MUMBAI, India, July 13 -- Intellectual Property India has published a patent application (202621069900 A) filed by Symbiosis International Deemed University on June 03, 2026, for Attention-Based Deep Learning System For Explainable Autism Spectrum Disorder Risk Prediction.
Inventors include Dr. Snehlata Wankhade; Reshal Age; Ritika Bhoyar; and Gunn Fulwani.
The application for the patent was published on July 10, 2026, under issue no. 28/2026.
Abstract: ABSTRACT ATTENTION-BASED DEEP LEARNING SYSTEM FOR EXPLAINABLE AUTISM SPECTRUM DISORDER RISK PREDICTION The present invention provides an attention-based deep learning system (100) for explainable and fair prediction of Autism Spectrum Disorder (ASD) risk levels. The system (100) comprises a data acquisition and preprocessing module (110) for merging and cleaning structured behavioral screening datasets, a feature engineering module (130) for computing composite behavioral scores and interaction features from AQ-10 questionnaire responses, a risk level assignment module (140) for multi-level risk categorization, and an ASDTransformerModel (150) incorporating an embedding layer (151), an attention-gating mechanism (152), a transformer encoder (153) with multi-head self-attention, and a softmax output classification layer (154). The system (100) further comprises an explainability module (160) employing SHAPley Additive Explanations for transparent feature importance analysis and a fairness evaluation module (170) utilizing the Equal Opportunity metric for demographic equity assessment. An interactive prediction interface (180) enables real-time clinical risk assessment. The system achieves 96 percent accuracy with equitable performance across demographic groups. [
Disclaimer: Curated by HT Syndication.
