No article text was recorded for this version.
signal-noise weight decomposition
Version 1 of 141 papers42 conceptsfrozen 2026-10-04
Papers in this version
- Blog: Gradient Descent Robustly Learns the Intrinsic Dimension of Data in Training Convolutional Neural Networks
- Blog: Quantifying the Optimization and Generalization Advantages of Graph Neural Networks Over Multilayer Perceptrons
- Blog: Provable Benefits of Local Steps in Heterogeneous Federated Learning for Neural Networks: A Feature Learning Perspective
- Blog: Max-Margin Works while Large Margin Fails: Generalization without Uniform Convergence
- Blog: Provably Neural Active Learning Succeeds via Prioritizing Perplexing Samples
- Blog: Asymptotics of feature learning in two-layer networks after one gradient-step
- Blog: LLMs as Noisy Channels: A Shannon Perspective on Model Capacity and Scaling Laws
- Blog: Benign overfitting in leaky ReLU networks with moderate input dimension
- Blog: Task structure and nonlinearity jointly determine learned representational geometry
- Blog: Towards Understanding Generalization via Decomposing Excess Risk Dynamics
- Blog: Federated Learning from Vision-Language Foundation Models: Theoretical Analysis and Method
- Blog: SGD with memory: fundamental properties and stochastic acceleration
- Blog: Inferring learning rules from animal decision-making
- Blog: Coherent Gradients: An Approach to Understanding Generalization in Gradient Descent-based Optimization
- Blog: Benign Overfitting in Two-layer Convolutional Neural Networks
- Blog: The Ladder in Chaos: Improving Policy Learning by Harnessing the Parameter Evolving Path in A Low-dimensional Space
- Blog: Benign Overfitting in Adversarial Training for Vision Transformers
- Blog: Learning how to explain neural networks: PatternNet and PatternAttribution
- Blog: Understanding the Generalization of Stochastic Gradient Adam in Learning Neural Networks
- Blog: Graph Neural Networks Provably Benefit from Structural Information: A Feature Learning Perspective
- Blog: Understanding Private Learning From Feature Perspective
- Blog: Removing the Feature Correlation Effect of Multiplicative Noise
- Blog: Understanding the Generalization of Adam in Learning Neural Networks with Proper Regularization
- Blog: Why Does Sharpness-Aware Minimization Generalize Better Than SGD?
- Blog: Balancing Learning Rates Across Layers: Exact Two-Step Dynamics and Optimal Scaling in Linear Neural Networks
- Blog: On the Role of Label Noise in the Feature Learning Process
- Blog: Understanding Decoupled and Early Weight Decay
- Blog: On the Optimization and Generalization of Two-layer Transformers with Sign Gradient Descent
- Blog: On the Clean Generalization and Robust Overfitting in Adversarial Training from Two Theoretical Views: Representation Complexity and Training Dynamics
- Blog: Benign Overfitting in Two-layer ReLU Convolutional Neural Networks
- Blog: Effect of Ambient-Intrinsic Dimension Gap on Adversarial Vulnerability
- Blog: How Classifier Features Transfer to Downstream: An Asymptotic Analysis in a Two-Layer Model
- Blog: Initialization Matters: Unraveling the Impact of Pre-Training on Federated Learning
- Blog: In Defense of Uniform Convergence: Generalization via Derandomization with an Application to Interpolating Predictors
- Blog: Per-example gradient regularization improves learning signals from noisy data
- Blog: On the Feature Learning in Diffusion Models
- Blog: Why Does Pruning during Training Work? A Signal-to-Noise Analysis of Sparse Neural Network Training
- Blog: Exact and Rich Feature Learning Dynamics of Two-Layer Linear Networks
- Blog: Understanding Convergence and Generalization in Federated Learning through Feature Learning Theory
- Blog: iNNvestigate neural networks!
- Blog: How Does Label Noise Gradient Descent Improve Generalization in the Low SNR Regime?
History & citation
Current · Permanent link
- This direction split from Weight Decomposition for Understanding Neural Generalization (76% overlap)
Cite this direction
Latest version (a living page):
@misc{lacuna_signal_noise_weight_decomposition_73f6e877,
title = {signal-noise weight decomposition},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/signal-noise-weight-decomposition/rdr_73f6e8776ba54c948ea32aa0031e0401},
note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-21},
}This exact version (version 1, frozen text):
@misc{lacuna_signal_noise_weight_decomposition_73f6e877_v1,
title = {signal-noise weight decomposition},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/signal-noise-weight-decomposition/rdr_73f6e8776ba54c948ea32aa0031e0401/v/1},
note = {Version 1, frozen 2026-10-04. Accessed 2026-11-21},
}