Helmholtz AI

Democratizing AI

Helmholtz Artificial Intelligence Cooperation Unit

Vision

We aim to reach an internationally visible leadership position in applied artificial intelligence (AI) / machine learning (ML) by combining unique research questions, data sets and expertise with newly developed AI/ML-based tools and democratized access to them in an open and dynamic community.

Mission

Our mission

We are a research-driven hub for applied artificial intelligence (AI) that:

  • fosters cross-field creativity by stimulating collaborative research projects
  • identifies and leverages similarities between applications to advance generalised AI / machine learning (ML) methods integrates field-specific excellence and AI/ML prowess
  • improves the quality, scalability and timely availability of emerging methods and tools
  • empowers and trains the current and next generation of scientists 

to enable the efficient and agile development and implementation of AI/ML assets across the whole Helmholtz Association.

Right Place, Right Time

Artificial intelligence (AI) and machine learning (ML) comprise technologies that have begun to impact industry, science and society in an unprecedented way. Speech and image recognition are prominent examples of applications that became reliable and commonplace over the last few years and that we find in devices we use every day. Ever more AI/ML-based technologies are employed in industrial settings, they shape how services are provided and products are manufactured, and they open entirely new avenues for research. Their impact on everyday life will make AI and ML one of the major transformative forces in the 21st century.

The scale of public and private investments, the surge in research and the intensity of policy debates highlight the relevance, potential and urgency of empowering researchers to apply AI/ML

We are convinced that coordinated efforts and cooperation are key for success: Integrating field-specific and technological excellence, ensuring quality and providing swift access - that is what Helmholtz AI is about. As Germany’s largest research organisation, the Helmholtz Association’s 19 research centres collect, store and analyse vast amounts of data - information and data science is both a natural and explicitly defined strategic focus. It is the ideal place to be and the time is right to tap this treasure trove of data to develop tools and applications based on AI and ML.

Participating organisations

German Aerospace Center (DLR)
Forschungszentrum Jülich
Helmholtz-Zentrum Hereon
Helmholtz-Zentrum Dresden-Rossendorf
Helmholtz Zentrum München
Karlsruhe Institute of Technology (KIT)
Helmholtz Association of German Research Centres

Impact

Output

Team

Related projects

Helmholtz Imaging

We catalyze scientific discovery from sensory measurements to knowledge.

Updated 4 days ago

HIFIS

Helmholtz Federated IT Services

Updated 5 months ago

Related software

AutoPQ

AU

Automated point forecast-based quantile forecasts

Updated 4 months ago
7

AutoPV

AU

Automated photovoltaic forecasts with limited information using an ensemble of pre-trained models.

Updated 4 months ago
3 5

AutoWP

AU

Automated wind power forecasts with limited computing resources using an ensemble of diverse wind power curves.

Updated 4 months ago
7

Heat

HE

Heat is a flexible and seamless open-source software for high performance data analytics and machine learning. It provides highly optimized algorithms and data structures for tensor computations using CPUs, GPUs and distributed cluster systems on top of MPI.

Updated 21 months ago
27 11

Propulate

PR

An asynchronous evolutionary optimization algorithm and software package for global optimization and hyperparameter search on high-performance computers.

Updated 2 weeks ago
11 6

SeisBench: A toolbox for machine learning in seismology

SE

SeisBench is an open-source Python toolbox for machine learning in seismology. It brings together the whole machine learning model lifecycle: datasets and benchmarks, models and training pipelines, and efficient implementations for deploying the models in production.

Updated 13 months ago
25 6