Nika Haghtalab

2023 Early Career Fellow

Nika Haghtalab is an Associate Professor in the Department of Electrical Engineering and Computer Sciences at UC Berkeley. She works on a broad range of core methodological questions in machine learning, with an emphasis on principled and often mathematical approaches. Her work contributes to an emerging foundation for machine learning and decision-making systems that explicitly accounts for the environments and contexts in which they operate, including questions of robustness and reliability, distribution shift and transfer learning, human–AI interaction and learning from human or agentic behavior, principled approach to generative modeling, and the social and economic forces that shape these systems and their interactions. Among her honors are Sloan fellowship, AI2050 fellowship, NSF CAREER, Google Research scholar award, NeurIPS and ICAPS best paper awards, EC exemplary track paper awards, and several other industry awards and fellowships.

AI2050 Project

Machine learning systems have become integral to our society, interacting with different sectors in unique and challenging ways. Yet, methods for addressing social and strategic considerations in ML are still in their infancy and lack provable performance. Nika Haghtalab’s AI2050 project aims to provide a unified perspective on a range of social and strategic considerations in ML, including robustness and fairness of learning systems.

Project Artifacts

D. Halawi, A. Wei, E. Wallace, T.T. Wang, N. Haghtalab, J. Steinhardt. Covert Malicious Finetuning: Challenges in Safeguarding LLM Adaptation. arXiv. 2024.

Affiliation

Assistant Professor, UC Berkeley

Hard Problem

Assurance, Responsibility