Collected molecules will appear here. Add from search or explore.
Provides provable finite-sample safety guarantees for Reinforcement Learning (RL) by applying adaptive conformal risk control to both offline and online RL settings.
stars
1
forks
0
The project addresses a high-value niche (provable safety in RL) using a sophisticated mathematical approach (conformal prediction). However, with only 1 star and no forks after 249 days, it is currently a dormant research repository or personal experiment rather than a living project. While the combination of CP and RL is relatively novel in open source, the lack of adoption makes it trivial to replicate or supersede.
TECH STACK
INTEGRATION
reference_implementation
READINESS