risk-sensitive-rl

Reference simulator and stored results for Lyapunov-guided safe RL with risk-budget feasibility in energy-aware O-RAN scheduling: CVaR tail constraint, LCB safety filter, PID dual, and a swept classical drift-plus-penalty frontier.

Does richer tail-risk feedback help a language model write a better trading reward? A pre-registered study across 11 models, five feedback arms and 568 seeds per comparison unit. MSc dissertation, UCL Institute of Finance and Technology.

Official code for "RiskZero: Plan More to Risk Less with a Learned Model"

A Reinforcement Learning MVP (Minimum Viable Product) for Condition-Based Maintenance (CBM) using industrial equipment temperature sensor data. This project implements a sophisticated QR-DQN (Quantile Regression Deep Q-Network) agent to learn optimal maintenance policies balancing risk mitigation and cost minimization.

Code for UCB-BQRL, a model-based optimistic reinforcement learning algorithm for risk-sensitive control with smoothed quantile objectives under unknown transition dynamics.

Adaptive risk-aware reinforcement learning framework for time-constrained navigation tasks. This MSc thesis studies how an RL agent can adapt its risk attitude when mission time changes, using a hierarchical controller that switches from a risk-aware CVaR-constrained policy to a risk-seeking policy when time pressure makes switching benificial.