Download as PDF, PPTX
Uploaded byYan Xu
Secrets behind AlphaGo
This document discusses deep reinforcement learning and how it was applied in AlphaGo to master the game of Go. It provides an overview of deep learning, reinforcement learning, and how AlphaGo combined the two approaches. AlphaGo used deep neural networks to mimic human expert moves and play games against itself to estimate win probabilities. It had a policy network to choose moves and a value network to estimate game outcomes. Through deep reinforcement learning, AlphaGo was able to achieve superhuman performance at the game of Go.
More Related Content
![[246]reasoning, attention and memory toward differentiable reasoning machines](https://cdn.slidesharecdn.com/ss_thumbnails/246reasoningattentionandmemory-towarddifferentiablereasoningmachines-171017054258-thumbnail.jpg?width=640&height=640&fit=bounds)
Similar to Secrets behind AlphaGo
Secrets behind AlphaGo
- 1. Behind AlphaGo - DeepReinforcement Learning Houston Machine Learning Yan Xu 08/05/2017
- 2. Roadmap • Introduction andFeature Engineering (2 lectures) • Supervised Learning (4 lectures) • Unsupervised Learning (3 lectures) • Deep Learning series (4 lectures) • Optimization in Deep learning • Behind AlphaGo • Mastering the game of Go with deep neural networks and tree search • Reinforcement learning 101 – Ravi • Deep reinforcement learning • Attention network • Cuda Programming Hands-on - Martin • Application of Deep Learning • Object recognition • Chatbot Slides posted on: http://www.slideshare.net/xuyangela
- 3. Outline • Go game •Deep learning recap • Reinforcement learning 101 • AlphaGo system overview: Deep reinforcement learning More depth into deep reinforcement learning in upcoming meetups
- 4.
- 5.
- 6. Why hard forcomputer to play
- 7. Deep Learning + Reinforcement Learning = Deepreinforcement learning
- 8. Deep Learning: BasicComponent Activation function
- 9.
- 10.
- 11. Deep Learning: Architecture Convolutionalneural network feature map Feature map
- 12.
- 13.
- 14.
- 15.
- 16.
- 17. AlphaGo: Deep ReinforcementLearning http://techtalks.tv/talks/deep-reinforcement-learning/62360/ Mimic human experts Play against self Estimate wins Policy Network Value Network
- 18.
- 19.
- 20.
- 21.
- 28.
- 29. More coming up! •Behind AlphaGo • Mastering the game of Go with deep neural networks and tree search • Reinforcement learning 101 – Ravi • Deep reinforcement learning • Attention network • Cuda Programming Hands-on - Martin • Application of Deep Learning • Object recognition • Chatbot • Any proposal?
- 30. HML Speaker Hallof Fame Recognize contribution to Houston Machine Learning Meetup
- 31. Thank you ~ Slideswill be posted at: http://www.slideshare.net/xuyangela Leave a group review please