Biologically Informed Neural Networks (BINNs): Principles and Practice
Date: 11 - 16 March 2026
Language of instruction: English
Application Deadline: 5 November 2025 @ 11:57
This one-week course introduces the theory and practical use of biologically informed neural networks (BINNs) in the life sciences. Participants will learn how to integrate molecular networks with deep learning approaches and apply them to biological data.
Each day combines lectures with hands-on workshops, covering topics such as: translating prior biological knowledge into model structures (e.g. pathway-guided layers); training, troubleshooting, and evaluating BINNs; interpreting trained models; and reasoning about when to constrain or expand models to better reflect cellular complexity. By the end of the week, participants will be equipped with the tools to adapt a BINN template to their own research questions.
Prior to the course, concise primers on core neural-network concepts and commonly used biological datasets will be provided. Participants should bring their own laptop and have basic programming experience in Python, but no advanced knowledge of neural networks is required. The emphasis is on conceptual understanding and hands-on intuition.
Coursework will be conducted in an online notebook environment using small biological datasets and a pre-trained BINN, enabling participants to run, visualize, and interpret models. Guest lectures from researchers applying BINNs in areas such as cancer and infectious disease will provide insight into current applications and future directions.
Contact: avlant.nilsson@ki.se
Keywords: Deep Learning, Systems Biology, Omics
Venue: SciLifeLab Solna
City: Solna
Country: Sweden
Prerequisites:
No prerequisite courses, or equivalent, demanded for this course. While prior experience in areas such as bioinformatics, statistics, or deep learning can be beneficial, the course is designed to provide the necessary background for students from diverse scientific disciplines.
Learning outcomes:
After completing this course, the participant is expected to be able to:
•Explain the principles, motivation, and applications of Biologically Informed NeuralNetworks (BINNs) in predictive and interpretative biological research.
• Apply core neural network concepts and architectures, including how biological networkstructure can be integrated into models.
•Retrieve, process, and visualize biological network data for use in BINNs.
•Recognize and address common challenges in biological data, such as high dimensionality,noise, and batch effects.
•Critically evaluate and interpret BINN performance using appropriate metrics, featureselection, and interpretation methods.
•Critically compare and evaluate BINN applications from current scientific literature.
Target audience: PhD students
Tech requirements:
Prior to the course, concise primers on core neural-network concepts and commonly used biological datasets will be provided. Participants should bring their own laptop and have basic programming experience in Python, but no advanced knowledge of neural networks is required. The emphasis is on conceptual understanding and hands-on intuition.
Cost basis: Free to all
Activity log
