Workshop scope

Generalizing materials descriptors – How can we interpret and develop novel materials descriptors?

Inverse learning problems and generative models – How can we take advantage of the latest algorithm developed in the field of computer science?

Latent space and reinforcement learning approaches – Which problems in material science can be tackle through an efficient use of exploration and exploitation steps ?

Quantifying uncertainty in ML predictions – How can uncertainty prediction be a fully integrated and beneficial step in the study of material properties?

Direct application of ML to MD simulations – What are the available platforms and the current challenges for machine learning potentials fully exploitable by the community?

Industry partners 

We will hear talks from industry leaders on how ML algorithms are employed in their research and product development:

IBM Research Zurich : Philippe Schwaller, “IBM RXN for Chemistry: Predicting Chemical Reactions using the Molecular Transformer”

IBM Finland : Jukka Remes, “IBM PowerAI/Watson Machine Learning (WML) Accelerator – infrastructure for open source -based scalable and productive deep learning”

Curious AI: Antti Rasmus, “Combining machine learning and reasoning tasks

Nvidia Finland: Timo Roman, “Towards autonomous driving with deep learning