AstroAI Lunch Talks - December 1, 2025 - Mugdha Polimera
01 Dec 2025 - Joshua Wing
The video can be found here: https://www.youtube.com/watch?v=8xC1HO-lb0E
Speaker: Mugdha Polimera (CfA)
Title: AI/ML enhancements to ADS and SciX
Abstract: The NASA-funded Astrophysics Data System (ADS) has evolved from a bibliographic search engine into an intelligent, cross-disciplinary research platform powered by domain-specific AI called the Science Explorer (SciX). We will present the language models, datasets, and pipelines developed by the ADS/SciX team to bring machine learning directly into the core of our search engine research infrastructure. We will showcase our in-house language models and the tasks they enable such as named entity recognition, semantic classification, and cross-disciplinary retrieval. These new SciX services will automatically identify, separate, and enrich collections across scientific domains. We will discuss how these new AI-enabled services will improve SciX’s search capability and enhance scientific discovery.