All organizations · GSoC 2026 · Science and medicine, Data

Machine Learning for Science (ML4SCI)

Machine learning applications in science

Machine Learning for Science (ML4SCI) is an umbrella organization for machine learning-related projects in science. ML4SCI brings together researchers from universities and scientific laboratories with motivated students to join existing scientific collaborations and contribute to cutting edge science projects across a wide variety of disciplines. Students work on existing problems to develop new machine learning-based approaches and produce open source code that directly contributes to solving these scientific challenges. ML4SCI currently includes projects from a variety of fields. For example, some of them explore the uses of machine learning for particle reconstruction and classification in high-energy physics, deep learning-based searches for dark matter in astrophysics, applications of machine learning techniques to data returned from planetary science missions, applications of quantum machine learning to science, and others. Machine learning ideas and approaches can be broadly applicable and transferable across the scientific domains. The goals of ML4SCI projects are to grow the open-source community in machine learning for science by addressing important scientific …

  • python
  • machine learning
  • c++
  • data analysis
  • artificial intelligence
  • machine learning
  • science and medicine
  • algorithms
  • physics
  • astronomy

Project ideas

70 ideas on its 2026 ideas page.

  1. Event Classification With Masked Transformer Autoencoders
  2. Super resolution at the CMS detector
  3. Agentic AI for High Energy Physics Analyses at the CMS Detector
  4. Agentic AI for Autonomous Gravitational Lensing Simulation Workflows
  5. Hybrid Quantum-Classical Representation Learning for Dark Matter Substructure Classification
  6. Neural Operators for Fast Simulation of Strong Gravitational Lensing
  7. Foundation Model for Gravitational Lensing
  8. Physics Guided Machine Learning on Real Lensing Images
  9. Gravitational Lens Finding
  10. Data Processing Pipeline for the LSST
  11. Physics-Informed Diffusion Models for Gravitational Lensing Simulation
  12. Unsupervised Super-Resolution and Analysis of Real Lensing Images
  13. Super resolution at the CMS detector
  14. End-to-End event classification with sparse neural networks
  15. Diffusion models for fast and accurate simulations of low level CMS experiment data.
  16. Deep Learning Inference for mass regression
  17. Linear attention vision transformers for end to end mass regression and classification
  18. End-to-End particle collision track reconstruction
  19. Foundation models for End-to-End event reconstruction
  20. Discovery of hidden symmetries and conservation laws
  21. Equivariant Vision Networks for Predicting Planetary Systems’ Architectures
  22. Denoising Astronomical Observations of Protoplanetary Disks
  23. Exoplanet Atmosphere Characterization
  24. Foundation Models for Exoplanet Characterization
  25. Quantum Machine Learning for Exoplanet Characterization
  26. Diffusion Models for Fast Detector Simulation
  27. Graph Representation Learning for Fast Detector Simulation
  28. Graph Transformers for Fast Detector Simulation
  29. Optimal Transport in High Energy Physics
  30. Fast Accurate Symbolic Empirical Representation Of Histograms
  31. Deep Graph anomaly detection with contrastive learning for new physics searches
  32. Learning Parametrization with Implicit Neural Representations
  33. Learning the Latent Structure with Diffusion Models
  34. Non-local GNNs for Jet Classification
  35. Physics-Informed Neural Network Diffusion Equation (PINNDE)
  36. Physics-Informed Neural Network Diffusion Equation (PINNDE)
  37. Symbolic Regression for Observable Event-Level Reweighting
  38. Jet Observable Library for Monte Carlo Validation
  39. ML-Based Simulation Bias Analysis - Pythia vs Herwig vs Data
  40. Agentic Exclusion Limit Extraction

Read the ideas page. Ideas are for scope reference; never copy one into a proposal.

History

In GSoC since 2021, 6 years. Projects per year: '21: 19, '22: 20, '23: 23, '24: 26, '25: 31, '26: 33.

On GitHub

12 repositories pushed to in the last year. Main languages: Jupyter Notebook (8), Python (3), HTML (1).

Links and channels

go deeper in your terminal
git-into-gsoc org "Machine Learning for Science (ML4SCI)" --shortlist
git-into-gsoc repos "Machine Learning for Science (ML4SCI)"
git-into-gsoc issues --org "Machine Learning for Science (ML4SCI)"
or ask your assistant:
"Which Machine Learning for Science (ML4SCI) repositories merge
newcomers' pull requests, and which beginner
issues are still free?"

Those commands read live data: which repositories merge outside pull requests, and which beginner issues nobody has claimed. Install in two minutes.