Leader(s)

Christophe Cerisara

Yannick Parmentier

Mailing list

In order to get information about the WG activities, you can subscribe to its mailing list at https://sympa.inria.fr/sympa/info/mosaiknlp.

Description

Members of the Natural Language Processing & Large Language Models (NLP & LLM) Working Group share a common interest in textual data. Their activities belong to the following axes :

  1. research work focussing either on LLMs themselves, or on the use of LLMs in the context of NLP applications ;
  2. research work focussing on multimodal deep learning models ;
  3. research work focussing on (deep and non-deep) Machine Learning techniques applied on texts ;
  4. research work focussing on NLP applications.

Current activities include LLM compression, (multilingual) text generation, merging of knowledge-based information and LLMs in the context of hybrid AI, federating learning techniques, and NLP applications such as literature analysis or Computer-Assisted Language Learning.

The main objectives of this WG is to provide researchers with a space to facilitate cooperation on the topics mentioned above, e.g. by sharing expertise and / or code.

Members

  • Abdelnur, Alejandra
  • Baki, Khadidja-wissal
  • Bannay, Aurelie
  • Blansché, Alexandre
  • Buet, Francois
  • Castagnos, Sylvain
  • Cerisara, Christophe
  • Cherif, Mohamed
  • Couceiro, Miguel
  • Cruz-Lara, Samuel
  • Daquin, Mathieu
  • Gardent, Claire
  • Gendron, Barbara
  • Ghebriout, Mohamed Imed Eddine
  • Guibon, Gael
  • Haurel, Maxime
  • Langlois, David
  • Lauer, Fabien
  • Lauzzana, Gabriel
  • Legrand, Joel
  • Mecharnia, Thamer
  • Michel, Gaspard
  • Nikiforovskaya, Anna
  • Parmentier, Yannick
  • Pennerath, Frederic
  • Prouteau, Thibault
  • Roussanaly, Azim
  • Sam, Irene-mary
  • Song, Yifei
  • Sow, Fatou
  • Sy, Yaya
  • Waffo-Dzuyo, Guy-Stéphane
  • Worm, Alexanne
  • Zheng, Estelle

Organisation

We propose to organise our works within “focal groups”, which are even totally bottom-up and free to “create”, self-organize and disappear. Focal groups are typically composed of 2 to 6 members who have decided to work or study together some precise topic related to LLM.

   Team              Work groups           Focal groups     Topic

┌────────┐             ┌─────┐              ┌───────┐
│ Mosaik ├─────┬───────┤ NLP ├───────┬──────┤ LLM42 │    growing/compression of LLMs
└────────┘     │       └─────┘       │      └───────┘
               │                     │      ┌───────┐
               │                     └──────┤  ...  │
               │                            └───────┘
               │       ┌────────┐
               ├───────┤ Frugal │
               │       └────────┘
               │
               │
               │         ┌─────┐
               └─────────┤ ... │
                         └─────┘
Back to top