Nicolas Ballas
Papers
1
Total Citations
3
H-Index
1
About
Nicolas Ballas is a leading researcher in self-supervised learning and video understanding, with a focus on building AI systems that learn from observation rather than massive labeled datasets. His work centers on developing models that can comprehend, predict, and plan within visual environments by leveraging internet-scale video data. Ballas is best known for his contributions to video representation learning, particularly through the V-JEPA series, which explores how self-supervised objectives can enable machines to understand the physical world. His recent landmark paper, "V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning" (2025), demonstrates how combining large-scale video pre-training with minimal interaction data—such as robot trajectories—can produce models capable of sophisticated reasoning and planning. This work represents a significant step toward more sample-efficient AI that learns like humans: primarily by watching the world. While still early in its citation impact, V-JEPA 2 has quickly become a touchstone for researchers in embodied AI and video understanding. Ballas’s research sits at the intersection of computer vision, robotics, and foundation models, pushing the boundaries of what self-supervised learning can achieve in dynamic, real-world settings.
Research Focus
Key Achievements
Top Papers
- 1