Lenka Zdeborová
@zdeborova.bsky.social
Professor at EPFL. Une mathémaphysinformaticienne. Passionate mushroom hunter. Tamer of two little dragons.
Scaling laws don’t just show up in test error — they leave fingerprints in the weight spectrum.
In the feature learning regime, we map this connection: phase diagrams of scaling exponents <-> spectral signatures of trained weights. The paper is: arxiv.org/abs/2509.24882
In the feature learning regime, we map this connection: phase diagrams of scaling exponents <-> spectral signatures of trained weights. The paper is: arxiv.org/abs/2509.24882
September 30, 2025 at 11:02 AM
Scaling laws don’t just show up in test error — they leave fingerprints in the weight spectrum.
In the feature learning regime, we map this connection: phase diagrams of scaling exponents <-> spectral signatures of trained weights. The paper is: arxiv.org/abs/2509.24882
In the feature learning regime, we map this connection: phase diagrams of scaling exponents <-> spectral signatures of trained weights. The paper is: arxiv.org/abs/2509.24882
Excited about our progress in characterizing The Computational Advantage of Depth in Learning with Neural Networks. Check out the number of samples that can be saved when GD runs on a multi-layer rather than on a two-layer neural network. arxiv.org/pdf/2502.13961
February 22, 2025 at 2:23 PM
Excited about our progress in characterizing The Computational Advantage of Depth in Learning with Neural Networks. Check out the number of samples that can be saved when GD runs on a multi-layer rather than on a two-layer neural network. arxiv.org/pdf/2502.13961
Our advance in statistical physics analysis of simple graph convolutional neural networks arxiv.org/pdf/2402.03818 will appear at the Learning on Graphs conference: logconference.org. My student Odilon has now extended this to the multi-layer case, so more is coming soon.
November 22, 2024 at 11:27 AM
Our advance in statistical physics analysis of simple graph convolutional neural networks arxiv.org/pdf/2402.03818 will appear at the Learning on Graphs conference: logconference.org. My student Odilon has now extended this to the multi-layer case, so more is coming soon.