Home page > Research axes
Research at CRIL deals with the design of autonomous intelligent systems. Depending of the available information, those systems should be able to take reasonable decisions in order to achieve as much as possible their goals. Hence, those systems need some inference capabilities. The main difficulties for building such systems are diverse. First, the information is usually heterogeneous and imperfect. It contains both knowledge and beliefs about the world in which the intelligent agent lives (for instances, the physics laws of that world and some data gathered from more or less reliable sensors), about the information that other agents of that world could have, description of available actions ad their effects, agent preferences about the state of the world or the actions to be done. The information can be imperfect in many ways: incompleteness, uncertainty, incoherence, context... Furthermore, there are many kinds of inference that can be used for designing an "intelligent" behavior. Finally, such inference or decision systems are often computationally intractable in the worst case. it is thus important to identify the sources of the complexity in order to deals with them, by developing algorithms efficient in practice, or by approximation or compilation methods.
Research at CRIL is following two research axes: on the one hand "Handling of imperfect, incomplete, context-sensitive, time-sensitive and multi-source knowledge" and on the other hand "Inference and decision process".