Description:
SustainML is an EU-funded project that stands for a new way of developing Machine Learning (ML) algorithms to reduce the CO2 footprint at all stages of the ML application lifecycle, from the design and exploration phase to the final training of the production systems. It also covers a continuous online recycling phase during deployment for the inference process.
SustainML will define a new framework with a different perspective on state-of-the-art processes to design and develop Machine Learning. It will contain unique features, such as:
- A software design assistant and interactive tools to guide developers through the entire process.
- A toolchain dedicated to real application testing and validation based on the SustainML framework.
eProsima and SustainML:
eProsima coordinates the SustainML project and will be responsible to design the GUI and the framework architecture. Additionally, the company will be in charge of communication and dissemination activities, aiming at the success of the project and the long-term maintenance results.
SustainML financing:
SustainML received funding from Horizon Europe research and innovation program under grant agreement No 101070408. Horizon Europe continues the work of the previous Horizon 2020 Research and Innovation program with 95.5€ billion of funding available from 2021 to 2027.
Acknowledgement:
This paper is supported by European Union’s Horizon Europe research and innovation programme (HORIZON-CL4-2021-HUMAN-01) under grant agreement No 101070408, project SustainML (Application Aware, Life-Cycle Oriented Model-Hardware Co-Design Framework for Sustainable, Energy Efficient ML Systems).
SustainML consortium:
The diversity of the participants in this project, which range from small to multinational companies and universities, guarantees the variety of research, innovation and application outcomes. The consortium is constituted by: DFKI, IBM, UPMEM, Inria, University of Copenhagen, Technische Universität Kaiserslautern, and eProsima.
More Information about SustainML:
For any questions please contact This email address is being protected from spambots. You need JavaScript enabled to view it.
