Michal Lukasik

About Me

I am a Research Scientist at Google DeepMind in New York.

Research Interests

LLMs for regression / numerical tasks.
Reinforcement Learning for LLMs.
Loss function design, e.g. ranking, distillation, label smoothing.
Fundamental research on deep learning and LLMs, e.g. memorization, bias-variance.

News and Highlights

Action Editor (equivalent to an Area Chair) for ACL and EMNLP 2024-2025 in the Machine Learning for NLP track.
Co-organized the workshop on Scaling Behavior of Large Language Models SCALE-LLM 2024 co-located with EACL 2024.
Reviewer for ICML, ICLR, Neurips 2020-2025. Outstanding reviewer ICML 2022 (top 10%).

Recent Publications

REAL: Regression-Aware Reinforcement Learning for LLM-as-a-Judge
Yasi Zhang, Tianyu Chen, Mingyuan Zhou, Oscar Leong, Ying Nian Wu, Michal Lukasik. 2026.

TRACT: Regression-Aware Fine-tuning Meets Chain-of-Thought Reasoning for LLM-as-a-Judge
Cheng-Han Chiang, Hung-yi Lee, Michal Lukasik. In ACL main, 2025.

Bipartite Ranking From Multiple Labels On Loss Versus Label Aggregation
Michal Lukasik, Lin Chen, Harikrishna Narasimhan, Aditya Krishna Menon, Wittawat Jitkrittum, Felix X. Yu, Sashank J. Reddi, Gang Fu, Mohammadhossein Bateni, Sanjiv Kumar. In ICML, 2025.

Better Autoregressive Regression via Regression-aware Fine-tuning
Michal Lukasik, Zhao Meng, Harikrishna Narasimhan, Yin-Wen Chang, Aditya Krishna Menon, Felix Yu, Sanjiv Kumar. In ICLR (spotlight), 2025.

Regression Aware Inference with LLMs
Michal Lukasik, Harikrishna Narasimhan, Aditya Krishna Menon, Felix Yu, Sanjiv Kumar. In EMNLP (findings), 2024.

It's an Alignment, Not a Trade-off: Revisiting Bias and Variance in Deep Models
Lin Chen, Michal Lukasik, Wittawat Jitkrittum, Chong You, Sanjiv Kumar. In ICLR (spotlight), 2024.

Two-stage LLM Fine-tuning with Less Specialization and More Generalization
Yihan Wang, Si Si, Daliang Li, Michal Lukasik, Felix Yu, Cho-Jui Hsieh, Inderjit S Dhillon, Sanjiv Kumar. In ICLR, 2024.