Resource Aware ML
by ,
DO: JvF25/3-303 | BN: b-it/1.047
Bridging Algorithms and Hardware: Towards Resource-Efficient Machine Learning
As machine learning models continue to grow in complexity, their computational and energy demands increasingly constrain scalability and real-world deployment. This lecture examines the close interplay between model design and compute architecture, showing how algorithmic choices directly shape hardware efficiency. We will explore emerging strategies for resource-aware machine learning — from hardware-specific model optimizations that preserve accuracy to advanced compression techniques that move beyond simple quantization. Particular focus will be given to binarized neural networks, an extreme form of quantization that paves the way toward sustainable and deployable AI on edge and embedded systems.
The talk will be given in two parts. Part one will be given by Prof. Dr. Jian-Jia Chen and part two will be given by Dr. Sebastian Buschjäger.
Vanessa Faber & Brendan Balcerak Jackson