A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.
New CTRL algorithm adapts to varying problem difficulty.
A new method learns multiple subspaces from data.
Paper proposes a controllable RANSAC method for anomaly detection.
CtrlNS learns latent factors and distribution shifts from sparse transitions without prior knowledge.
We introduce a control-tutored reinforcement learning (CTRL) algorithm. The idea is to enhance tabular learning algorithms so as to improve the exploration of the state-space, and substantially reduce learning times by leveraging some limited knowledge of the plant encoded into a tutoring model-based control strategy. …
A RL-based method adds conditional controls to pre-trained diffusion models.
Default-ERM shortcut learning persists even without additional information.
This work develops efficient methods for continuous-time distributional reinforcement learning.
When learning behavior, training data is often generated by the learner itself; this can result in unstable training dynamics, and this problem has particularly important applications in safety-sensitive real-world control tasks such as robotics. In this work, we propose a principled and model-agnostic approach to miti…