[CogSci] Interdisciplinary seminar cycle: Benchmarking Speech Acquisition in Humans and Machines

Marvin Lavechin marvinlavechin at gmail.com
Tue Feb 15 09:08:07 PST 2022

Please join us for an interdisciplinary online seminar cycle about
Speech Acquisition in Humans and Machines*.

Recent advances in machine learning make unsupervised learning of speech
But bridging the gap between psycholinguistics and unsupervised machine
learning will take theoretical and empirical progress.
This is particularly obvious in cases where naturalistic audio that
faithfully represents infants' experiences is used for training.
Additionally, true benchmarking involves trying to directly relate machines
to humans, which is far from trivial.

The goal of the present cycle is to give an overview of the issues in
benchmarking humans and machines in terms of their acquisition of speech
perception for members of both communities and engage in discussions, as
well as potentially empirical studies, to resolve open issues.
The focus will be put exclusively on work that attempts to approach speech
perception in its full complexity, meaning using data sets as realistic as
possible (for humans and/or machines).

Kick-off meeting: March 8, time 4-6:15pm CET
Following meetings: 2nd Tue of each month, 4-5pm CET

The description of the seminar can be found here

You can subscribe to the mailing list here

Feel free to forward this email to anyone who might be interested, and
apologies for cross-posting.


Alejandrina Cristia, Marvin Lavechin
Laboratoire de Sciences Cognitives et Psycholinguistique
29, rue d'Ulm, 75005, Paris, FRANCE
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cognitivesciencesociety.org/pipermail/announcements-cognitivesciencesociety.org/attachments/20220215/558eb948/attachment-0001.html>

More information about the Announcements mailing list