ipi seminar [オンライン開催] 2021年12月16日(木)10:30~12:00

知の物理学研究センター / Institute for Physics of Intelligence (iπ)


Sho YAIDA (Facebook AI)

"Effective Theory of Deep Neural Networks"

Large neural networks perform extremely well in practice, providing the backbone of modern machine learning. The goal of this talk is to provide a blueprint for theoretically analyzing these large models from first principles. In particular, we'll overview how the statistics and dynamics of deep neural networks drastically simplify at large width and become analytically tractable. In so doing, we'll see that the idealized infinite-width limit is too simple to capture several important aspects of deep learning such as representation learning. To address them, we'll step beyond the idealized limit and systematically incorporate finite-width corrections.

※この講演に関するZoomのリンク等の案内を受け取ることを希望されるかたは、下記のgoogle formからメールアドレスをご記入ください。こちらに登録頂いた情報は、案内の配信のみに利用いたします。
                     世話人:知の物理学研究センター Tilman HARTWIG, 髙橋昂, 中西健
  • このエントリーをはてなブックマークに追加