Self-Supervised Learning and Data Curation at Scale in Computer Vision

Discover the game-changing power of self-supervised learning (SSL) in this exciting webinar, where we’ll explore how SSL is revolutionizing model training by utilizing large-scale datasets without manual annotations. Learn about its remarkable success in NLP with models like BERT and GPT and its growing influence in computer vision through techniques like DINO, MAE, and AIM. Dive into how Lightly is advancing SSL in computer vision with its popular open-source framework, LightlySSL, used in over 800,000 installations. Uncover how SSL-trained models simplify data curation, enabling scalable, efficient workflows for modern AI applications.

Key Takeaways:

• Discover how self-supervised learning (SSL) leverages large-scale datasets without manual annotations, revolutionizing NLP and computer vision.
• Explore advancements in SSL techniques, including DINO, MAE, and AIM, and their transformative impact on computer vision.
• Learn about LightlySSL, a widely-used open-source framework with over 800,000 installs, driving SSL adoption in computer vision.
• Understand how SSL-trained models streamline data curation, enabling efficient and scalable workflows.

Igor holds an electrical engineering degree from ETH Zurich and has over five years of specialized experience in machine learning. He co-founded Lightly, a Y-Combinator-backed startup focusing on data curation and active learning that powers data pipelines at several Fortune500 companies. You can reach him on LinkedIn.