This section provides a curated overview of Self-Supervised Learning (SSL) techniques and their applications to simulation data in Computer-Aided Engineering (CAE). It covers fundamentals, tutorials, review papers, example applications, online courses, books, articles, open-source libraries, and personal recommendations.
- Learning Representations:
Leverage SSL to learn features from unlabelled simulation data. - Pre-training for Downstream Tasks:
Use SSL to pre-train models on large simulation datasets before fine-tuning on limited labelled data.
- Self-Supervised Learning (SSL) – A Gentle Introduction:
A comprehensive blog post by Lilian Weng that explains SSL concepts and applications.
Lilian Weng's Self-Supervised Learning Blog - The Self-Supervised Learning Cookbook: Book Link
A practical guide for researchers and practitioners (search for the latest version online).
- SelfEEG: A Python library for SSL in electroencephalography.
- MMSelfSup: An open-source toolbox for self-supervised representation learning based on PyTorch.
- VISSL: A library for state-of-the-art self-supervised learning from images.
- Lightly: A Python library focused on SSL for images.
- When I started with Self-Supervised Learning for my master's thesis, I discovered a wonderful blog by my supervisor that greatly enhanced my understanding.
Lilian Weng's Self-Supervised Learning Blog - I highly recommend exploring his other related blog posts for additional insights.
- Community Contributions:
You are encouraged to contribute additional resources, improvements, and feedback. Check out the initial repository at cae_ai_resources and submit a pull request with your suggestions.
Happy learning and contributing!