The LIUM (Laboratoire d'Informatique de l'Université du Mans) was established about 35 years ago. It gathers most of researchers in computer science at Le Mans University and is located on two sites: the IC2 building in Le Mans and the IUT buildings in Laval. Since its creation, LIUM's main research themes have been Technology-Enhanced Learning (TEL) and Language and Speech Technology (LST). Its current director is Sébastien George and its deputy director is Sylvain Meignier.

LIUM in numbers

50 people among which:
- 25 Academic staff
- 20 Contract researchers and PhD students
- 4 Support staff

Partnerships: researchers cooperate actively with other French, European and international laboratories. LIUM has also formed industrial partnerships with Airbus, Orange Lab, Facebook, Systran, Allo-media, OpenClassrooms, 44screens.
- Member of Atlanstic 2020 (grouping of digital science players in the Pays de la Loire region)
- Member of the Claude Chappe Institute

- CPU and GPU computing cluster (1900 CPU cores, 90 GPUs)
- Experimental room for research work in IT environment for Human Learning (audio and video recording)
- Software and corpus: SideKit/s4d, TEDLium, nmtpy, TurtleTable(t), Albiziapp, TGRIS, EscapeIt


Production, protection and distribution of software and resources

LIUM produces a remarkable number of softwares related to its research activities. This includes the TurtleTable(t) environment, which is a IT environment for Human Learning dedicated to the learning of the basics of algorithmic distributed on the playstore, the creation of an integrated development environment dedicated to UTL (Usage Tracking Language), Sidekit/s4d toolkits for speaker recognition, and linguistic resources such as the TED-LIUM corpus. Most of these tools are distributed under open license.

Excellent results in international evaluation campaigns

LIUM regularly takes part in international evaluation campaigns in speech and speaker recognition and machine translation. The results rank LIUM among the academic labs that best master these technologies worldwide.

Some results:

  • Speech transcription: 1st at REPERE 2013 (French), 1st to IWSLT 2014 (Italian) and 2015 (German), 2nd at MGB 2015 (English) and 2017 (Arabic)
  • Text / speech translation: 1st multimodal translation at IWSLT 2017, 2022
  • Speaker segmentation and clustering: 2nd at MGB 2016, 1st at MediaEval 2015 and 2016


The TEL team owns a research platform in Laval (financed by the Laval agglomeration and FEDER funds) with a surface area of 146m2 including an experimental room of 53m2. This room can accommodate 18 learners and one teacher. The adjoining office is used to simulate distance learning and make observations of instrumented pedagogical situations.