Descripción del proyecto
De la IA basada en datos a la basada en modelos: aprendizaje representativo para la planificación
Dos de las principales líneas de investigación en inteligencia artificial (IA) giran en torno al desarrollo de aprendices basados en datos capaces de inferir comportamientos y funciones a partir de la experiencia y los datos, y de aprendices basados en modelos capaces de abordar modelos inmanejables como SAT, planificación clásica y redes bayesanas. Los aprendices y, en particular, los aprendices profundos, han logrado un éxito considerable, pero acaban siendo cajas negras inflexibles. Los solucionadores, por otro lado, necesitan modelos que son difíciles de crear a mano. El proyecto financiado con fondos europeos RLeap tiene por objeto conseguir integrarlos a ambos en el contexto de la planificación, al abordar y resolver el problema del aprendizaje de representaciones simbólicas de primer orden únicamente a partir de percepciones brutas. El proyecto puede marcar la diferencia en la forma en que se puede entender y lograr una IA general, explicable y de confianza.
Objetivo
Two of the main research threads in AI revolve around the development of data-based learners capable of inferring behavior and functions from experience and data, and model-based solvers capable of tackling well-defined but intractable models like SAT, classical planning, and Bayesian networks. Learners, and in particular deep learners, have achieved considerable success but result in black boxes that do not have the flexibility, transparency, and generality of their model-based counterparts. Solvers, on the other hand, require models which are hard to build by hand. RLeap is aimed at achieving an integration of learners and solvers in the context of planning by addressing and solving the problem of learning first-order planning representations from raw perceptions alone without using any prior symbolic knowledge. The ability to construct first-order symbolic representations and using them for expressing, communicating, achieving, and recognizing goals is a main component of human intelligence and a fundamental, open research problem in AI. The success of RLeap requires the development of radically new ideas and methods that will build on those of a number of related areas that include planning, learning, knowledge representation, combinatorial optimization and SAT. The approach to be pursued is based on a clear separation between learning the symbolic representations themselves, that is cast as a combinatorial problem, and learning the interpretations of those representations, that is cast as a supervised learning problem from targets obtained from the first part. RLeap will address both problems, not just in the planning setting but in the generalized planning setting as well where plans are general strategies. The project can make a significant difference in how general, explainable, and trustworthy AI can be understood and achieved. The PI has made key contribution to the main themes of the project that make him uniquely qualified to carry it forward.
Ámbito científico
Palabras clave
Programa(s)
Régimen de financiación
ERC-ADG - Advanced GrantInstitución de acogida
52062 Aachen
Alemania