The University of Helsinki and the Finnish Society for Synthetic Chemistry invite you to participate in a summer school focused on the application of machine learning and artificial intelligence in synthetic chemistry. This five-day event combines expert lectures and hands-on exercises, providing a comprehensive overview of the latest methods in the field.
The summer school is aimed at graduate students, researchers, and advanced chemistry students who want to develop their skills in applying machine learning and artificial intelligence in synthetic chemistry. Previous programming experience is not required, as the course will also cover the basics of Python programming.
Join us to learn, network, and deepen your understanding of the tools shaping the future of chemistry!
Time and venue of the Summer School
The summer school will be organized from Monday June 9th to Friday June 13th at the Department of chemistry of the University of Helsinki, Finland.
How to enroll
Maximum of twenty-five (25) participants will be selected for the summer school based on a short application and a motivation letter. The application will open on 1st of April. The applications must be submitted no later than 28th of April and the selection will be done before 5th of May based on the applications and motivation letter. Members of the Finnish Society For Synthetic Chemistry are prioritized in the selection.
Price
The fee for the summer school is 300 € for academics and 1100 € for industrial participants. Memebers of the Society will got discount of 50€/100€ (academic/industry). For both industrial and academic participants the fee includes lunches, refreshments during the coffee breaks, the school dinner on Thursday and all the lecture materials.
Lecturers
Lecturers
- Dr. Susana Abellan (University of Zaragoza, Spain)
- Dr. Juan Alegre (University of Zaragoza, Spain)
- Dr. David Dalmau (University of Zaragoza, Spain)
- Dr. Diego Garay (Institute of Chemical Research of Catalonia)
- Dr. Samuel Genheden (AstraZeneca)
- Prof. Kari Laasonen (Aalto University, Finland)
- Dr. Ruben Laplaza (EPFL, Switzerland)
- Dr. Lucia Moran (University of Oslo, Norway)
- Prof. Robert Paton (Colorado State University, USA)
- Prof. Thijs Stuyver (PSL University in Paris, France)
- Prof. Philippe Schwaller (EPFL, Switzerland)
Organizers
- Dr. Juan Alegre (University of Zaragoza, Spain)
- Dr. Juho Helaja (University of Helsinki, Finland)
- Dr. Juri Timonen (University of Helsinki & University of Eastern Finland, Finland)
Program of the week
| Mon 9th | Tue 10th | Wed 11th | Thu 12th | Fri 13th | |
|---|---|---|---|---|---|
| Morning session 1 9.00 -10.30 | Arrival, practicalities & Get to know each other | Automation of reaction exploration with QM | ML to predict chemical outcomes | Chemical space exploration, dimensionality reduction and clustering. | TBA |
| Coffee break 1 | |||||
| Morning session 2 11-12.30 | Prediction of properties for organic compounds | Data visualization | Where to find data? FAIR principles and open-source projects | Neural network-based models in chemistry | TBA |
| Lunch 12.30-13.30 | |||||
| Hands-on exercises 1 13.30-15.00 | Basics of Python, Notebooks & ML | Descriptors and featurization | Data curation and evaluation of ML models | Generative models and inverse design | Time for students’ projects |
| coffee break | |||||
| Hands-on exercises 2 15.30-17.00 | Digital chemical reactions | Automated descriptor generation with QM | ML automation | Active Learning with ML models | Time for students’ projects |
Content of the sessions
Basics of Python & ML: introduction to scientific modules in Python and ML libraries
Digital chemical reactions: introduction to SMARTS strings, using RDKit to set up digital reactions and automating reactions with multiple substrates
Automation of reaction exploration with QM: automated reaction exploration and TS searches from SMILES strings with the TS-Tools program
Data visualization: examples of data visualization for general and chemical problems
Descriptors and featurization: explanation about molecular descriptors, the different ways to generate then and how to featurize molecules with chemical properties, electronic and steric parameters, etc.
Automated descriptor generation with QM: how to use AQME to generate QM descriptors in an automated manner, including generation of conformers, input generation, analysis of outputs and data mining.
ML to predict chemical outcomes: supervised ML methods and examples to predict chemical outcomes such as reactions results and molecular properties
Where to find data? FAIR principles and open-source projects: description of FAIR principles as quality standards for digital protocols, and examples of open-source projects and databases such as ioChem
Data curation and evaluation of ML models: good practices for data curation and methods to evaluate ML models.
ML automation: automation of ML protocols using ROBERT
Chemical space exploration, dimensionality reduction and clustering: introduction to the concept of chemical space, using dimensionality reduction to simplify visualization, and using molecular clustering to select and organize molecules using their properties.
Neural networks in chemistry: introduction to neural network algorithms and their chemical applications
Generative models and inverse design: examples of generative models and inverse design to suggest new molecules digitally during chemical discovery
Active learning with ML models: examples of active learning (evolving ML models through iterative additions of data) in chemical discovery and reaction optimization
Learning Objectives
- Understand the key concepts of machine learning and artificial intelligence, such as regression, classification, and clustering, and apply these methods to solve problems in synthetic chemistry.
- Acquire practical skills in Python programming, including data preprocessing, building and evaluating machine learning models, and visualizing results in chemical applications.
- Learn to use machine learning techniques, such as random forests and neural networks, for predicting reaction mechanisms and evaluating new reaction outcomes.
- Gain proficiency in using computational tools to explore and analyze reaction spaces, including algorithms and databases for reaction prediction.
- Be able to automate computational methods, such as optimization tasks and simulations, to accelerate and enhance synthesis processes.
- Apply learned skills in hands-on exercises, such as reaction modeling and synthesis process simulation, using real chemical examples.
- Deepen your understanding of the perspectives of international experts and the latest tools for applying machine learning and artificial intelligence in synthetic chemistry.
Contact information
Dr. Juri Timonen (juri.timonen@helsinki.fi)
Dr. Juho Helaja (juho.helaja@helsinki.fi)
Sponsored by:
The University of Helsinki Doctoral Programme in Chemistry and Molecular Sciences (CHEMS)
If your company/organization is willing to sponsor the meeting, please contact us either here of via e-mail!