معرفی
Jean-Baptiste Alayrac is a Researcher at DeepMind, focusing on structured learning from video and natural language. His academic background includes a PhD from the Sierra and Willow groups at Ecole Normale Supérieure and Telecom ParisTech, where he explored machine learning and computer vision. He has held teaching roles as a Teaching Assistant at Ecole Normale Supérieure and other universities, contributing to courses in statistical machine learning and mathematics.
His research interests span multimodal learning, vision-language models, self-supervised learning, and efficient retrieval systems. Notable projects include the Flamingo model for few-shot learning and the Perceiver IO architecture for structured data processing. He has also contributed to foundational works like HowTo100M, leveraging large-scale video-text embeddings.
Alayrac's publications emphasize cross-modal interactions, with key contributions in adversarial robustness, layered video representations, and weakly supervised learning. His work often bridges computer vision and natural language processing, with applications in instructional video analysis and cross-lingual translation.
