From Synthesis to Clinical Assistance: A Strategy-Aware Agent Framework for Autism Intervention based on Real Clinical Dataset

TL;DR

This paper presents SAagent, a strategy-aware AI framework for autism intervention that synthesizes high-fidelity dialogues, supports clinical decisions, and improves small language models using synthetic data.

Contribution

The paper introduces SAagent, a novel framework combining explicit strategy control and probabilistic behavior modeling for autism intervention, addressing data scarcity and strategy adherence issues.

Findings

Dialogue strategy distribution closely matches human therapists (KL divergence: 0.083)

Achieves nearly 80% strategic consistency with human experts in real interventions

Synthetic data enhances small language models' therapeutic capabilities

Abstract

The development of AI-assisted Early Intensive Behavioral Intervention (EIBI) for Autism Spectrum Disorder (ASD) is severely constrained by data scarcity. Furthermore, while Applied Behavior Analysis (ABA) serves as the gold standard for clinical intervention, general-purpose Large Language Models (LLMs) struggle to strictly adhere to its standardized procedures, often resulting in interactions that are linguistically fluent but strategically inconsistent. To address these challenges, we introduce \textsc{ASDAgent}, a strategy-aware framework designed to unify high-fidelity intervention dialogue synthesis and clinical decision support. \textsc{ASDAgent} incorporates two specialized components to solve distinct problems: (i) a \textsc{DoctorAgent} equipped with an Observe-Think-Act-Correct (O-T-A-C) reasoning loop, which resolves the issue of strategy collapse in LLMs by making ABA execution explicit and controllable; and (ii) a \textsc{ChildAgent} that utilizes probabilistic behavior modeling to mitigate data homogeneity, simulating diverse and non-deterministic ASD response patterns. Experiments demonstrate that dialogues generated by \textsc{ASDAgent} closely mirror the strategy distribution of human therapists (KL divergence: 0.083). In real autism intervention, \textsc{ASDAgent} achieves nearly 80\% strategic consistency with human experts. Moreover, we show that synthetic data produced by \textsc{ASDAgent} effectively distills professional clinical knowledge into small language models (SLMs), significantly enhancing their therapeutic capabilities.