Image de couverture de livre

Using physics-inspired Singular Learning Theory to understand grokking & other phase transitions in modern neural networks

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteur principal:	Lakkapragada, Anish
Format:	Preprint
Publié:	2025
Sujets:	Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

A Novel Explanation Against Linear Neural Networks
par: Lakkapragada, Anish
Publié: (2023)

Exponential Family Discriminant Analysis: Generalizing LDA-Style Generative Classification to Non-Gaussian Models
par: Lakkapragada, Anish
Publié: (2026)

A rationale from frequency perspective for grokking in training neural network
par: Zhou, Zhangchen, et autres
Publié: (2024)

Comprehensive OOD Detection Improvements
par: Lakkapragada, Anish, et autres
Publié: (2024)

How to explain grokking
par: Kozyrev, S. V.
Publié: (2024)

Learning words in groups: fusion algebras, tensor ranks and grokking
par: Shutman, Maor, et autres
Publié: (2025)

Emergence in non-neural models: grokking modular arithmetic via average gradient outer product
par: Mallinar, Neil, et autres
Publié: (2024)

CosmicFish-HRM: Adaptive Reasoning via Hierarchical Recurrent Mechanisms in Compact Language Models
par: Lakkapragada, Venkat Akhil
Publié: (2026)

Detecting overfitting in Neural Networks during long-horizon grokking using Random Matrix Theory
par: Prakash, Hari K., et autres
Publié: (2026)

Late-Stage Generalization Collapse in Grokking: Detecting anti-grokking with Weightwatcher
par: Prakash, Hari K, et autres
Publié: (2026)

Quadratic models for understanding catapult dynamics of neural networks
par: Zhu, Libin, et autres
Publié: (2022)

Differential equation and probability inspired graph neural networks for latent variable learning
par: Shi, Zhuangwei
Publié: (2022)

Order parameters and phase transitions of continual learning in deep neural networks
par: Shan, Haozhe, et autres
Publié: (2024)

A quatum inspired neural network for geometric modeling
par: Du, Weitao, et autres
Publié: (2024)

Theory of periodic convolutional neural network
par: Liu, Yuqing
Publié: (2025)

Learning topological defects formation with neural networks in a quantum phase transition
par: Shi, Han-Qing, et autres
Publié: (2022)

A Kaczmarz-inspired approach to accelerate the optimization of neural network wavefunctions
par: Goldshlager, Gil, et autres
Publié: (2024)

Energy-convergence trade off for the training of neural networks on bio-inspired hardware
par: Garg, Nikhil, et autres
Publié: (2025)

Visualizing the loss landscapes of physics-informed neural networks
par: Rowan, Conor, et autres
Publié: (2026)

Enforcing hidden physics in physics-informed neural networks
par: Chen, Nanxi, et autres
Publié: (2025)

Sharp-PINNs: staggered hard-constrained physics-informed neural networks for phase field modelling of corrosion
par: Chen, Nanxi, et autres
Publié: (2025)

Strategies for training point distributions in physics-informed neural networks
par: Humagain, Santosh, et autres
Publié: (2025)

To be or not to be stable, that is the question: understanding neural networks for inverse problems
par: Evangelista, Davide, et autres
Publié: (2022)

$Δ$-PINNs: physics-informed neural networks on complex geometries
par: Costabal, Francisco Sahli, et autres
Publié: (2022)

Dilated convolutional neural network for detecting extreme-mass-ratio inspirals
par: Zhao, Tianyu, et autres
Publié: (2023)

Quality-factor inspired deep neural network solver for solving inverse scattering problems
par: Du, Yutong, et autres
Publié: (2025)

Precipitation nowcasting of satellite data using physically-aligned neural networks
par: Catão, Antônio, et autres
Publié: (2025)

Progressive multi-fidelity learning with neural networks for physical system predictions
par: Conti, Paolo, et autres
Publié: (2025)

Engineering application of physics-informed neural networks for Saint-Venant torsion
par: Jo, Su Yeong, et autres
Publié: (2025)

An efficient wavelet-based physics-informed neural network for multiscale problems
par: Pandey, Himanshu, et autres
Publié: (2024)

A Basin-Selection Perspective on Grokking via Singular Learning Theory
par: Cullen, Ben, et autres
Publié: (2026)

PAC-Bayes Bounds for Gibbs Posteriors via Singular Learning Theory
par: Wang, Chenyang, et autres
Publié: (2026)

Training deep physical neural networks with local physical information bottleneck
par: Wang, Hao, et autres
Publié: (2026)

Pseudo-differential-enhanced physics-informed neural networks
par: Gracyk, Andrew
Publié: (2026)

Learning solutions of parametric Navier-Stokes with physics-informed neural networks
par: Naderibeni, M., et autres
Publié: (2024)

GaborPINN: Efficient physics informed neural networks using multiplicative filtered networks
par: Huang, Xinquan, et autres
Publié: (2023)

Learning characteristic parameters and dynamics of centrifugal pumps under multiphase flow using physics-informed neural networks
par: Carvalho, Felipe de Castro Teixeira, et autres
Publié: (2023)

Compressibility Measures Complexity: Minimum Description Length Meets Singular Learning Theory
par: Urdshals, Einar, et autres
Publié: (2025)

Residual resampling-based physics-informed neural network for neutron diffusion equations
par: Zhang, Heng, et autres
Publié: (2024)

Phase transitions reveal hierarchical structure in deep neural networks
par: Ersoy, Ibrahim Talha, et autres
Publié: (2025)