Scholar
Andreas Kirsch
Google Scholar ID: WYQVZpYAAAAJ
Unknown affiliation
Citations & Impact
Citations
1,978
H-index
13
i10-index
13
Publications
20
Co-authors
28
list available
Publications
Resume (English only)
Academic Achievements
- Published 'All models are wrong, some are useful: Model Selection with Limited Labels', accepted to AISTATS 2025. Another paper, 'Turning Up the Heat: Min-p Sampling for Creative and Coherent LLM Outputs', was accepted to ICLR 2025.
Research Experience
- Currently a Research Scientist at Google DeepMind working on agentic LLMs. Previously, a Researcher (Research Scientist) at Midjourney for a year. Worked as a software engineer at Google (YouTube Monetization) and as a performance research engineer at DeepMind for a year in 2016/17.
Education
- Obtained a DPhil (PhD) from the University of Oxford in the OATML group and as a student in the AIMS CDT program, supervised by Prof Yarin Gal.
Background
- Research interests include information theory and its applications, information bottlenecks, and active learning using Bayesian deep learning, uncertainty quantification. Also interested in AI ethics and AI safety.
Miscellany
- Originally from Romania, grew up in Southern Germany. Twitter handle @blackhc.
