Tag
A paper investigates the contribution of agent harnesses versus models to agent behavior, collapsing traces into a compact finite-state machine validated across twelve datasets.
This paper introduces a novel automata-based approach for control synthesis from signal temporal logic specifications using reinforcement learning, improving robustness scores and satisfaction rates over existing methods.
This technical note from 1994 describes methods for constructing deterministic and nondeterministic finite state automata in Forth, emphasizing a one-to-one mapping between definitions and state tables to avoid slow nested IFs.
This paper introduces a white-box instrument using hidden deterministic finite automata to separately measure reward success and latent-state learning in reinforcement learning agents, finding that high reward does not imply task understanding.
An introductory literate Haskell document exploring the relationship between coalgebras and automata, demonstrating how state machines can be modeled using fold and unfold operations within category theory.