AI vs Climate Change: Insights from the Cutting Edge
summary
DeepMind applies reinforcement learning and neural nets to real systems and reports concrete wins: 30% reduction in CPU wake‑ups on Android Pie that extends Pixel battery life, a cloud model using ~2,500 sensor inputs and ~20 control actions that achieves a 40% reduction in data center cooling energy via safe autonomous control, and a wind portfolio forecaster that makes ~700 MW of capacity 20% more valuable by improving 36‑hour production predictions.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 2:20The Atari Challenge
- 6:24AlphaGo Zero
- 10:21Our Purpose
- 11:01Challenge of Climate Change
- 13:28Consumer AI
- 15:42Google Data Centers
- 18:35Initial Results
- 19:39Lessons Learned
- 20:45Safety First Automation
- 22:42Performance Improvement
- 23:07Data Center Engineers
- 23:24Wind Energy
- 25:05How it works
- 26:39Summary
- 27:44Closing Thoughts
related talks
22mAI On The Farthest Edge
14mAI for Science
45mGreen AI: Making Machine Learning Environmentally Sustainable
29mHow AI and Deep Learning could effort the Climate Change
22mLet's Terminate Climate Change with AI
12mKeynote: What's New in AI
01h 51mAI and Social Entrepreneurship- 12mKeynote: What's New in AI