Universities face several challenges to enable their students and researchers to dive into Artificial Intelligence and Data Science, fom providing suitable content to provisioning adequate hardware infrastructure. Through the Data Science Developer Platform implemented on IBM ecosystem along with well-designed curricula for Machine Learning and Data Science, universities can focus more on delivering high-quality content and rich resources to their students and researchers.
The OpenPOWER Project at TUM in cooperation with our industry partner IBM is aimed at connecting industry with academia, providing technical infrastructure and know-how to progress the development in AI and Data Science, and making leading IT-resources more accessible to universities. Our target is to bring AI Machine learning and deep learning education and know-how, tools and technologies, infrastructure and support accessibility to universities and educational institutions in Europe and all over the world.
The hardware infrastructure is provided by our industry partner IBM based on their POWER architecture with hardware and software technologies that facilitate and enable AI-driven application development and research.
The infrastructure is utilized both in research and education, where the platform facilitates teaching practical courses where the students get to understand concepts and collect hands-on experience around AI and machine learning. Several successful projects have been already accomplished in cooperation with university research chairs from different domains incorporating AI ideas and research in their work. Additionally, we also support students working on their bachelor, master, or PhD thesis experimenting and developing new AI-driven applications.
- Development of a Data Science Developer Platform
- Development of Machine Learning and Data Science Curricula
- Cooperating with universities in adopting curricula and realization of research projects on the platform.
Companies have economic values in their data that remain largely unused at present since the necessary processes are not yet established and corresponding applications for analysis and evaluation with powerful hardware are missing. As a result, the sensible use and economic value of data cannot yet be fully exploited.
Therefore, the ability to create targeted data analyses is one of the core tasks of the “Data Scientist”. This includes a fundamental understanding of the actual machine learning algorithms, their application to modern hardware architectures with the aid of modern accelerators and the associated processes (DevOps). Only if an optimal coordination between development and operation is established, the best possible analyses can be achieved, and the full potential of the hardware can be exploited.
With its PowerAI products, IBM offers a comprehensive range of software and hardware that offer and support solutions for precisely such emerging scenarios.
Technische Universität München
Prof Dr. Helmut Krcmar
Dr. Holger Wittges