Publications

Speech and Speaker Recognition from Raw Waveform with SincNet
Mirco Ravanaelli
Yoshua Bengio
Deep neural networks can learn complex and abstract representations, that are progressively obtained by combining simpler ones. A recent tre… (voir plus)nd in speech and speaker recognition consists in discovering these representations starting from raw audio samples directly. Differently from standard hand-crafted features such as MFCCs or FBANK, the raw waveform can potentially help neural networks discover better and more customized representations. The high-dimensional raw inputs, however, can make training significantly more challenging. This paper summarizes our recent efforts to develop a neural architecture that efficiently processes speech from audio waveforms. In particular, we propose SincNet, a novel Convolutional Neural Network (CNN) that encourages the first layer to discover meaningful filters by exploiting parametrized sinc functions. In contrast to standard CNNs, which learn all the elements of each filter, only low and high cutoff frequencies of band-pass filters are directly learned from data. This inductive bias offers a very compact way to derive a customized front-end, that only depends on some parameters with a clear physical meaning. Our experiments, conducted on both speaker and speech recognition, show that the proposed architecture converges faster, performs better, and is more computationally efficient than standard CNNs.
2018-12-12
ArXiv (prépublication)
arxiv.org
arxiv.org
The effects of negative adaptation in Model-Agnostic Meta-Learning
Tristan Deleu
Yoshua Bengio
The capacity of meta-learning algorithms to quickly adapt to a variety of tasks, including ones they did not experience during meta-training… (voir plus), has been a key factor in the recent success of these methods on few-shot learning problems. This particular advantage of using meta-learning over standard supervised or reinforcement learning is only well founded under the assumption that the adaptation phase does improve the performance of our model on the task of interest. However, in the classical framework of meta-learning, this constraint is only mildly enforced, if not at all, and we only see an improvement on average over a distribution of tasks. In this paper, we show that the adaptation in an algorithm like MAML can significantly decrease the performance of an agent in a meta-reinforcement learning setting, even on a range of meta-training tasks.
2018-12-04
ArXiv (prépublication)
arxiv.org
arxiv.org
Recurrent transition networks for character locomotion
Félix Harvey
Christopher Pal
We present a novel approach, based on deep recurrent neural networks, to automatically generate transition animations given a past context o… (voir plus)f a few frames, a target character state and optionally local terrain information. The proposed Recurrent Transition Network (RTN) is trained without any gait, phase, contact or action labels. Our system produces realistic and fluid transitions that rival the quality of Motion Capture-based animations, even without any inverse-kinematics post-process. Our system could accelerate the creation of transition variations for large coverage or even replace transition nodes in a game's animation graph. The RTN also shows impressive results on a temporal super-resolution task.
2018-12-03
SIGGRAPH Asia 2018 Technical Briefs (publié)
doi.org
arxiv.org
Deep Learning recognizes weather and climate patterns
Karthik Kashinath
M. Prabhat
Mayur Mudigonda
Ankur Mahesh
Sookyung Kim
Yunjie Liu
S Ebrahimi Kahou
B. Toms
Evan Racah
Christopher Beckham
Christopher Pal
Tegan Maharaj
Jim Biard
K. Kunkel
Dean Nesbit Williams
Travis O'Brien
M. Wehner
W. Collins
2018-11-30
(publié)
www.semanticscholar.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… (voir plus)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.
2018-11-30
Transactions of the Association for Computational Linguistics (publié)
doi.org
Contextual Bandits for Adapting Treatment in a Mouse Model of de Novo Carcinogenesis
Audrey Durand
Charis Achilleos
Demetris Iacovides
Katerina Strati
Georgios D. Mitsis
Joelle Pineau
2018-11-28
Proceedings of the 3rd Machine Learning for Healthcare Conference (publié)
proceedings.mlr.press
proceedings.mlr.press
Understanding the impact of entropy in policy learning
Zafarali Ahmed
Nicolas Roux
Mohammad Norouzi
Dale Schuurmans
Entropy regularization is commonly used to improve policy optimization in reinforcement learning. It is believed to help with \emph{explorat… (voir plus)ion} by encouraging the selection of more stochastic policies. In this work, we analyze this claim using new visualizations of the optimization landscape based on randomly perturbing the loss function. We first show that even with access to the exact gradient, policy optimization is difficult due to the geometry of the objective function. Then, we qualitatively show that in some environments, a policy with higher entropy can make the optimization landscape smoother, thereby connecting local optima and enabling the use of larger learning rates. This paper presents new tools for understanding the optimization landscape, shows that policy entropy serves as a regularizer, and highlights the challenge of designing general-purpose policy optimization algorithms.
