0.25
0.5
0.75
1.25
1.5
1.75
2
Explorations in Computer Go, Web Search, and Online Advertising
Published on Jul 25, 20114313 Views
In computer go, the goal is to find a good move in a given position by exploring the associated game tree, which is far too large to enumerate and hence requires sophisticated methods for navigation.
Related categories
Chapter list
Explorations in Computer Go, Web Search, and Online Advertising00:00
Overview - 101:08
Overview - 201:35
Maximising Long-Term Reward is Hard01:37
Exploration-Exploitation Trade-Off07:24
Overview - 308:39
Tabular Q-Learning08:48
Results12:28
Learning Aggressive Fighting13:32
Learning “Aikido” Style Fighting14:37
Lesson 116:18
Overview - 417:03
Reinforcement Learning for Car Racing: AMPS (Kochenderfer, 2005)17:59
Project Gotham Racing 319:43
Balancing Abstraction Complexity19:52
State Representation and Reward19:52
Actions20:21
Project Gotham Racing20:29
Lesson 223:30
Overview - 524:00
Computer Go24:10
The Game of Go24:55
Computer Go - 126:08
Computer Go - 226:52
Key Insights for Monte-Carlo Go28:20
Monte Carlo Go - 130:17
Monte Carlo Go - 231:07
Monte Carlo Go - 331:29
Monte Carlo Go - 431:34
Monte Carlo Go - 531:39
Upper Confidence Intervals32:24
Success of Monte-Carlo Go33:44
Lesson 335:18
Overview - 635:52
Traditional Web Search Paradigm35:56
User Feedback and Clicks38:29
Generalisation across Documents39:44
Diversity of Search Results41:18
Dynamics and Mortality42:06
Lesson 442:50
Overview - 743:12
bing43:13
AdPredictor: Bayesian Probit Regression44:11
The Causal Loop44:20
Thompson Heuristic47:05
Lesson 548:16
Lessons Learnt48:49