Research on intelligent systems within the CDERT includes the application of the three major foundational aspects of artificial intelligence, namely logical and statistical reasoning, multi-agent systems and machine learning, in the development of novel intelligent and interactive systems. Our work in AI/ML serves to underpin all the other themes within CDERT.
Our current research activities involve the development of networks for expert agents that are capable of considering inconsistencies in the input data, providing more informed automated decisions [1]; the investigation of machine learning algorithms for identifying maritime objects (including vessels and mine-like objects) from acoustic data [2]; grounding language to vision [3, 4]; and the combination of classic control with reinforcement learning for underwater autonomous vessels [5].
Paulo E. Santos is a full member of the CNRS International Research lab CROSSING and the team co-leader of the trust New Paradigms for Autonomous Agents/Human Interaction.
[1] CORTES, H.M., SANTOS, P.E., DA SILVA FILHO, J.I., Monitoring electrical systems data-network equipment by means of Fuzzy and Paraconsistent Annotated Logic, Expert Systems with Applications, Volume 187, 2022.
[2] DOMINGOS, L.C.F., SANTOS, P.E., SKELTON, P.S.M., BRINKWORTH, R.S.A. and SAMMUT, K. "A Survey of Underwater Acoustic Data Classification Methods Using Deep Learning for Shoreline Surveillance" Sensors 22, no. 6: 2181, 2022
[3] TRAN, T., SANTOS, P.E., POWERS, D. Contrastive Visual and Language Translational Embeddings for Visual Relationship Detection. AAAI-MAKE: Machine Learning and Knowledge Engineering for Hybrid Intelligence, 2022
[4] NEAU, M., SANTOS, P.E., BOSSER, A-G., BEU, N., BUCHE, C. Commonsense Reasoning for Identifying and Understanding the Implicit Need of Help and Synthesizing Assistive Actions. AAAI-MAKE: Machine Learning and Knowledge Engineering for Hybrid Intelligence, 2022
[5] CHAFFRE, T.; SANTOS, P.E.; LE CHENADEC, G.; CHAUVEAU, E.; CLEMENT, B.; SAMMUT, K.: Learning Stochastic Adaptive Control using a Bio-Inspired Experience Replay. TechRxiv. Preprint (under submission). https://doi.org/10.36227/techrxiv.19297577.v1, 2022
