Publications

Accounting for Variance in Machine Learning Benchmarks
Xavier Bouthillier
Pierre Delaunay
Mirko Bronzi
Assya Trofimov
Brennan Nichyporuk
Justin Szeto
Naz Sepah
Edward Raff
Kanika Madan
Vikram Voleti
Samira Ebrahimi Kahou
Vincent Michalski
Dmitriy Serdyuk
Tal Arbel
Chris Pal
Gael Varoquaux
Pascal Vincent
Strong empirical evidence that one machine-learning algorithm A outperforms another one B ideally calls for multiple trials optimizing the l… (see more)earning pipeline over sources of variation such as data sampling, data augmentation, parameter initialization, and hyperparameters choices. This is prohibitively expensive, and corners are cut to reach conclusions. We model the whole benchmarking process, revealing that variance due to data sampling, parameter initialization and hyperparameter choice impact markedly the results. We analyze the predominant comparison methods used today in the light of this variance. We show a counter-intuitive result that adding more sources of variation to an imperfect estimator approaches better the ideal estimator at a 51 times reduction in compute cost. Building on these results, we study the error rate of detecting improvements, on five different deep-learning tasks/architectures. This study leads us to propose recommendations for performance comparisons.
2021-01-01
MLSys (published)
arxiv.org
arxiv.org
Active Learning for Capturing Human Decision Policies in a Data Frugal Context
Loïc Grossetête
Alexandre Marois
Bénédicte Chatelais
Christian Gagné
Daniel Lafond
2021-01-01
International Conference on Machine Learning, Optimization, and Data Science (published)
doi.org
ADEPT: An Adjective-Dependent Plausibility Task
Ali Emami
Ian Porada
Alexandra Olteanu
Kaheer Suleman
Adam Trischler
Jackie Cheung
2021-01-01
Annual Meeting of the Association for Computational Linguistics (published)
doi.org
Adversarial Feature Desensitization
Pouya Bashivan
Reza Bayat
Adam Ibrahim
Kartik Ahuja
Mojtaba Faramarzi
Touraj Laleh
Blake Richards
Irina Rish
Neural networks are known to be vulnerable to adversarial attacks -- slight but carefully constructed perturbations of the inputs which can … (see more)drastically impair the network's performance. Many defense methods have been proposed for improving robustness of deep networks by training them on adversarially perturbed inputs. However, these models often remain vulnerable to new types of attacks not seen during training, and even to slightly stronger versions of previously seen attacks. In this work, we propose a novel approach to adversarial robustness, which builds upon the insights from the domain adaptation field. Our method, called Adversarial Feature Desensitization (AFD), aims at learning features that are invariant towards adversarial perturbations of the inputs. This is achieved through a game where we learn features that are both predictive and robust (insensitive to adversarial attacks), i.e. cannot be used to discriminate between natural and adversarial data. Empirical results on several benchmarks demonstrate the effectiveness of the proposed approach against a wide range of attack types and attack strengths. Our code is available at https://github.com/BashivanLab/afd.
openreview.net
An Analysis of the Adaptation Speed of Causal Models
Rémi LE PRIOL
Reza Babanezhad Harikandeh
Yoshua Bengio
Simon Lacoste-Julien
2021-01-01
AISTATS (published)
proceedings.mlr.press
arxiv.org
Analyzing the Contribution of Ethical Charters to Building the Future of Artificial Intelligence Governance
Lyse Langlois
Catherine Régis
2021-01-01
Reflections on Artificial Intelligence for Humanity (published)
doi.org
Batch Reinforcement Learning Through Continuation Method
Yijie Guo
Shengyu Feng
Nicolas Le Roux
Ed Chi
Honglak Lee
Minmin Chen
Many real-world applications of reinforcement learning (RL) require the agent to learn from a fixed set of trajectories, without collecting … (see more)new interactions. Policy optimization under this setting is extremely challenging as: 1) the geometry of the objective function is hard to optimize efficiently; 2) the shift of data distributions causes high noise in the value estimation. In this work, we propose a simple yet effective policy iteration approach to batch RL using global optimization techniques known as continuation. By constraining the difference between the learned policy and the behavior policy that generates the fixed trajectories, and continuously relaxing the constraint, our method 1) helps the agent escape local optima; 2) reduces the error in policy evaluation in the optimization procedure. We present results on a variety of control tasks, game environments, and a recommendation task to empirically demonstrate the efficacy of our proposed method.
