CS 285: Lecture 13, Part 2 - YouTube

Follow along using the transcript. Follow along using the transcript. New New New New

all right so let's talk about some actual exploration algorithms and for now we'll still be in the multi-arm banded setting and we will be concerned with theoretically principled strategies strategies that theoretically get good regret good meaning not too far off from actually solving the palmdp so how can we beat the band how can we minimize this measure of regret well it turns out there are a variety of relatively simple strategies that provably get regret that is optimal in the big o sense and we can often provide theoretical guarantees on their regret now these algorithms are optimal up…

saved by

• Sebi S

related reading

• CS 285: Lecture 13, Part 3youtube.com
• The Multi-Armed Bandit Problem and Its Solutions | Lil'Loglilianweng.github.io
• A (Long) Peek into Reinforcement Learning | Lil'Loglilianweng.github.io
• Part 3: Intro to Policy Optimization - Spinning Up documentationspinningup.openai.com
• The Bayes Banditfrancesco215.github.io
• course_stat_rl.pdfmit.edu
• Part 2: Kinds of RL Algorithms - Spinning Up documentationspinningup.openai.com
• [2506.22401] Exploration from a Primal-Dual Lens: Value-Incentivized Actor-Critic Methods for Sample-Efficient Online RLarxiv.org
• RLAlgsInMDPs.pdfsites.ualberta.ca
• Exploration for the Efficient Deployment of Reinforcement Learning Agentsopenreview.net
• Exploration Strategies in Deep Reinforcement Learning | Lil'Loglilianweng.github.io
• RL_Notes__final_.pdfjubayer-ibn-hamid.github.io