Research map / Bandits and online learning / Online convex optimization and games
Safe Learning in Tree-Form Sequential Decision Making: Handling Hard and Soft Constraints
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add known always-strictly-safe strategy and mix it with LP solution (hard-threshold Algorithm 2) | hard-threshold safe learning in SDMCs · Kuhn K5, K7, L2 Poker | cumulative regret and per-round safety | No clear effect |
| use the LP-based hard-threshold algorithm instead of COX-UCB (which needs private-information observability) | constrained opponent exploitation · Kuhn K5, K7, L2 Poker | cumulative regret | No clear effect |
| relax hard threshold to soft threshold (cumulative violation only) | safe learning in SDMCs · theory: regret and violation | regret and cumulative violation bounds | Helped |
| require per-round hard safety guarantee without extra assumption | general repeated SDMCs · theory | regret | Hurt |
Closest papers by meaning
- Model-Free Online Learning in Unknown Sequential Decision Making Problems and Games (AAAI 2021 · similarity 0.90)
- Bandit Linear Optimization for Sequential Decision Making and Extensive-Form Games (AAAI 2021 · similarity 0.89)
- Mediator Interpretation and Faster Learning Algorithms for Linear Correlated Equilibria in General Sequential Games (ICLR 2024 · similarity 0.87)
- Learning in Markov Games with Adaptive Adversaries: Policy Regret, Fundamental Barriers, and Efficient Algorithms (NeurIPS 2024 · similarity 0.87)
- Impact of Decentralized Learning on Player Utilities in Stackelberg Games (ICML 2024 · similarity 0.87)
- Policy-Regret Minimization in Markov Games with Function Approximation (ICML 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.