2021-01-01
ICLR (published)
openreview.net
openreview.net
Can Open Source Licenses Help Regulate Lethal Autonomous Weapons?
Cheng Lin
AJung Moon
Lethal autonomous weapon systems (LAWS, ethal autonomous weapon also known as killer robots) are a real and emerging technology that have th… (see more)e potential to radically transform warfare. Because of the myriad of moral, legal, privacy, and security risks the technology introduces, many scholars and advocates have called for a ban on the development, production, and use of fully autonomous weapons [1], [2].
2021-01-01
IEEE technology & society magazine (published)
doi.org
Capacity Expansion in the College Admission Problem
Federico Bobbio
Margarida Carvalho
Andrea Lodi
Alfredo Torrico
2021-01-01
arXiv.org (preprint)
dblp.uni-trier.de
Consistency and Rate of Convergence of Switched Least Squares System Identification for Autonomous Switched Linear Systems
Borna Sayedana
Mohammad Afshari
Peter E. Caines
Aditya Mahajan
In this paper, we investigate the problem of system identification for autonomous switched linear systems with complete state observations.… (see more) We propose switched least squares method for the identification for switched linear systems, show that this method is strongly consistent, and derive data-dependent and data-independent rates of convergence. In particular, our data-dependent rate of convergence shows that, almost surely, the system identification error is O (cid:0)(cid:112) log( T ) /T (cid:1) where T is the time horizon. These results show that our method for switched linear systems has the same rate of convergence as least squares method for non-switched linear systems. We compare our results with those in the literature. We present numerical examples to illustrate the performance of the proposed system identification method.
2021-01-01
arXiv.org (preprint)
dblp.uni-trier.de
Continual Learning via Local Module Composition
Oleksiy Ostapenko
Pau Rodriguez
Massimo Caccia
Laurent Charlin
Modularity is a compelling solution to continual learning (CL), the problem of modeling sequences of related tasks. Learning and then compos… (see more)ing modules to solve different tasks provides an abstraction to address the principal challenges of CL including catastrophic forgetting, backward and forward transfer across tasks, and sub-linear model growth. We introduce local module composition (LMC), an approach to modular CL where each module is provided a local structural component that estimates a module's relevance to the input. Dynamic module composition is performed layer-wise based on local relevance scores. We demonstrate that agnosticity to task identities (IDs) arises from (local) structural learning that is module-specific as opposed to the task- and/or model-specific as in previous works, making LMC applicable to more CL settings compared to previous works. In addition, LMC also tracks statistics about the input distribution and adds new modules when outlier samples are detected. In the first set of experiments, LMC performs favorably compared to existing methods on the recent Continual Transfer-learning Benchmark without requiring task identities. In another study, we show that the locality of structural learning allows LMC to interpolate to related but unseen tasks (OOD), as well as to compose modular networks trained independently on different task sequences into a third modular network without any fine-tuning. Finally, in search for limitations of LMC we study it on more challenging sequences of 30 and 100 tasks, demonstrating that local module selection becomes much more challenging in presence of a large number of candidate modules. In this setting best performing LMC spawns much fewer modules compared to an oracle based baseline, however, it reaches a lower overall accuracy. The codebase is available under https://github.com/oleksost/LMC.
openreview.net
Contrastive Behavioral Similarity Embeddings for Generalization in Reinforcement Learning
Rishabh Agarwal
Marlos C. Machado
Pablo Samuel Castro
Marc Gendron-Bellemare
Reinforcement learning methods trained on few environments rarely learn policies that generalize to unseen environments. To improve generali… (see more)zation, we incorporate the inherent sequential structure in reinforcement learning into the representation learning process. This approach is orthogonal to recent approaches, which rarely exploit this structure explicitly. Specifically, we introduce a theoretically motivated policy similarity metric (PSM) for measuring behavioral similarity between states. PSM assigns high similarity to states for which the optimal policies in those states as well as in future states are similar. We also present a contrastive representation learning procedure to embed any state similarity metric, which we instantiate with PSM to obtain policy similarity embeddings (PSEs). We demonstrate that PSEs improve generalization on diverse benchmarks, including LQR with spurious correlations, a jumping task from pixels, and Distracting DM Control Suite.
2021-01-01
ICLR (published)
openreview.net
openreview.net