English
Related papers

Related papers: Eigenvalues of the Hessian in Deep Learning: Singu…

200 papers

We study the properties of common loss surfaces through their Hessian matrix. In particular, in the context of deep learning, we empirically show that the spectrum of the Hessian is composed of two parts: (1) the bulk centered near zero,…

Machine Learning · Computer Science 2018-05-08 Levent Sagun , Utku Evci , V. Ugur Guney , Yann Dauphin , Leon Bottou

The loss function of deep networks is known to be non-convex but the precise nature of this nonconvexity is still an active area of research. In this work, we study the loss landscape of deep networks through the eigendecompositions of…

Machine Learning · Computer Science 2019-02-08 Guillaume Alain , Nicolas Le Roux , Pierre-Antoine Manzagol

The eigenvalue distribution of the Hessian matrix plays a crucial role in understanding the optimization landscape of deep neural networks. Prior work has attributed the well-documented ``bulk-and-spike'' spectral structure, where a few…

Machine Learning · Computer Science 2026-05-26 Shenyang Deng , Boyao Liao , Zhuoli Ouyang , Tianyu Pang , Yaoqing Yang

It is well-known that the Hessian of deep loss landscape matters to optimization, generalization, and even robustness of deep learning. Recent works empirically discovered that the Hessian spectrum in deep learning has a two-component…

Machine Learning · Computer Science 2022-08-02 Zeke Xie , Qian-Yuan Tang , Yunfeng Cai , Mingming Sun , Ping Li

To understand the dynamics of optimization in deep neural networks, we develop a tool to study the evolution of the entire Hessian spectrum throughout the optimization process. Using this, we study a number of hypotheses concerning…

Machine Learning · Computer Science 2019-01-30 Behrooz Ghorbani , Shankar Krishnan , Ying Xiao

The Hessian of neural networks can be decomposed into a sum of two matrices: (i) the positive semidefinite generalized Gauss-Newton matrix G, and (ii) the matrix H containing negative eigenvalues. We observe that for wider networks,…

Machine Learning · Computer Science 2020-01-15 Etai Littwin , Lior Wolf

This study delves into the intricate dynamics of trained deep neural networks and their relationships with network parameters. Trained networks predominantly continue training in a single direction, known as the drift mode. This drift mode…

Machine Learning · Computer Science 2023-11-02 David Haink

The Hessian of a neural network captures parameter interactions through second-order derivatives of the loss. It is a fundamental object of study, closely tied to various problems in deep learning, including model design, optimization, and…

Machine Learning · Computer Science 2021-07-02 Sidak Pal Singh , Gregor Bachmann , Thomas Hofmann

Characterizing the loss of a neural network with respect to model parameters, i.e., the loss landscape, can provide valuable insights into properties of that model. Various methods for visualizing loss landscapes have been proposed, but…

Near an optimal learning point of a neural network, the learning performance of gradient descent dynamics is dictated by the Hessian matrix of the loss function with respect to the network parameters. We characterize the Hessian…

Machine Learning · Statistics 2025-12-18 Carlos Couto , José Mourão , Mário A. T. Figueiredo , Pedro Ribeiro

Loss landscape analysis is extremely useful for a deeper understanding of the generalization ability of deep neural network models. In this work, we propose a layerwise loss landscape analysis where the loss surface at every layer is…

Machine Learning · Computer Science 2020-12-09 Adepu Ravi Sankar , Yash Khasbage , Rahul Vigneswaran , Vineeth N Balasubramanian

We consider the optimization problem associated with fitting two-layers ReLU networks with respect to the squared loss, where labels are generated by a target network. We leverage the rich symmetry structure to analytically characterize the…

Machine Learning · Computer Science 2020-10-19 Yossi Arjevani , Michael Field

The energy landscape of multiverse cosmology is often modeled by a multi-dimensional random Gaussian potential. The physical predictions of such models crucially depend on the eigenvalue distribution of the Hessian matrix at potential…

High Energy Physics - Theory · Physics 2018-04-04 Masaki Yamada , Alexander Vilenkin

We empirically demonstrate that full-batch gradient descent on neural network training objectives typically operates in a regime we call the Edge of Stability. In this regime, the maximum eigenvalue of the training loss Hessian hovers just…

Machine Learning · Computer Science 2022-11-24 Jeremy M. Cohen , Simran Kaur , Yuanzhi Li , J. Zico Kolter , Ameet Talwalkar

Recent progress has been made in understanding optimisation dynamics in neural networks trained with full-batch gradient descent with momentum with the uncovering of the edge of stability phenomenon in supervised learning. The edge of…

Machine Learning · Computer Science 2023-07-11 Rares Iordan , Marc Peter Deisenroth , Mihaela Rosca

We show that learning can be improved by using loss functions that evolve cyclically during training to emphasize one class at a time. In underparameterized networks, such dynamical loss functions can lead to successful training for…

Machine Learning · Computer Science 2021-06-24 Miguel Ruiz-Garcia , Ge Zhang , Samuel S. Schoenholz , Andrea J. Liu

We apply state-of-the-art tools in modern high-dimensional numerical linear algebra to approximate efficiently the spectrum of the Hessian of modern deepnets, with tens of millions of parameters, trained on real data. Our results…

Machine Learning · Computer Science 2019-06-04 Vardan Papyan

Neural networks (NNs) are central to modern machine learning and achieve state-of-the-art results in many applications. However, the relationship between loss geometry and generalization is still not well understood. The local geometry of…

Machine Learning · Computer Science 2026-04-15 Yuto Omae , Kazuki Sakai , Yohei Kakimoto , Makoto Sasaki , Yusuke Sakai , Hirotaka Takahashi

We establish universal Gaussian fluctuations for the mesoscopic linear eigenvalue statistics in the vicinity of the cusp-like singularities of the limiting spectral density for Wigner-type random matrices. Prior to this work, the linear…

Probability · Mathematics 2023-08-25 Volodymyr Riabov

Recent empirical and theoretical work has shown that the dynamics of the large eigenvalues of the training loss Hessian have some remarkably robust features across models and datasets in the full batch regime. There is often an early period…

Machine Learning · Computer Science 2025-02-04 Atish Agarwala , Jeffrey Pennington
‹ Prev 1 2 3 10 Next ›