“I am a simple man, I love beautiful things, including my beloved wife, the nature and mathematics. “
Hi there 👋
My name is NGUYEN Tu Anh (my pronouns are he/him/his). I am a PhD Candidate at the ALMAnaCH team (INRIA Paris) and Meta AI under the supervision of Prof. Benoit Sagot and Prof. Emmanuel Dupoux starting from 2021. I am also a visiting student at the CoML team (ENS Paris).
Before that, I graduated from the master MVA (Mathematics, Vision, Learning) at ENS Cachan (currently ENS Paris-Saclay). I was also a polytechnicien (engineering student at École polytechnique).
I have interned at La Javaness, SYSTRAN, ENS Paris and Facebook AI (currently Meta AI).
My research interests include Speech Processing, Natural Language Processing and Unsupervised Learning.
Personally, I like playing games, cooking and enjoying delicious food. I used to make some noob music for me as well as my wife. I am proud to be born in Vietnam .
If you ever wonder what is the creature on the left. It’s a leo-pig (I designed it 😛), simply because I was born in the Year of the Pig, and my Zodiac Sign is Leo.
- April 2022: We share our new TextlessNLP research milestones. 1/ We open-source our Textless Python Library. 2/ We can model and create emotional speech. 3/ We can generate chit-chat dialogues with pauses, ‘ums,’ and overlapping speech. Read more on our blog post.
- September 2021: We proudly present TextlessNLP, the first work that generates expressive speech from raw audio without using any text. Our work is tweeted by Yann LeCun, Jerome Pesenti and Mike Schroepfer. Read more on our blog post and our demo page.
- August 2021: Me and my beloved wife had an amazing wedding at Mairie de Masssy. More images can be found on our wedding website.
- April 2021: I will begin a joint PhD between Inria and Facebook, under the supervision of Prof. Benoit Sagot and Prof. Emmanuel Dupoux.
- October 2020: We release the 2021 Zero Resource Speech Challenge, which focuses on Spoken Language Modelling - the task of learning language models directly from raw audio without any annotation or text. Our metrics and baseline paper has been accepted at NeurIPS SAS 2020 Workshop.
- August 2019: SYSTRAN participated in the News Commentary Translation Task at the 6th Workshop on Asian Translation. Our translation system achieved the first rank in both directions ja-ru and ru-ja.
You can contact me through my personal email at
nguyentuanh208 [at] gmail.com.