Découvrez le dernier rapport d'impact de Mila, qui met en lumière les réalisations exceptionnelles des membres de notre communauté au cours de la dernière année.
Rapport et guide politique GPAI: Vers une réelle égalité en IA
Rejoignez-nous à Mila le 26 novembre pour le lancement du rapport et du guide politique qui présente des recommandations concrètes pour construire des écosystèmes d'IA inclusifs.
Nous utilisons des témoins pour analyser le trafic et l’utilisation de notre site web, afin de personnaliser votre expérience. Vous pouvez désactiver ces technologies à tout moment, mais cela peut restreindre certaines fonctionnalités du site. Consultez notre Politique de protection de la vie privée pour en savoir plus.
Paramètre des cookies
Vous pouvez activer et désactiver les types de cookies que vous souhaitez accepter. Cependant certains choix que vous ferez pourraient affecter les services proposés sur nos sites (ex : suggestions, annonces personnalisées, etc.).
Cookies essentiels
Ces cookies sont nécessaires au fonctionnement du site et ne peuvent être désactivés. (Toujours actif)
Cookies analyse
Acceptez-vous l'utilisation de cookies pour mesurer l'audience de nos sites ?
Multimedia Player
Acceptez-vous l'utilisation de cookies pour afficher et vous permettre de regarder les contenus vidéo hébergés par nos partenaires (YouTube, etc.) ?
Publications
Veille sur les outils numériques en santé dans le contexte de COVID-19
Syntax is fundamental to our thinking about language. Although neural networks are very successful in many tasks, they do not explicitly mod… (voir plus)el syntactic structure. Failing to capture the structure of inputs could lead to generalization problems and over-parametrization. In the present work, we propose a new syntax-aware language model: Syntactic Ordered Memory (SOM). The model explicitly models the structure with a one-step look-ahead parser and maintains the conditional probability setting of the standard language model. Experiments show that SOM can achieve strong results in language modeling and syntactic generalization tests, while using fewer parameters then other models.
Modeling joint probability distributions over sequences has been studied from many perspectives. The physics community developed matrix prod… (voir plus)uct states, a tensor-train decomposition for probabilistic modeling, motivated by the need to tractably model many-body systems. But similar models have also been studied in the stochastic processes and weighted automata literature, with little work on how these bodies of work relate to each other. We address this gap by showing how stationary or uniform versions of popular quantum tensor network models have equivalent representations in the stochastic processes and weighted automata literature, in the limit of infinitely long sequences. We demonstrate several equivalence results between models used in these three communities: (i) uniform variants of matrix product states, Born machines and locally purified states from the quantum tensor networks literature, (ii) predictive state representations, hidden Markov models, norm-observable operator models and hidden quantum Markov models from the stochastic process literature,and (iii) stochastic weighted automata, probabilistic automata and quadratic automata from the formal languages literature. Such connections may open the door for results and methods developed in one area to be applied in another.
Feed-forward neural networks consist of a sequence of layers, in which each layer performs some processing on the information from the previ… (voir plus)ous layer. A downside to this approach is that each layer (or module, as multiple modules can operate in parallel) is tasked with processing the entire hidden state, rather than a particular part of the state which is most relevant for that module. Methods which only operate on a small number of input variables are an essential part of most programming languages, and they allow for improved modularity and code re-usability. Our proposed method, Neural Function Modules (NFM), aims to introduce the same structural capability into deep learning. Most of the work in the context of feed-forward networks combining top-down and bottom-up feedback is limited to classification problems. The key contribution of our work is to combine attention, sparsity, top-down and bottom-up feedback, in a flexible algorithm which, as we show, improves the results in standard classification, out-of-domain generalization, generative modeling, and learning representations in the context of reinforcement learning.
