AGH University of Science and Technology in Kraków is opening the largest Artificial Intelligence Centre of Excellence in Poland. On this occasion, Sano is invited to give several AI workshops for the students. Sano scientists will provide workshops on 4 exclusive topics on 10 of May:
- Deep Reinforcement Learning for Surgical Robotics;
- Brain Connectome and graph convolutional networks, merging the two types of networks;
- From small data to big impact;
- Introduction to Federated Learning with Flower.dev Open-Source Framework.
More about the topics.
“Deep Reinforcement Learning for Surgical Robotics” – the workshop is organised by Przemysław Korzeniowski, Michał Naskręt, Sabina Kamińska, and Diego Dall’Alba.
Reinforcement Learning (RL) is an exciting area of study that has applications in various fields. Our workshop provides an opportunity for students to learn how RL algorithms can be implemented in robots to perform a wide range of tasks. One field that is particularly interesting is healthcare, where robots can assist surgeons with various visuomotor tasks. During the workshop, students will have the chance to work with virtual environments for simulating surgical scenarios. Students will develop an understanding of how RL algorithms work and how they can be applied to create robotic policies that can be transferred to real-world scenarios (Sim-To-Real). The workshop provides a unique opportunity for students to gain practical skills that can be applied to a variety of fields.
The next workshop title is “Brain Connectome and graph convolutional networks, merging the two types of networks”, itis organized by Alessandro Crimi and Joan Falco Roget.
Advancements in neuroimaging led us to represent the brain as a graph or network. At the same time particular convolutional neural networks can handle data based on graphs. In this workshop, we delve into how to construct those representations and how we can use them to study diseases such as brain tumours or neurodegeneration.
“From small data to big impact” – the workshop proposed by Jose Sousa.Machine Learning developments have a central need for high volumes of data to allow for pattern descriptions. From these patterns perspective models of the world are developed to allow for classifications. However, despite its huge success on several classification problems it still struggles with the capability to deal with unseen events. In the healthcare decision-making process, these two properties are conflicting, big data is difficult to obtain, and unseen events are common. In this workshop, we will try to explore approaches to deal with the development of the continuous learning process capable to mitigate these limitations.
Maciej Malawski and Michał Daniłowski, in their turn, will provide “Introduction to Federated Learning with Flower.dev Open-Source Framework”. Federated Learning (FL) is an interesting approach to distributed machine learning (ML), in which multiple clients can collaboratively train a ML model without sharing their data with a central server it is important in all the applications where data privacy is concerned: ranging from personal data from mobile devices to medical data from hospitals. In this workshop we will provide an overview of typical problems related to FL, explain some basic concepts and algorithms for FL, and give a hands-on experience with Flower.dev open-source framework.