2 papers
cs.CV2026
Target-Checked Reliability Score Refinement for Video Question Answering
Guoxiang Ren, Rohitash Chandra
Video-language models can answer multiple-choice questions with high confidence yet be wrong. We study whether answer-level reliability scores can be improved under target shift wi…
cs.CL2025
Longitudinal Abuse and Sentiment Analysis of Hollywood Movie Dialogues using Language Models
Rohitash Chandra, Guoxiang Ren, Group-H
Over the past decades, there has been an increase in the prevalence of abusive and violent content in Hollywood movies. In this study, we use language models to explore the longitu…