About me
I am an ELLIS PhD student currently in the Computational Principles of Intelligence Lab in Munich. I am supervised by both Eric Schulz (LMU & Helmholtz AI) and Jane X. Wang (Google DeepMind).
Broadly, I am interested in computational models of human behavior to try to leverage these insights in Machine Learning. So far in my PhD, I have been exploring how Large Language Models behave and take decision using the cogntive science lens. I am also interested to use these insights to inject human-like priors into RL agents.
When not doing research, I enjoy football, bouldering, skiing, surfing, calisthenics and dancing.
News
- May 2024: Two papers accepted at ICML 2024: CogBench: a large language model walks into a psychology lab and Ecologically rational meta-learned inference explains human category learning.
- April 2024: I gave a talk about Meta-learning in deep neural networks for the Harvard Efficient-ML seminar series as the ‘rising star speaker’.
- Feb 2023: We released CogBench, a benchmark for using cognitive psychology tasks to evaluate the behavior of LLMs.
- Sep 2023: Our paper on Meta-in-context learning in large language models was accepted at NeurIPS 2023.
- Aug 2023: I attended the MIT Brains, Minds & Machines Summer School in Woods Hole, MA.