Publication
Interpretable end-to-end Neurosymbolic Reinforcement Learning agents
Nils Grandien; Quentin Delfosse; Kristian Kersting
In: Computing Research Repository eprint Journal (CoRR), Vol. abs/2410.14371, Pages 1-19, arXiv, 2024.
Abstract
Deep reinforcement learning (RL) agents rely on shortcut learning, preventing them
from generalizing to slightly different environments [1 ]. To address this problem,
symbolic method, that use object-centric states, have been developed. However,
comparing these methods to deep agents is not fair, as these last operate from raw
pixel-based states. In this work, we instantiate the symbolic Successive Concept
Bottlenecks Agents (SCoBots) framework [ 2]. SCoBots decompose RL tasks into
intermediate, interpretable representations, culminating in action decisions based on
a comprehensible set of object-centric relational concepts. This architecture aids
in demystifying agent decisions. By explicitly learning to extract object-centric
representations from raw states, object-centric RL, and policy distillation via rule
extraction, this work places itself within the neurosymbolic AI paradigm, blending
the strengths of neural networks with symbolic AI. We present the first implemen-
tation of an end-to-end trained SCoBot, separately evaluate of its components, on
different Atari games. The results demonstrate the framework’s potential to create
interpretable and performing RL systems, and pave the way for future research
directions in obtaining end-to-end interpretable RL agents.