REaDY team seminar

10.30-11.10: Chotiga Pattamadilok

From lip- to script-reading: An integrative view of Audio-Visual Associations in language processing (AVA)

During the talk, I will present the general idea of our new ANR project that proposes to explore the relationships between the two main forms of audio-visual association in language processing, i.e., the associations between speech and articulatory gestures and between speech and orthography. Given their distinct properties, these natural and artificial audio-visual associations have been considered as two cognitive processes that are explained by different theoretical models. The present proposal adopts a novel perspective that seeks to establish the missing link between them. The aim is to elaborate a unified framework explaining how different inputs jointly contribute to forming coherent language representations. A new study that we conducted to address this issue will be presented.

11.10-11.50: Amie Fairs

Can we successfully carry out speech production experiments online?

In this age of COVID, more and more psychological experiments need to be carried out online so that data can still be collected. While much research has shown that typical language comprehension studies, such as lexical decision, can be carried out online, to our knowledge there are no online language production studies. Anecdotally, many language production researchers are skeptical about whether online production data are reliable. We sought in this experiment to carry out a typical production study – picture naming – online, and to determine a) whether we could replicate the well-known production effect of word frequency, b) whether the response patterns were similar to a lab-based experiment, and c) whether online-related parameters, such as internet speed, would have an effect on response times or errors. Preliminary data analysis suggests that we can replicate the word frequency effect, yet the distributions of responses and amount of errors are different to lab based experiments. While this analysis is preliminary, this suggests that online production studies are valuable and find similar sized effects to the lab. In addition, in the course of testing this experiment we have learnt a lot of practical information useful for online production studies, which I will discuss.

Seminar of Francesca Carbone

What is the impact of prosody and semantics on emotion processing?

Francesca Carbone

Seminar Anna Marczyk & Ben O’Brien

How Spectrotemporal Modulations can be used in phonetic research?

Anna Marczyk & Ben O’Brien

Seminar Laurent Besacier (LIG)

An introduction to computational language documentation: example of the BULB project (presentation in French)
By Laurent Besacier of the LIG (IT Lab Grenoble)

Abstract: In this talk, I will introduce computational language documentation which is an emerging interdisciplinary field where linguistics leverage machine learning in order to help language documentation. First contributions on this topic were done during the BULB* project which addressed oral Bantu languages. I will present some insights of BULB in this talk: (1) how to ease or speed up speech data collection using mobile apps (2) how to leverage neural sequence-to-sequence models for automatic data processing and analysis.

*BULB: Breaking the Unwritten Language Barrier

Seminar of Clément François

Seminar by

Clément François (LPL)

Cortical and subcortical processing of baby speech sounds

Within the next meeting of the REaDY research team

Seminar Thibault Cattelain

On the occasion of its next meeting on Monday, October 7th, the POP team will welcome Thibault Cattelain from Gipsa-Lab for a conference starting at 4pm.

This presentation is open to all !

 
Thibault Cattelain

(Gipsa-Lab, Grenoble)
A new technique to analyze electromyographic signals during an emotional perception task

 

Seminar Pierre Lison

Meeting “Dialogue and NLP”

Modélisation du dialogue: contrôle du dialogue et corpus multilingues (seminar held in French)
Pierre Lison (Norwegian Computing Center)

Wednesday May 22, 10:00
Laboratoire Parole et Langage

In the framework of PHC Aurora.

Contact at the lab: Laurent Prévot

Seminar Lina Maria Rojas Barahona

A glance through Conversational agents

Wednesday May 22 11:30

Lina Maria Rojas Barahona, Orange Labs, Lannion

Conversational agents are once and again gaining great interest from academics and industrials. The availability of big data as well as the advances in processing units have made deep learning approaches feasible and promising, reviving the dream of creating artificial agents that can easily converse to people. Several solutions have been proposed since the first psychoanalyst chatbot (1966). From regular expressions and symbolic approaches (formal grammars and formal logics) to statistical approaches (probabilistic models and data-driven techniques, e.g. machine learning or deep learning). We have already obtained promising results. For instance, we know that machines can learn optimal strategies for simple tasks in small domains (task-oriented dialogue systems). Moreover, we are treating open domain dialogues by asking questions to online encyclopedias such as Wikipedia (conversational reading comprehension) and we are able to predict the best answer in chitchats (end to end neural approaches). In this talk I will review briefly the state of the art and highlight the open research problems on conversational agents.

Contact at the lab: Noël Nguyen