After successful completion of the course, students are able to:
This course focuses on two aspects: (1) Sustainable AI (2) AI for Sustainablity.
Introduction: Hardware advancements, data explosion, and its energy impacts
Energy Challenge of AI models: Cost of Training and Inference
Large Language Models and their energy consumption
A path forward: Methods to address energy consumption of AI models
Understand, apply, and engineer large-scale geographically distributed ML/AL applications
Understand resource efficient mechanisms for ML/AL applications with strict latency and/or data quality constraints
Understand Geographically distributed inference and learning in Ai/ML
AI-driven Smart ICT management (data centre optimizations, chip architecture)
AI-driven renewable energy and grid energy management
AI for climate change modelling use cases
Understand the common AI/ML applications used for combating climate change (e.g., sensing of water pollution, flood sensing, etc.)
Facilitate computation and communication in rural and uninhabited areas
Understand how to apply different ML/AI methods for the implementation of applications combating climate change
The theoretical concepts are presented and discussed on the basis of slides and scientific literature. Practical tasks are carried out in the laboratory on the basis of these concepts. There will be two practical projects.
Oral exam at the end of the term
Knwoldge about Machine Learning
Ability to understand neural network, training the models and deploying in runtime
