morgan
@lgrobol.eu
The clown formerly known as Loïc Grobol. Associate professor of #ComputationalLinguistics at Université Paris Nanterre. #nlp for #MinorityLanguages and stuff 🏳️⚧️ 🏳️🌈
they/gender neutral
they/gender neutral
on présente notre projet ce soir! viendez!
Aujourd'hui ! Fenoz e Lannuon!
Demain à Lannion, à 20h à l'Espace Sainte-Anne rue de Kerampont, @melaniejouitteau.bsky.social (chercheuse en linguistique au CNRS) présentera la conférence « L’IA au service du breton : projet YAR pour Yezh Ar Vro ». #brezhoneg #IA www.letelegramme.fr/agenda/debat...
October 7, 2025 at 11:54 AM
on présente notre projet ce soir! viendez!
don't mind this I'm just trying to see something
September 20, 2025 at 10:06 AM
don't mind this I'm just trying to see something
currently in the virtual #CMCL2025 poster session with @virmalised.us and a sousaphone
May 3, 2025 at 2:50 PM
currently in the virtual #CMCL2025 poster session with @virmalised.us and a sousaphone
also incidentally my first paper under my actual name. I'll probably keep using both, but it seemed appropriate for this one
. @lgrobol.eu and I finally got our pronouns/neopronouns modeling paper up on arxiv ahead of NAACL! We conceptualize pronoun as speaker priors on the flexibility of pronoun use depending on community norms. Let us know your thoughts! arxiv.org/abs/2504.02973
A Bayesian account of pronoun and neopronoun acquisition
A major challenge to equity among members of queer communities is the use of one's chosen forms of reference, such as personal names or pronouns. Speakers often dismiss their misuses of pronouns as "u...
arxiv.org
April 10, 2025 at 2:57 PM
also incidentally my first paper under my actual name. I'll probably keep using both, but it seemed appropriate for this one
I'm really really happy about this one
yayyyyy we're going to @queerinai.com at #NAACL2025! stay tuned for the camera ready describing our Bayesian pronoun/neopronoun model 👀
March 8, 2025 at 12:44 AM
I'm really really happy about this one
Helping on this one was really cool, BGMMs are so good!
Welp! We responded to reviews, so the paper's prettier! Check out some very interesting findings showing non-linear dynamics in the lexicon that influence response production in the cloze task. Featuring a very clean (imo) modeling framework -- mixture models -- for semantics! osf.io/preprints/ps...
January 3, 2025 at 9:13 AM
Helping on this one was really cool, BGMMs are so good!
This is a great thread that you should definitely read. I wrote a little something on diversity-washing this spring.
In French, about how fake Breton-French machine translation systems are actually harmful
In French, about how fake Breton-French machine translation systems are actually harmful
December 7, 2024 at 9:41 AM
This is a great thread that you should definitely read. I wrote a little something on diversity-washing this spring.
In French, about how fake Breton-French machine translation systems are actually harmful
In French, about how fake Breton-French machine translation systems are actually harmful
hej i don't anticipate posting a lot here but uuuh this account exists i guess
November 10, 2024 at 1:06 AM
hej i don't anticipate posting a lot here but uuuh this account exists i guess
Reposted by morgan
-- email template to be sent to legal@wikimedia.org and copy to ca@wikimedia.org (trust&safety) --
Subject: Transphobic rules on French speaking wikipedia
To the attention of the Legal Department of the Wikimedia Foundation,
Subject: Transphobic rules on French speaking wikipedia
To the attention of the Legal Department of the Wikimedia Foundation,
March 4, 2024 at 11:33 PM
-- email template to be sent to legal@wikimedia.org and copy to ca@wikimedia.org (trust&safety) --
Subject: Transphobic rules on French speaking wikipedia
To the attention of the Legal Department of the Wikimedia Foundation,
Subject: Transphobic rules on French speaking wikipedia
To the attention of the Legal Department of the Wikimedia Foundation,