DRS automatically selects decision-relevant concepts from candidates for RL agents, supplies performance bounds, recovers manual sets, and improves test-time interventions on benchmarks and healthcare tasks.
Base features include continuous-valued positions, velocities, and relative offsets of the agent paddle, ball, and opponent paddle
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Selecting Decision-Relevant Concepts in Reinforcement Learning
DRS automatically selects decision-relevant concepts from candidates for RL agents, supplies performance bounds, recovers manual sets, and improves test-time interventions on benchmarks and healthcare tasks.