|
数十億のパラメーターを持つ大規模ニューラルネットワークに進化戦略 (ES)を適用する際の計算上およびメモリ上の障壁を克服するために設計されたEGGROLL(Evolution Guided General Optimization via Low-rank Learning)という最適化アルゴリズムが出現したことにより、AGIは2年早まるのではないかと言われ始めているそうです。 Evolution Strategies at the Hyperscale https://arxiv.org/pdf/2511.16652 NVIDIA×オックスフォード:AGIは2年早まったか? https://www.youtube.com/watch?v=xZr5is0MFrI Will AGI Arrive Two Years Earlier Thanks to EGGROLL? With the emergence of a new optimization algorithm called EGGROLL (Evolution Guided General Optimization via Low-rank Learning)—designed to overcome the computational and memory barriers involved in applying Evolution Strategies (ES) to large-scale neural networks with billions of parameters—there is growing speculation that AGI may arrive two years earlier than previously expected. Your browser does not support viewing this document. Click here to download the document. Your browser does not support viewing this document. Click here to download the document. Your browser does not support viewing this document. Click here to download the document.
0 Comments
Leave a Reply. |
著者萬秀憲 アーカイブ
September 2025
カテゴリー |
RSS Feed