AI laboratory
In our AI Lab, we provide human and computational resources for research and development of state-of-the-art AI (neural network) models.
The computational resources are provided by a cluster of high-end Linux workstations with several powerful GPUs of the latest generation (NVIDIA RTX A6000, Geforce RTX 3090 etc.) and a large file server for storing the training datasets. This allows state-of-the-art neural networks, which have billions of parameters trained in a reasonable amount of time.
Our teams have a long history of research in AI and machine learning and have a pool of AI and ML experts in all research groups. We keep up to date with the latest neural network models and artificial intelligence software frameworks and publish regularly at trade shows and conferences. Our AI experts conduct research and development in a wide range of research areas, such as transfer learning, learning from little data, unsupervised learning, compact neural networks, explainable AI, and more. We have also gained knowledge on how to run efficiently neural networks on a variety of platforms, from embedded devices (e.g., smartphones) to high-end workstations. Ethical and legal AI aspects as well as Data privacy are also considered.
Our DIGITAL institute is a member of national and European AI research and innovation networks (e.g. CLAIRE) and is involved in the development of new AI standards, e.g. the MPEG CDVA standard, which is working on compact descriptors for video analysis.
