Research
What we work on
Selected projects from the group, explained for a general technical audience. For the full list of papers, see Publications.
arXiv preprint 2025
Forecasting by Learning Dynamic External Influence
Moving beyond history: forecasting models that learn from time-aware external events, with applications to financial markets.
Read the post
AAAI 2025
Forecasting Multi-Relational and Recursive Interactions in Temporal Networks
A temporal point process on recursive hyperedges for forecasting higher-order interactions in dynamic networks.
Read the post
AAAI 2025 (Oral)
Active Reinforcement Learning Strategies for Offline Policy Improvement
Choosing which new trajectories to collect so that an offline-trained policy improves with as few environment interactions as possible.
Read the post