COMPUTER SCIENCE AND SYSTEMS LABORATORY (LIS)
The LIS “Laboratoire d’Informatique et Systèmes” is a Joint Research Unit (UMR) under the supervision of the Centre National de la Recherche Scientifique (CNRS) attached to the Univer-sité de Toulon (UTLN) and Institut des sciences de l’information et de leurs interactions (INS2I), of the Université d’Aix-Marseille (AMU). The LIS conducts fundamental and applied research in the fields of computing, automation, signal and image.
- Director : Mustapha Ouladsine, Pr. AMU
- Deputy Director : Frédéric Bechet, Pr. AMU
- Name of local referent : Éric Busvelle, Pr. UTLN
- UTLN Administrative and Financial Service : Adoration Di Santi
Composition of the laboratory
LIS brings together more than 375 members : 190 permanent researchers and research profes-sors, more than 125 doctoral students, more than 40 post-docs and 20 technical and administra-tive staff..
In Toulon there are 24 research professors and about as many doctorants, post-docs and engi-neers.
At LIS- UTLN key areas of research include, AI for marine and maritime surveillance applica-tions, AI for large scale bioacoustics, AI for speech and hearing, and AI for robotics. LIS has es-tablished partnerships with industry technological clusters and innovation centers, as well as maintains strong ties with blue growth industries in the Region.
The members of the LIS Toulon site are divided into four teams :
- DYNI : Information DYNamics (Led by Ass Prof Ricard Marxer)
DYNI’s research in Artificial Intelligence and representation learning aims to cover the data ac-quisition, transmission and processing chain from sensors to users. The research is applied to diverse fields such as, marine and maritime robotics, bioacoustics, speech & hearing, and multi-modal information analysis in physics, health or cognition. Through its technological platform, SMIoT, DYNI innovates the scientific instrumentation for smart long-term data acquisition and embedded data processing.
DYNI tackles the main challenges of data-driven approaches applied to the experimental scienc-es : a) high cost of data acquisition, b) human bias in manual annotations, c) complexity of the underlying phenomena, and d) explicatory power of advanced machine learning models. Thus, our work focuses on unsupervised/semi-supervised learning, reinforcement learning, Deep Learning and explainable AI applied to a variety of topics ranging from environmental monitoring to the study of vocal interactivity.
- CDE : Control & Diagnosis for the Environment (led by Ass Prof Frédéric Lafont)
The group main research interests are estimation, control, diagnosis and decision support sys-tems. Two main approaches are considered. In the first one, the participants collaborate in theo-retical projects, developing methodologies for estimation, state reconstruction, control, diagnosis and monitoring. With the second approach, the research group contributes to the study of real systems coming from vehicles (underwater robots), renewable energies and agriculture sys-tems.
- SIIM : Signal and Images (directed by Prof Nadège THIRION-MOREAU)
The researches of the members of the SIIM team are gearing towards data sciences & analysis, and concern more precisely : signal and image processing, computer vision, Artificial Intelligence (machine learning, neural networks) and optimization.
They range from the observation or the physical study of the data for modeling or characteriza-tion purposes or their processing and analysis. The main aims are to retrieve meaningful infor-mation or latent variables, to help the end-users with the interpretation of the results or to develop automatic decision or diagnostic assistance systems. This involves the development of innova-tive theoretical approaches and the effective implementation of these methods through numerical computation and high performance algorithms. Their main fields of application are environmental data observation and monitoring (hyperspectral imaging, fluorescence spectroscopy imaging), environmental data analysis/mining, marine surveillance, written document analysis, biomedical engineering (diagnostic assistance)
- DIAMS : Data Integration, Analysis, and Management as Services (Led by Prof Omar Boucelma) Web site : https://diams.lis-lab.fr
DIAMS’ focus is on automating and orchestrating data processing, integration, and analysis tasks in analytics workflows of massive, multi-modal, and multi-source data.
Main research topic in Toulon is data management. Our objective is to build and develop models and tools to index, manipulate, analyze, and recommend multi-modal data from heterogeneous sources in a personalized way. We seek to represent and exploit the content of potentially mas-sive and heterogeneous data jointly with the knowledge related to the field.
Our main applications deal with :
-* Multi-modal linguistic data
-* Web intelligence data
-* Data from sensor networks (Defense and Environment)
-* Sentiment and emotions analysis for automatic recommendation of content based on text and data mining approaches (in collaboration with R2I team at LIS lab).
We are also working on data security and privacy by proposing models to define appropriate access control policies. More specifically, we are working on the security of publica-tion/subscription networks with applications using the mqtt protocol, which is one of the stand-ards of the Internet of Things. This work is carried out in collaboration with the GePaSud labora-tory of the University of French Polynesia. Applications in the field of environmental monitoring are planned.
As part of these teams, members of LIS focus their research on the following themes :
- in computer science : information dynamics, representation learning, multimodal information research (Video, Image, Speech, (Bio)Acoustics, Text), XML databases, ontology ;
- in systems : modelling and control of continuous systems, diagnosis of dynamic systems, trajectory planning, sub-Riemannian geometry, quantum control, tracking, human vision, hu-man movement, observability and filtering ;
- in image processing : low-level image processing (segmentation, edge detection), merging, demixing, stereovision ;
- in signal processing : blind processing (identification, source separation, deconvolution), time-frequency/time-scale analysis, decision/classification, optimization, tensor decompositions.
Collaborative research and teaching
The LIS laboratory participates in multidisciplinary collaborative research projects through its participation in the strategic research poles on “Sea, Environment and Sustainable Development” and the pole on “Information and Survey”
The LIS laboratory collaborates with the Mechanical and Robotics Systems Lab COSMER, and its staff teaches on the Master’s degree in Complex Systems Engineering Robotics and Con-nected Objects (ROC), and computer sciences master DID.