Syllabus
Sylabus:
Date
Title
Speaker 26.02
Intro and Mathematical foundation to DL
Bartek Wilczynski (Warsaw) 04.03
Convolutional and Recurrent neural networks
Marco Frasca (Milano) 11.03
Autoencoders and variational autoencoders
Carl Herrmann (Heidelberg) 18.03
Attention mechanisms and transformers
Dario Malchiodi (Milano) 08.04
Transformers and RNN for sequence analysis
Dario Malchiodi (Milano) 15.04
Models for multimodal data integration
Britta Velten (Heidelberg) 22.04
VAE in single-cell genomics
Carl Herrmann (Heidelberg) 29.04
AF, EMSFold to predict structure of proteins
Joanna Sulkowska (Warsaw) 06.05
RNN, CNN models for topology/graph analysis in biopolymers
Joanna Sulkowska (Warsaw) 13.05
Deep learning models for protein-ligand binding site prediction
David Hoksza (Prague) 23.05
Diffusion models for protein design
Elodie Laine (Paris) 27.05
Intro to BioImage Analysis and Deep Learning Utilization
Martin Schatz (Prag) 03.06
Deep Architectures for sampling macromolecules
Grégoire Sergeant-Perthuis (Paris) 10.06
Deep learning for segmentation
Karl Rohr (Heidelberg)
Annotation
Project DeepLife intends to be a comprehensive and application-oriented teaching project which will introduce students to the most advanced algorithms and applications of deep learning in life sciences. For this, we will build on existing complementarities between the bioinformatics master programs of the five universities identified during the first educational project, and offer a new and comprehensive course covering the different aspects of Deep learning approaches in life sciences. We will set a strong focus on complementary application areas in life sciences and practical implementations. We will cover three very active application areas of deep learning in life sciences: (1) structural bioinformatics, (2) application of deep-learning to single-cell genomics, (3) biomedical image analysis. This course will be held in hybrid mode, with online lectures by teachers from different institutions and on-site practical exercises in the form of Q&A sessions in smaller groups. Our objective is to bring together the strong and complementary expertise in structural bioinformatics (Paris, Prag, Warsaw), single-cell genomics (Heidelberg, Warsaw) and image analysis (Heidelberg, Prag, Milano).
The course is composed of online lectures and online exercises in the form of notebooks that are students supposed to succesfully complete.
The course will be extended for best 15 students with an on-site 2-day hackathon during which mixed teams will work on small implementation projects around selected topics of deep-learning.