Laboratory for Artificial Intelligence and Semantic Data (LIASD)

Affiliated Research areas

• Web and information systems, data management systems, information retrieval and digital libraries, data fusion
• Machine learning, statistical data processing and applications using signal processing (e.g. speech, image, video)
• Artificial intelligence, intelligent systems, natural language processing
• Computer graphics, computer vision, multimedia, computer games
• Human computer interaction and interface, visualisation

Scientific Areas

• Computer science
• Artificial intelligence
• Ubiquitous Computing and Data Science (IUSD)
• Smart Data Spaces (EID)
• Programming, AI, Security, Texts, Image, Simulation (PASTIS)

Keywords

• Artificial Intelligence
• modeling and simulation
• data semantics
• data science

Summary

The LIASD brings together researchers and PhD candidates in computer science. The lab is composed of three research groups named EID, IUSD and PASTIS, respectively.

The EID research group stands for Espaces Intelligents des Données (Smart Data Spaces) and specialises in massive data engineering. Its theme draws on the skills and experience of the group members, centred on data and knowledge. Their research is divided between deductive systems, producing knowledge by reasoning from symbolic models, and inductive systems, producing knowledge from massive data, by statistical and neural learning. In this context, the group promotes the emergence of hybrid models and systems, integrating induction into deductive systems and vice versa, to varying degrees, which also includes the control of personal data, now a fundamental information issue. Their projects are organised along multiple directions: deductive reasoning on knowledge and data; induction learning on massive data; dynamic multimodal description models; multilingual semantic resources; or formal approach to personal data.

The IUSD research group stands for Ubiquitous Computing and Data Science (in French). The group's research focuses on the development of algorithms and methodologies for human-centred or human-in-the-loop applications. The different tackled themes are biometrics, brain-machine interfaces and smart/communicating connected objects. In the case of biometrics, the group is working on topical issues to offer solutions in fields such as forensics, identification and authentication. On the topic of brain-machine interfaces, members of the group are developing algorithms for interpreting thought to assist disabled people and algorithms for classifying brain activity in an attempt to understand its behaviour. As far as connected objects are concerned, the group is working on methodologies to use them in various applications such as smart cities and buildings, home automation and e-health.

The PASTIS research group's name stands for Programming, Artificial intelligence, Security, Texts, Images, Simulations. This reflects the diversity of subjects the team members involved in the group work on. The group focuses on understanding and modelling cognitive human activities using computers. The group's philosophy is to rely heavily on programming, typically through simulations and implementation of concrete and reusable tools. Members of PASTIS implement this common approach in various fields of computer science: they work on creating training datasets for NLP models (GLAÇON), solving combinatorial problems and games, expressive rendering, educational computer graphics, trust modelling in non-classical logics, emancipatory security and privacy.