Research groups at the Faculty of Technology

The ‘Sustainable AI’ research group, led by Jun.-Prof. Dr David Kappel, is dedicated to investigating the computational complexity of machine learning algorithms with a view to significantly reducing their energy consumption whilst maintaining the same level of performance.

Our mission

Modern machine learning (ML) architectures consume unprecedented amounts of energy, with a single training session often exceeding the energy and carbon footprint of a car over its entire lifetime. At the current rate of growth, ML models could overtake the transport sector in the global energy balance within 10 to 20 years. Biological brains, by contrast, are extremely energy-efficient, demonstrating that efficient learning systems are, in principle, possible. The ‘Sustainable AI’ research group identifies the mechanisms that enable the remarkable energy efficiency of biological brains and explores new approaches to significantly reduce the energy consumption of machine learning using hybrid ML/bio-inspired models.

Research projects

• EVENTS (Energy-efficient distributed sensor systems for computer vision: event-based distributed AI algorithms) was launched in October 2022 as part of the ‘BMBF – OCTOPUS – Electronic Systems for Trustworthy and Energy-Efficient Decentralised Data Processing in Edge Computing’ funding programme. The aim of the project is to develop efficient AI algorithms that can be adapted for use on energy-efficient neuromorphic systems for computer vision. The project consortium, led by TU Dresden, will implement and test the developed algorithms on innovative neuromorphic hardware in various pilot applications. Further information can be found on the EVENTS project website.
• ESCADE (Energy-Efficient Large-Scale Artificial Intelligence for Sustainable Data Centres) launched in May 2023 and is funded by the Federal Ministry for Economic Affairs and Climate Action (BMWK). The aim of the project is to develop modern, large-scale, distributed and energy-efficient machine learning models for complex applications such as natural language processing. Further information can be found on the ESCADE project website.