INFuture2017 - Integrating ICT in Society
International audience
International audience
Deep Neural Networks (DNNs) are the current state-of-the-art models in many speech related tasks. There is a growing interest, though, for more biologically realistic , hardware friendly and energy efficient models, named Spiking Neural Networks (SNNs). Recently, it has been shown that SNNs can be trained efficiently, in a supervised manner, using ...
Published in International journal of molecular sciences
Free nerve endings are key structures in sensory transduction of noxious stimuli. In spite of this, little is known about their functional organization. Transient receptor potential (TRP) channels have emerged as key molecular identities in the sensory transduction of pain-producing stimuli, yet the vast majority of our knowledge about sensory TRP ...
International audience
Biomedical entity linking aims to map biomedical mentions,such as diseases and drugs, to standard entities in a givenknowledge base. The specific challenge in this context isthat the same biomedical entity can have a wide range ofnames, including synonyms, morphological variations, andnames with different word orderings. Recently, BERT-basedmethods...
Abstract model-checking methods are efficient to prove properties on systems with infinite state-space dimension. When considering continuoustime dynamical systems, adequate temporal logic has to be used to express specifications of systems. In this short paper, we present our current work on designing a subset of signal temporal logic (STL) amenab...
Words with the suffix-ism are reductionist terms that help us navigate complex social issues by using a simple one-word label for them. On the one hand they are often associated with political ideologies, but on the other they are present in many other domains of language, especially culture, science, and religion. This has not always been the case...
Before they even speak, infants become attuned to the sounds of the language(s) they hear, processing native phonetic contrasts more easily than non-native ones. For example, between 6-8 months and 10-12 months, infants learning American English get better at distinguishing English [ɹ] and [l], as in ‘rock’ vs ‘lock’, relative to infants learning J...
International audience
This paper reports on progress integrating the speech recognition toolkit ESPnet into Elpis, a web front-end originally designed to provide access to the Kaldi automatic speech recognition toolkit. The goal of this work is to make end-to-end speech recognition models available to language workers via a user-friendly graphical interface. Encouraging...