Two papers accepted at AAAI Fall Symposium on AI Trustworthiness and Risk Assessment for Challenged Contexts, (ATRACC) on time series and explainability.

Biplav completes the first VAIBHAV Fellowship (2025-2028; 17 among 216 selected globally) visit for summer work with IIT-BHU on using AI (group recommendation) and sensors for personalized recommendation and monitoring of Yoga's Surya Namaskara (SN; sun salutation) regimen. See paper.

Kausik Lakkaraju successfully defended his dissertation proposal on July 23 - Rating AI Models for Robustness Through a Causal Lens

Major update to gaico! Enhanced support for specialized text metrics - time-series and automated planning - added for LLM output comparison and analysis [Python, Foundation Models, Library].

Biplav gave a keynote talk, Towards Sustainable Communities with Big, Open, Data and Group Recommendations Promoting User Trust' at , 'Big Data and AI (BDA) conference, Bengaluru, India on 19 July 2025.

Second patent of the group, USC 1587 - US 12,332,873, Multimodal Retrieval and Execution Monitoring Using Rich Recipe Representation, was awarded.

We released the python package, gaico, GenAI Results Comparator, to help compare, analyze and visualize outputs from Large Language Models (LLMs) [Python, Foundation Models, Library].

Three proposals get USC internal funding (1 as PI, 2 as Co-PI) to develop concepts of safe chatbots using neuro-symbolic methods with applications in medical student training and archival (library) information accessibility [Chatbots, Neuro-symbolic].