MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641126768 A) filed by S. Hrushikesava Raju; Dr. K. Padmanaban; and Koneru Lakshmaiah Education Foundation on October 23, 2026, for Ewtc-Insight: An Adaptive, Evidence-Weighted Trust Correction System And Method For Explainable, Evidence-Backed Multilingual Youtube Audience Intelligence.
Inventors include Dr. K. Padmanaban; Gonu Sai Varshitha; Adusumalli Keerthi Priya; Bommu Madhu Sreyan; and Doddaka Sai Teja.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: A computer-implemented system and method are disclosed for generating evidence-backed audience intelligence from multilingual YouTube comments. A pretrained multilingual transformer generates a sentiment label and confidence for each comment; confident predictions are accepted directly, while uncertain predictions are routed to evidence retrieval, in which semantically similar real comments are retrieved from a vector index. An Evidence-Weighted Trust Correction (EWTC) algorithm derives similarity weights from the median of the retrieved distances, computes weighted agreement between the prediction and the retrieved comments, derives correction strength from the model's own uncertainty, and outputs a trust-adjusted confidence without manually tuned parameters. Related topics are linked into clusters to surface root causes of complaints, and sentiment, emotion, engagement, topic, language and evidence-trust characteristics are combined into an Audience Health Score displayed on an interactive dashboard.
Disclaimer: Curated by HT Syndication.
