News

[NeurIPS] Improved regret analysis for variance-adaptive linear bandits

[NeurIPS] Improved regret analysis for variance-adaptive linear bandits

The paper "Improved regret analysis for variance-adaptive linear bandits and horizon-free linear mixture MDPs," co-authored by Yeoneung Kim and Kwang-Sung Jun, has been accepted to Advances in Neural Information Processing Systems (NeurIPS) ...
[T-RO] Distributionally robust risk map for learning-based motion planning and control

[T-RO] Distributionally robust risk map for learning-based motion planning and control

The paper "Distributionally robust risk map for learning-based motion planning and control: A semidefinite programming approach" has been accepted for publication in the IEEE Transactions on Robotics. Distributionally robust risk map for learning-based motion planning and ...
[CDC 2 papers] Distributionally robust partially observable control, Distributionally robust unit commitment

[CDC 2 papers] Distributionally robust partially observable control, Distributionally robust unit commitment

The papers "Wasserstein distributionally robust control of partially observable linear systems: Tractable approximation and performance guarantee" and "On affine policies for Wasserstein distributionally robust unit commitment" have been accepted to 2022 IEEE Conference on Decision and ...
[RA-L & IROS] Infusing MPC into meta-RL

[RA-L & IROS] Infusing MPC into meta-RL

The paper "Infusing model predictive control into meta-reinforcement learning for mobile robots in dynamic environments" has been accepted to IEEE Robotics and Automation Letters (RA-L) and 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) ...
[ICML] Accelerated gradient methods for geodesically convex optimization

[ICML] Accelerated gradient methods for geodesically convex optimization

The paper "Accelerated gradient methods for geodesically convex optimization: Tractable algorithms and convergence analysis" has been accepted to the International Conference on Machine Learning (ICML). Accelerated gradient methods for geodesically convex optimization: Tractable algorithms and convergence ...