Michal Rolínek

News

• Moved to London to work as a researcher at a trading firm G-Research
• Invited talk at the NeurIPS workshop on Learning Meets Combinatorial Algorithms
• Invited talk at INFORMS 2020!
• A fresh RL paper Sample-efficient Cross-Entropy Method for Real-time Planning!
• Our paper Optimizing Rank-based Metrics with Blackbox Differentiation was among the ~25 nominees for the best paper award at CVPR2020 (out of ~1400 papers). Yay!

Who am I? Machine-learning researcher, Python geek, math olympiad enthusiast, occasionally problem-solving teacher in Africa, mountain lover, amateur Waltz dancer

My research mission:

• Black magic for extracting signal from financial markets :).
• Use combinatorial algorithms to improve deep learning.
• Use deep learning to improve combinatorial algorithms.
• With the above two, compete on computer vision benchmarks.

Other research topics: keypoint matching, metric learning, constraint satisfaction, reinforcement learning, generative models, disentanglement, stochastic optimization, graph neural networks, operations research