The Next Decade of Software Is About Climate - What Is the Role of ML?

Explore the critical intersection of artificial intelligence and climate change in this thought-provoking conference talk. Delve into the substantial carbon footprint of AI systems and their impact on global warming targets. Examine the rising interest in assessing and mitigating carbon emissions throughout the AI lifecycle. Learn about the Green Software Foundation's Software Carbon Intensity specification and its importance in measuring environmental impact. Discover strategies for reducing the carbon footprint of AI, including optimizing model size, implementing carbon-aware features, and leveraging sustainable computing practices. Gain insights into the future of sustainable AI development and practical steps to create more environmentally responsible software. Engage with topics such as data collection efficiency, federated learning, and edge computing as potential solutions for greener AI implementations.

Introduction
The Age of AI
Bill Gates AMA
Sara Bergman
Agenda
Why Software and Climate
Machine Learning and Climate
Model size is exploding
State of AI
Carbon Footprint
Green Software
Green Software Foundation
Measuring Software
Carbon Efficiency
Quantification
Quantification Methods
Core Characteristics
Net Zero
SEI Score
Machine Learning Climate
Patterns
Life Cycle
Project Planning
Carbon Aware Features
Service Level Objectives
Data Collection
Data Hoarders
Data Green
The Next Phase
Graph
Training Models
Federated Learning
Optimizing Size
Time Shifting
Sustainable Regions
Hardware
Deployment Maintenance
Azure Machine Learning
Variable Power
Hosting Efficiency
Serverless Architecture
AI Models on the Edge
Where to Start
Summary
Free course
Questions
What about software in general
What is the Green Software Foundation doing