of the Association for Computational Linguistics (published)
doi.org
Contextual Bandits for Adapting Treatment in a Mouse Model of de Novo Carcinogenesis
Audrey Durand
Charis Achilleos
Demetris C Iacovides
Katerina Strati
Georgios D. Mitsis
Joelle Pineau
In this work, we present a specific case study where we aim to design effective treatment allocation strategies and validate these using a m… (see more)ouse model of skin cancer. Collecting data for modelling treatments effectiveness on animal models is an expensive and time consuming process. Moreover, acquiring this information during the full range of disease stages is hard to achieve with a conventional random treatment allocation procedure, as poor treatments cause deterioration of subject health. We therefore aim to design an adaptive allocation strategy to improve the efficiency of data collection by allocating more samples for exploring promising treatments. We cast this application as a contextual bandit problem and introduce a simple and practical algorithm for exploration-exploitation in this framework. The work builds on a recent class of approaches for non-contextual bandits that relies on subsampling to compare treatment options using an equivalent amount of information. On the technical side, we extend the subsampling strategy to the case of bandits with context, by applying subsampling within Gaussian Process regression. On the experimental side, preliminary results using 10 mice with skin tumours suggest that the proposed approach extends by more than 50% the subjects life duration compared with baseline strategies: no treatment, random treatment allocation, and constant chemotherapeutic agent. By slowing the tumour growth rate, the adaptive procedure gathers information about treatment effectiveness on a broader range of tumour volumes, which is crucial for eventually deriving sequential pharmacological treatment strategies for cancer.
2018-11-29
Proceedings of the 3rd Machine Learning for Healthcare Conference (published)
proceedings.mlr.press