Decoupling Turn-Taking from Semantics: A Decoupled Data Approach for Finite-State-Machine-Based Full-Duplex Dialogue
This work proposes a decoupled data approach that learns turn-taking from real Human-Human spoken dialogues while shaping semantic behavior through configurable Human-Agent text dialogues and introduces a rule-based event-guided data transformation method that serializes HH spoken dialogues into FSM tapes.
Yi-Hang Li, Chen-Hui Chu
· 1 citation