Descrizione del progetto
Diffondere l’apprendimento per rinforzo teoricamente valido
L’apprendimento per rinforzo (RL, reinforcement learning) è una branca dell’apprendimento automatico che si occupa del modo in cui gli agenti intelligenti interagiscono con ambienti sconosciuti per massimizzare le loro ricompense. La potenziale applicazione delle tecniche di RL su problemi complessi del mondo reale, come il controllo autonomo dei veicoli o le reti energetiche intelligenti, ha accresciuto l’attenzione verso questo settore. Tuttavia, gli algoritmi di RL allo stato dell’arte non sono applicabili nei settori più promettenti, soprattutto a causa della mancanza di garanzie formali di prestazione. Il progetto SCALER, finanziato dall’UE, intende affrontare questa sfida adottando un approccio di principio per sviluppare una nuova generazione di algoritmi di apprendimento per rinforzo provatamente efficienti e scalabili. La metodologia si baserà sull’identificazione di nuove proprietà strutturali dei processi decisionali di Markov su larga scala che permettono un apprendimento efficiente dal punto di vista computazionale e statistico.
Obiettivo
Reinforcement learning (RL) is an intensely studied subfield of machine learning concerned with sequential decision-making problems where a learning agent interacts with an unknown reactive environment while attempting to maximize its rewards. In recent years, RL methods have gained significant popularity due to being the key technique behind some spectacular breakthroughs of artificial intelligence (AI) research, which renewed interest in applying such techniques to challenging real-world problems like control of autonomous vehicles or smart energy grids. While the RL framework is clearly suitable to address such problems, the applicability of the current generation of RL algorithms is limited by a lack of formal performance guarantees and a very low sample efficiency. This project proposes to address this problem and advance the state of the art in RL by developing a new generation of provably efficient and scalable algorithms. Our approach is based on identifying various structural assumptions for Markov decision processes (MDPs, the main modeling tool used in RL) that enable computationally and statistically efficient learning. Specifically, we will focus on MDP structures induced by various approximation schemes including value-function approximation and relaxations of the linear-program formulation of optimal control in MDPs. Based on this view, we aim to develop a variety of new tools for designing and analyzing RL algorithms, and achieve a deep understanding of fundamental performance limits in structured MDPs. While our main focus will be on rigorous theoretical analysis of algorithms, most of our objectives are inspired by practical concerns, particularly by the question of scalability. As a result, we expect that our proposed research will have significant impact on both the theory and practice of reinforcement learning, bringing RL methods significantly closer to practical applicability.
Campo scientifico
- engineering and technologymechanical engineeringvehicle engineeringautomotive engineeringautonomous vehicles
- engineering and technologyenvironmental engineeringenergy and fuelsrenewable energy
- natural sciencescomputer and information sciencesartificial intelligencemachine learningreinforcement learning
Programma(i)
Argomento(i)
Meccanismo di finanziamento
ERC-STG - Starting GrantIstituzione ospitante
08002 Barcelona
Spagna