Cover Image

Spectral Edge Dynamics Reveal Functional Modes of Learning

Fuente: arXiv

Saved in:

Bibliographic Details
Main Author:	Xu, Yongzhong
Format:	Preprint
Published:	2026
Subjects:	Machine Learning Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Spectral Edge Dynamics of Training Trajectories: Signal--Noise Geometry Across Scales
by: Xu, Yongzhong
Published: (2026)

Spectral Edge Dynamics: An Analytical-Empirical Study of Phase Transitions in Neural Network Training
by: Xu, Yongzhong
Published: (2026)

Gradient-Direction Sensitivity Reveals Linear-Centroid Coupling Hidden by Optimizer Trajectories
by: Xu, Yongzhong
Published: (2026)

Feature Repulsion and Spectral Lock-in: An Empirical Study of Two-Layer Network Grokking
by: Xu, Yongzhong
Published: (2026)

Spectral Probe-Circuits: A Three-Step Recipe for Identifying Attention-Head Circuits in Pretrained Transformers
by: Xu, Yongzhong
Published: (2026)

Low-Dimensional and Transversely Curved Optimization Dynamics in Grokking
by: Xu, Yongzhong
Published: (2026)

Optimizer-Induced Low-Dimensional Drift and Transverse Dynamics in Transformer Training
by: Xu, Yongzhong
Published: (2026)

Global Low-Rank, Local Full-Rank: The Holographic Encoding of Learned Algorithms
by: Xu, Yongzhong
Published: (2026)

Low-Dimensional Execution Manifolds in Transformer Learning Dynamics: Evidence from Modular Arithmetic Tasks
by: Xu, Yongzhong
Published: (2026)

Early-Warning Signals of Grokking via Loss-Landscape Geometry
by: Xu, Yongzhong
Published: (2026)

The Geometry of Multi-Task Grokking: Transverse Instability, Superposition, and Weight Decay Phase Structure
by: Xu, Yongzhong
Published: (2026)

When Do Attention Circuits Form? Developmental Trajectories of Capability and Attention-Sink Emergence Across Three 1B-ClassArchitectures
by: Xu, Yongzhong
Published: (2026)

The Lifecycle of the Spectral Edge: From Gradient Learning to Weight-Decay Compression
by: Xu, Yongzhong
Published: (2026)

Revealing Interpretable Failure Modes of VLMs
by: Chaudhary, Isha, et al.
Published: (2026)

Dynamical Systems Analysis Reveals Functional Regimes in Large Language Models
by: Ugail, Hassan, et al.
Published: (2026)

Approaching Deep Learning through the Spectral Dynamics of Weights
by: Yunis, David, et al.
Published: (2024)

Learning Robust Spectral Dynamics for Temporal Domain Generalization
by: Yu, En, et al.
Published: (2025)

Spectral Invariant Learning for Dynamic Graphs under Distribution Shifts
by: Zhang, Zeyang, et al.
Published: (2024)

Spectral Theory for Edge Pruning in Asynchronous Recurrent Graph Neural Networks
by: Bessone, Nicolas
Published: (2025)

Shape-aware Graph Spectral Learning
by: Xu, Junjie, et al.
Published: (2023)

When Does Learning Renormalize? Sufficient Conditions for Power Law Spectral Dynamics
by: Zhang, Yizhou
Published: (2025)

From Order to Distribution: A Spectral Characterization of Forgetting in Continual Learning
by: Xu, Zonghuan, et al.
Published: (2026)

GraphEdge: Dynamic Graph Partition and Task Scheduling for GNNs Computing in Edge Network
by: Xiao, Wenjing, et al.
Published: (2025)

Addressing Spectral Bias of Deep Neural Networks by Multi-Grade Deep Learning
by: Fang, Ronglong, et al.
Published: (2024)

EdgeGFL: Rethinking Edge Information in Graph Feature Preference Learning
by: Zhuo, Shengda, et al.
Published: (2025)

Can Data-Driven Dynamics Reveal Hidden Physics? There Is A Need for Interpretable Neural Operators
by: Gao, Wenhan, et al.
Published: (2025)

DyEdgeGAT: Dynamic Edge via Graph Attention for Early Fault Detection in IIoT Systems
by: Zhao, Mengjie, et al.
Published: (2023)

Dynamic Spectral Backpropagation for Efficient Neural Network Training
by: Muthuraman, Mannmohan
Published: (2025)

Graph Learning with Distributional Edge Layouts
by: Zhao, Xinjian, et al.
Published: (2024)

Influence Functions for Edge Edits in Non-Convex Graph Neural Networks
by: Heo, Jaeseung, et al.
Published: (2025)

Rethinking Spectral Augmentation for Contrast-based Graph Self-Supervised Learning
by: Jian, Xiangru, et al.
Published: (2024)

Pattern-Matching Dynamic Memory Network for Dual-Mode Traffic Prediction
by: Weng, Wenchao, et al.
Published: (2024)

Diffusion Spectral Representation for Reinforcement Learning
by: Shribak, Dmitry, et al.
Published: (2024)

Spectral Representation-based Reinforcement Learning
by: Gao, Chenxiao, et al.
Published: (2025)

Symmetry Reveals Layerwise Dynamics: How Transformers Perform In-Context Classification
by: Lutz, Patrick, et al.
Published: (2026)

DynMuon: A Dynamic Spectral Shaping View of Muon
by: Wu, Fangzhou, et al.
Published: (2026)

Compressing Chemistry Reveals Functional Groups
by: Sharma, Ruben, et al.
Published: (2025)

Scalable Meta-Learning via Mixed-Mode Differentiation
by: Kemaev, Iurii, et al.
Published: (2025)

Lightweight Edge Learning via Dataset Pruning
by: Ale, Laha, et al.
Published: (2026)

Belief Dynamics Reveal the Dual Nature of In-Context Learning and Activation Steering
by: Bigelow, Eric, et al.
Published: (2025)