Université de Montréal

About: Université de Montréal is a education organization based out in Montreal, Quebec, Canada. It is known for research contribution in the topics: Population & Context (language use). The organization has 45641 authors who have published 100476 publications receiving 4004007 citations. The organization is also known as: University of Montreal & UdeM.

Representational power of restricted boltzmann machines and deep belief networks

Abstract: Deep belief networks (DBN) are generative neural network models with many layers of hidden explanatory factors, recently introduced by Hinton, Osindero, and Teh (2006) along with a greedy layer-wise unsupervised learning algorithm. The building block of a DBN is a probabilistic model called a restricted Boltzmann machine (RBM), used to represent one layer of the model. Restricted Boltzmann machines are interesting because inference is easy in them and because they have been successfully used as building blocks for training deeper models. We first prove that adding hidden units yields strictly improved modeling power, while a second theorem shows that RBMs are universal approximators of discrete distributions. We then study the question of whether DBNs with more layers are strictly more powerful in terms of representational power. This suggests a new and less greedy criterion for training RBMs within DBNs.

Individual Differences in Rhythmic Cortical Entrainment Correlate with Predictive Behavior in Sensorimotor Synchronization

Abstract: The current study aims at characterizing the mechanisms that allow humans to entrain the mind and body to incoming rhythmic sensory inputs in real time. We addressed this unresolved issue by examining the relationship between covert neural processes and overt behavior in the context of musical rhythm. We measured temporal prediction abilities, sensorimotor synchronization accuracy and neural entrainment to auditory rhythms as captured using an EEG frequency-tagging approach. Importantly, movement synchronization accuracy with a rhythmic beat could be explained by the amplitude of neural activity selectively locked with the beat period when listening to the rhythmic inputs. Furthermore, stronger endogenous neural entrainment at the beat frequency was associated with superior temporal prediction abilities. Together, these results reveal a direct link between cortical and behavioral measures of rhythmic entrainment, thus providing evidence that frequency-tagged brain activity has functional relevance for beat perception and synchronization.

The Ubuntu Dialogue Corpus: A Large Dataset for Research in Unstructured Multi-Turn Dialogue Systems

Abstract: This paper introduces the Ubuntu Dialogue Corpus, a dataset containing almost 1 million multi-turn dialogues, with a total of over 7 million utterances and 100 million words. This provides a unique resource for research into building dialogue managers based on neural language models that can make use of large amounts of unlabeled data. The dataset has both the multi-turn property of conversations in the Dialog State Tracking Challenge datasets, and the unstructured nature of interactions from microblog services such as Twitter. We also describe two neural learning architectures suitable for analyzing this dataset, and provide benchmark performance on the task of selecting the best next response.

The MC-Fold and MC-Sym pipeline infers RNA structure from sequence data

Abstract: The classical RNA secondary structure model considers A.U and G.C Watson-Crick as well as G.U wobble base pairs. Here we substitute it for a new one, in which sets of nucleotide cyclic motifs define RNA structures. This model allows us to unify all base pairing energetic contributions in an effective scoring function to tackle the problem of RNA folding. We show how pipelining two computer algorithms based on nucleotide cyclic motifs, MC-Fold and MC-Sym, reproduces a series of experimentally determined RNA three-dimensional structures from the sequence. This demonstrates how crucial the consideration of all base-pairing interactions is in filling the gap between sequence and structure. We use the pipeline to define rules of precursor microRNA folding in double helices, despite the presence of a number of presumed mismatches and bulges, and to propose a new model of the human immunodeficiency virus-1 -1 frame-shifting element.

The BioGRID interaction database: 2015 update

Abstract: The Biological General Repository for Interaction Datasets (BioGRID: http://thebiogrid.org) is an open access database that houses genetic and protein interactions curated from the primary biomedical literature for all major model organism species and humans. As of September 2014, the BioGRID contains 749 912 interactions as drawn from 43 149 publications that represent 30 model organisms. This interaction count represents a 50% increase compared to our previous 2013 BioGRID update. BioGRID data are freely distributed through partner model organism databases and meta-databases and are directly downloadable in a variety of formats. In addition to general curation of the published literature for the major model species, BioGRID undertakes themed curation projects in areas of particular relevance for biomedical sciences, such as the ubiquitin-proteasome system and various human disease-associated interaction networks. BioGRID curation is coordinated through an Interaction Management System (IMS) that facilitates the compilation interaction records through structured evidence codes, phenotype ontologies, and gene annotation. The BioGRID architecture has been improved in order to support a broader range of interaction and post-translational modification types, to allow the representation of more complex multi-gene/protein interactions, to account for cellular phenotypes through structured ontologies, to expedite curation through semi-automated text-mining approaches, and to enhance curation quality control.