home » news » Revolutionary AI initiative GAIn makes artificial intelligence energy-efficient

Revolutionary AI initiative GAIn makes artificial intelligence energy efficient

With the GAIn (Next Generation AI Computing) project, Bavaria is sending a strong signal for cooperation and sustainability between the Free States of Bavaria and Saxony. The initiative combines the strengths of the federal states in order to master the challenges of the fourth industrial revolution. The focus is particularly on drastically reducing the energy consumption of artificial intelligence. As a pilot project, GAIn aims to develop innovative AI hardware and corresponding software concepts in order to find new solutions to the challenges of AI transformation, such as energy consumption, limited predictability and reliability as well as compliance with legal requirements.

GAIn is a cooperation between the Ludwig-Maximilians-Universität München (LMU), the Technische Universität München (TUM) and the Technische Universität Dresden (TU Dresden). Together with baiosphere and the Digital Agency Saxony as network partners as well as various industrial partners, the cooperation between the two federal states is being intensified. Prof. Gitta Kutyniok (LMU) is leading the project: “GAIn is of essential importance for the future of our society and Germany as an economic location, and therefore has enormous political and social significance.”

Potential problems of current hardware platforms

In recent years, significant challenges have become noticeable in current hardware, caused in particular by exploding energy consumption.
The Landau principle states that every irreversible calculation in a computer system consumes a minimum amount of energy, depending on the temperature of the system.
The principle makes it clear that the energy efficiency of computers is determined by the energy costs of information processing and the efficiency of data transmission.
In GAIn, possibilities of modern platforms and post-Shannon approaches such as quantum computing will be found to further increase the efficiency of AI systems.

Network as the key to innovation

The GAIn project was only possible thanks to a strong network and the combination of expertise from both locations.
baiosphere is the strong Bavarian AI network in which researchers, funding institutions, industry partners and political stakeholders work closely together to get the project off the ground.
Dr. Michael Klimke (Managing Director of the Bavarian AI Agency) expects great things from GAIn: “We are delighted that we, as the Bavarian AI Agency, were able to provide the spark for the project and bring the right people together at the right time. GAIn is a particularly promising project that combines excellent AI research from Bavaria with new ideas for AI hardware from Saxony at the highest level.”
Networks such as baiosphere and the Digital Agency Saxony initiate collaborations that facilitate access to important resources and markets and promote the social relevance of the innovation process.
This strengthens the innovative power and sustainability of AI technologies and the regions as locations for pioneering technologies.

Strengths of the federal states combined

With the High-Tech Agenda Bavaria, the Free State of Bavaria has already defined forward-looking technologies such as AI, medical technology, quantum technology and robotics as strategically important topics for its development in 2019.
The focus is therefore on the theoretical foundations and practical application of these technologies.
Topics such as sustainability, reliability and ethical and legal requirements are also given special consideration.

Saxony contributes its expertise in the field of pioneering hardware and microelectronics to GAIn.
By attracting leading chip factories and progressive start-ups, Saxony has taken on a leading role in Europe.

By combining the strengths of both federal states, GAIn promises considerable advantages and widespread use of the technologies developed.

Energy-saving concepts and practical applications

Prof. Holger Boche (TUM) is working on computing and information theory in order to reduce energy consumption despite increasing data volumes: “The key objective is to significantly reduce energy consumption for computing despite the exponentially increasing data volumes observed in computing and AI applications.”
Prof. Stefanie Speidel (TU Dresden) is focusing on the practical implementation of theoretical concepts in AI medical robotics: “How can we transfer the theoretical concepts in combination with new hardware to a specific use case in the field of AI medical robotics?“.
Prof. Frank Fitzek (TU Dresden) is developing new, energy-saving concepts for AI.
He sees an urgent need to reduce the energy consumption of future AI methods.
“Without GAIn methods, the energy available worldwide could be consumed by AI alone. Our approach is to use alternatives to digital computing for AI.”

GAIn bundles competencies

The topics of AI and robotics are complex and require collective efforts.
The sustainable added value of these technologies can only be achieved through the cooperation of experts with complementary fields of experience.
Through the close cooperation between Bavaria and Saxony on the topic of AI and robotics and resource-saving predictability, the two states are aiming for strong location development.

The GAIn project is being funded with a total of 6 million euros by the Bavarian State Ministry of Science and the Arts and the Saxon State Ministry of Science, Culture and Tourism.
You can find the official press release on the project launch here: https://www.stmwk.bayern.de/pressemitteilung/12814/nr-63-vom-06-08-2024.html