DELTA: Deliberative Multi-Agent Reasoning with Reinforcement Learning for Multimodal Psychological Counseling

TL;DR

DELTA is a multi-agent reinforcement learning framework that enhances multimodal psychological counseling by explicitly reasoning over verbal, visual, and vocal cues to improve empathy and response quality.

Contribution

It introduces a structured reasoning process over multimodal signals and employs reinforcement learning guided by emotion attunement to advance counseling AI.

Findings

DELTA outperforms existing models in counseling quality and emotion attunement.

Explicit multimodal reasoning improves empathic responses.

Structured mental state representations complement multimodal cues.

Abstract

Psychological counseling is a fundamentally multimodal cognitive process in which clinicians integrate verbal content with visual and vocal cues to infer clients' mental states and respond empathically. However, most existing language-model-based counseling systems operate on text alone and rely on implicit mental state inference. We introduce DELTA, a deliberative multi-agent framework that models counseling as a structured reasoning process over multimodal signals, separating evidence grounding, mental state abstraction, and response generation. DELTA further incorporates reinforcement learning guided by a distribution-level Emotion Attunement Score to encourage emotionally attuned responses. Experiments on a multimodal counseling benchmark show that DELTA improves both counseling quality and emotion attunement across models. Ablation and qualitative analyses suggest that explicit multimodal reasoning and structured mental state representations play complementary roles in supporting empathic human-AI interaction.