About
I am a PhD Student in Computer Science at the University of Stuttgart, Germany, advised by Prof. Andreas Bulling.
My research focuses on computational theory of mind, with a specific emphasis on deep learning approaches.
In my work, I design and assess innovative AI models for theory of mind tasks that span multiple modalities, including vision and natural language.
My ultimate goal is to improve human-AI collaboration.
I hold a BSc and a MSc degree in Physics.
During my studies I focused on theoretical physics, in particular theoretical ecology and neuroscience.
News
- [24.06.2024] 🔥 Paper accepted at ICML 2024 Workshop on Mechanistic Interpretability (Vienna, Austria)
- [16.05.2024] 🔥 Paper accepted at ACL 2024 (Bangkok, Thailand)
- [01.05.2024] ☀️ I got accepted at the ELLIS CoGenAI Summer School (Helsinki, Finland)
- [25.03.2024] 👨🏼🏫 Constantin Ruhdorfer, Anna Penzkofer and I are organising a practical course “Computational Theory of Mind and Cognition”. Our course has limited spots available, don't miss out on this opportunity!
- [01.03.2024] 🏆 My student Constantin Ruhdorfer's master's thesis was awarded the Rul Gunzenhäuser Award
- [09.12.2023] 🔥 Paper accepted at AAAI 2024 (Vancouver, Canada)
- [29.09.2023] 📑 Reviewing for CHI 2024
- [07.04.2023] 🔥 Paper accepted at CogSci 2023 (Sydney, Australia)
- [27.03.2023] 🔥 Paper accepted at INTERACT 2023 (York, UK)
Research
These days I'm especially interested in:
- Probing LLMs' internal representations, in particular those of mental states of others
- Designing socially-aware LLMs with advanced theory of mind capabilities
- Building causal models for planning using theory of mind
- Learning theory of mind through world models
- Improving human-AI collaboration using theory of mind
I maintain a page where I summarize every new paper involving theory of mind and LLMs.
Publications
-
C Ruhdorfer, M Bortoletto, A Penzkofer, A Bulling "The Overcooked Generalisation Challenge." Preprint, 2024
-
M Bortoletto, C Ruhdorfer, L Shi, A Bulling "Benchmarking Mental State Representations in Language Models." ICML Workshop on Mechanistic Interpretability, 2024
-
M Bortoletto, C Ruhdorfer, A Abdessaied, L Shi, A Bulling "Limits of Theory of Mind Modelling in Dialogue-Based Collaborative Plan Acquisition." ACL, 2024
-
M Bortoletto, L Shi, A Bulling "Neural Reasoning about Agents' Goals, Preferences, and Actions." AAAI, 2024
-
E Sood, L Shi, M Bortoletto, Y Wang, P Müller, A Bulling "Improving Neural Saliency Prediction with a Cognitive Model of Human Visual Attention." CogSci, 2023
-
G Zhang, M Bortoletto, Z Hu, L Shi, M Bâce, A Bulling "Exploring Natural Language Processing Methods for Interactive Behaviour Modelling." INTERACT, 2023
Padawans
I regularly supervise Master and Bachelor theses.
If you have an idea for your thesis, feel free to contact me.
If you are looking for available projects, check if there are Open Thesis Projects here.