A Proactive Multi-Agent Dialogue Framework for Assessing Social Language Disorder Traits in Autism
Researchers have developed a novel multi-agent dialogue framework called TPA (Think, Plan, Ask) designed to proactively assess traits of Social Language Disorder (SLD) in individuals with autism. This AI system simulates a doctor agent that reasons about unobserved diagnostic traits and formulates targeted questions to elicit them from a simulated patient agent. In evaluations using real clinical data, TPA demonstrated significantly higher diagnostic efficiency and trait coverage compared to existing dialogue planning methods and even automated replays of clinician-led assessments. AI
IMPACT This AI framework could lead to more scalable and efficient AI-assisted clinical screening for autism-related language disorders.