Benoit R. Cottereau

Event streams can dramatically enhance driving intelligence by improving temporal precision and motion awareness, outperforming traditional frame-based perception methods.

Event-based vision gets a lightweight, efficient boost: E-TIDE matches state-of-the-art forecasting accuracy while slashing model size and training costs.

SNNs can now learn robust visual representations from unlabeled event data, rivaling supervised learning in low-data regimes, thanks to a new contrastive self-supervised learning framework.