Abstract The bulk of social neuroscience takes a ‘stimulus-brain’ approach, typically comparing brain responses to different types of so… (voir plus)cial stimuli, but most of the time in the absence of direct social interaction. Over the last two decades, a growing number of researchers have adopted a ‘brain-to-brain’ approach, exploring similarities between brain patterns across participants as a novel way to gain insight into the social brain. This methodological shift has facilitated the introduction of naturalistic social stimuli into the study design (e.g. movies) and, crucially, has spurred the development of new tools to directly study social interaction, both in controlled experimental settings and in more ecologically valid environments. Specifically, ‘hyperscanning’ setups, which allow the simultaneous recording of brain activity from two or more individuals during social tasks, has gained popularity in recent years. However, currently, there is no agreed-upon approach to carry out such ‘inter-brain connectivity analysis’, resulting in a scattered landscape of analysis techniques. To accommodate a growing demand to standardize analysis approaches in this fast-growing research field, we have developed Hyperscanning Python Pipeline, a comprehensive and easy open-source software package that allows (social) neuroscientists to carry-out and to interpret inter-brain connectivity analyses.
2020-10-08
Social Cognitive and Affective Neuroscience (publié)
Continual learning (CL) is a setting in which an agent has to learn from an incoming stream of data during its entire lifetime. Although maj… (voir plus)or advances have been made in the field, one recurring problem which remains unsolved is that of Catastrophic Forgetting (CF). While the issue has been extensively studied empirically, little attention has been paid from a theoretical angle. In this paper, we show that the impact of CF increases as two tasks increasingly align. We introduce a measure of task similarity called the NTK overlap matrix which is at the core of CF. We analyze common projected gradient algorithms and demonstrate how they mitigate forgetting. Then, we propose a variant of Orthogonal Gradient Descent (OGD) which leverages structure of the data through Principal Component Analysis (PCA). Experiments support our theoretical findings and show how our method reduces CF on classical CL datasets.
The coronavirus disease 2019 (COVID-19) pandemic has quickly become a global public health crisis unseen in recent years. It is known that t… (voir plus)he structure of the human contact network plays an important role in the spread of transmissible diseases. In this work, we study a structure aware model of COVID-19 CGEM. This model becomes similar to the classical compartment-based models in epidemiology if we assume the contact network is a Erdos-Renyi (ER) graph, i.e. everyone comes into contact with everyone else with the same probability. In contrast, CGEM is more expressive and allows for plugging in the actual contact networks, or more realistic proxies for it. Moreover, CGEM enables more precise modelling of enforcing and releasing different non-pharmaceutical intervention (NPI) strategies. Through a set of extensive experiments, we demonstrate significant differences between the epidemic curves when assuming different underlying structures. More specifically we demonstrate that the compartment-based models are overestimating the spread of the infection by a factor of 3, and under some realistic assumptions on the compliance factor, underestimating the effectiveness of some of NPIs, mischaracterizing others (e.g. predicting a later peak), and underestimating the scale of the second peak after reopening.
Most of today's popular deep architectures are hand-engineered for general purpose applications. However, this design procedure usually lead… (voir plus)s to massive redundant, useless, or even harmful features for specific tasks. Such unnecessarily high complexities render deep nets impractical for many real-world applications, especially those without powerful GPU support. In this paper, we attempt to derive task-dependent compact models from a deep discriminant analysis perspective. We propose an iterative and proactive approach for classification tasks which alternates between (1) a pushing step, with an objective to simultaneously maximize class separation, penalize co-variances, and push deep discriminants into alignment with a compact set of neurons, and (2) a pruning step, which discards less useful or even interfering neurons. Deconvolution is adopted to reverse `unimportant' filters' effects and recover useful contributing sources. A simple network growing strategy based on the basic Inception module is proposed for challenging tasks requiring larger capacity than what the base net can offer. Experiments on the MNIST, CIFAR10, and ImageNet datasets demonstrate our approach's efficacy. On ImageNet, by pushing and pruning our grown Inception-88 model, we achieve better-performing models than smaller deep Inception nets grown, residual nets, and famous compact nets at similar sizes. We also show that our grown deep Inception nets (without hard-coded dimension alignment) can beat residual nets of similar complexities.