AI Speech Therapy Agent for Stuttering Assessment and Personalized Planning
A new paper introduces Virtual Speech Therapist (VST), an AI-driven platform for stuttering assessment and personalized therapy planning. VST integrates deep learning classification of stuttering types with multi-agent large language model (LLM) reasoning to generate, critique, and refine individualized therapy plans. A dedicated critic agent ensures clinical safety, methodological soundness, and adherence to professional guidelines. The system automates workflows from speech sample acquisition to evidence-based decision-making.
Key facts
- VST uses deep learning for stuttering classification.
- Multi-agent LLM reasoning generates therapy plans.
- A critic agent evaluates plans for safety and guidelines.
- System automates from speech sample to therapy plan.
Entities
—