Tanya Goyal

Tanya Goyal is an Assistant Professor in the Department of Computer Science at Cornell University. Previously, she was a postdoctoral scholar at the Princeton Language and Intelligence Center (2023-2024) and earned her Ph.D. in Computer Science from the University of Texas at Austin in 2023, advised by Greg Durrett. Her doctoral research focused on evaluation tools for text generation models and won the UTCS Bert Kay Dissertation Award. She holds a B.Tech in Mathematics and Computing from the Indian Institute of Technology, Guwahati (2011-2015).

Her research interests center on Natural Language Processing, particularly reliable evaluation frameworks for large language models (LLMs), factuality assessment, and understanding LLM behaviors influenced by training data or alignment strategies. She teaches advanced language technologies (CS 6740) and undergraduate NLP (CS 4740) at Cornell. Notable contributions include the FALTE toolkit for fine-grained annotation, the WildHallucinations evaluation framework, and studies on length optimization in RLHF and reward model dynamics.

Recent achievements include three COLM 2024 paper acceptances on topics like reward model dynamics and book summarization, a talk at Weill Cornell Medicine on generative AI, and participation in ICLR 2024. She actively contributes to open-source projects, including repositories for factuality evaluation and dependency-based entailment analysis.