2018-11-26
(publié)
www.semanticscholar.org
Environments for Lifelong Reinforcement Learning
Khimya Khetarpal
Shagun Sodhani
Sarath Chandar
Doina Precup
To achieve general artificial intelligence, reinforcement learning (RL) agents should learn not only to optimize returns for one specific ta… (voir plus)sk but also to constantly build more complex skills and scaffold their knowledge about the world, without forgetting what has already been learned. In this paper, we discuss the desired characteristics of environments that can support the training and evaluation of lifelong reinforcement learning agents, review existing environments from this perspective, and propose recommendations for devising suitable environments in the future.
2018-11-25
ArXiv (prépublication)
doi.org
arxiv.org
Multi-task Learning over Graph Structures
Pengfei Liu
Jie Fu
Yue Dong
Xipeng Qiu
Jackie CK Cheung
We present two architectures for multi-task learning with neural sequence models. Our approach allows the relationships between different ta… (voir plus)sks to be learned dynamically, rather than using an ad-hoc pre-defined structure as in previous work. We adopt the idea from message-passing graph neural networks and propose a general \textbf{graph multi-task learning} framework in which different tasks can communicate with each other in an effective and interpretable way. We conduct extensive experiments in text classification and sequence labeling to evaluate our approach on multi-task learning and transfer learning. The empirical results show that our models not only outperform competitive baselines but also learn interpretable and transferable patterns across tasks.
2018-11-25
ArXiv (prépublication)
arxiv.org
arxiv.org
Planning in Dynamic Environments with Conditional Autoregressive Models
Johanna Hansen
Kyle Kastner
Aaron Courville
Gregory Dudek
We demonstrate the use of conditional autoregressive generative models (van den Oord et al., 2016a) over a discrete latent space (van den Oo… (voir plus)rd et al., 2017b) for forward planning with MCTS. In order to test this method, we introduce a new environment featuring varying difficulty levels, along with moving goals and obstacles. The combination of high-quality frame generation and classical planning approaches nearly matches true environment performance for our task, demonstrating the usefulness of this method for model-based planning in dynamic environments.
2018-11-24
ArXiv (prépublication)
arxiv.org
arxiv.org
A Survey of Mobile Computing for the Visually Impaired
Martin Weiss
Margaux Luck
Roger Girgis
Christopher Pal
Joseph Paul Cohen
The number of visually impaired or blind (VIB) people in the world is estimated at several hundred million. Based on a series of interviews … (voir plus)with the VIB and developers of assistive technology, this paper provides a survey of machine-learning based mobile applications and identifies the most relevant applications. We discuss the functionality of these apps, how they align with the needs and requirements of the VIB users, and how they can be improved with techniques such as federated learning and model compression. As a result of this study we identify promising future directions of research in mobile perception, micro-navigation, and content-summarization.
2018-11-24
ArXiv (prépublication)
arxiv.org
arxiv.org
Keep Drawing It: Iterative language-based image generation and editing
Alaaeldin El-Nouby
Shikhar Sharma
Hannes Schulz
R Devon Hjelm
Layla El Asri
S Ebrahimi Kahou
Yoshua Bengio
Graham W. Taylor
Conditional text-to-image generation approaches commonly focus on generating a single image in a single step. One practical extension beyond… (voir plus) one-step generation is an interactive system that generates an image iteratively, conditioned on ongoing linguistic input / feedback. This is significantly more challenging as such a system must understand and keep track of the ongoing context and history. In this work, we present a recurrent image generation model which takes into account both the generated output up to the current step as well as all past instructions for generation. We show that our model is able to generate the background, add new objects, apply simple transformations to existing objects, and correct previous mistakes. We believe our approach is an important step toward interactive generation.
2018-11-23
arXiv.org (prépublication)
dblp.uni-trier.de