Copertina

Learning Shrinks the Hard Tail: Training-Dependent Inference Scaling in a Solvable Linear Model

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autore principale:	Levi, Noam
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Machine Learning Disordered Systems and Neural Networks Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

More Bang for the Buck: Improving the Inference of Large Language Models at a Fixed Budget using Reset and Discard (ReD)
di: Meir, Sagi, et al.
Pubblicazione: (2026)

Grokking in Linear Estimators -- A Solvable Model that Groks without Understanding
di: Levi, Noam, et al.
Pubblicazione: (2023)

Grokking at the Edge of Linear Separability
di: Beck, Alon, et al.
Pubblicazione: (2024)

Probing the Latent Hierarchical Structure of Data via Diffusion Models
di: Sclocchi, Antonio, et al.
Pubblicazione: (2024)

Scaling Laws and Spectra of Shallow Neural Networks in the Feature Learning Regime
di: Defilippis, Leonardo, et al.
Pubblicazione: (2025)

Enhancing Noise-Robust Losses for Large-Scale Noisy Data Learning
di: Staats, Max, et al.
Pubblicazione: (2023)

Predictive Coding Networks and Inference Learning: Tutorial and Survey
di: van Zwol, Björn, et al.
Pubblicazione: (2024)

Class Imbalance in Anomaly Detection: Learning from an Exactly Solvable Model
di: Pezzicoli, F. S., et al.
Pubblicazione: (2025)

A Scalable Measure of Loss Landscape Curvature for Analyzing the Training Dynamics of LLMs
di: Kalra, Dayal Singh, et al.
Pubblicazione: (2026)

Representation Learning on a Random Lattice
di: Brill, Aryeh
Pubblicazione: (2025)

Spectral Dynamics in Deep Networks: Feature Learning, Outlier Escape, and Learning Rate Transfer
di: Lauditi, Clarissa, et al.
Pubblicazione: (2026)

Applications of Statistical Field Theory in Deep Learning
di: Ringel, Zohar, et al.
Pubblicazione: (2025)

Grokking vs. Learning: Same Features, Different Encodings
di: Manning-Coe, Dmitry, et al.
Pubblicazione: (2025)

Parameter Symmetry Potentially Unifies Deep Learning Theory
di: Ziyin, Liu, et al.
Pubblicazione: (2025)

Quantifying Hyperparameter Transfer and the Importance of Embedding Layer Learning Rate
di: Kalra, Dayal Singh, et al.
Pubblicazione: (2026)

A Geometric Perspective on the Difficulties of Learning GNN-based SAT Solvers
di: Skenderi, Geri
Pubblicazione: (2025)

Connecting NTK and NNGP: A Unified Theoretical Framework for Wide Neural Network Learning Dynamics
di: Avidan, Yehonatan, et al.
Pubblicazione: (2023)

Sampling Data with Chains of Forward-Backward Diffusion Steps
di: Kang, Hyunmo, et al.
Pubblicazione: (2026)

The Underlying Scaling Laws and Universal Statistical Structure of Complex Datasets
di: Levi, Noam, et al.
Pubblicazione: (2023)

Scaling Laws Do Not Scale
di: Diaz, Fernando, et al.
Pubblicazione: (2023)

On the origin of neural scaling laws: from random graphs to natural language
di: Barkeshli, Maissam, et al.
Pubblicazione: (2026)

Generalization through variance: how noise shapes inductive biases in diffusion models
di: Vastola, John J.
Pubblicazione: (2025)

Identifying internal patterns in (1+1)-dimensional directed percolation using neural networks
di: Parkhomenko, Danil, et al.
Pubblicazione: (2025)

A Spin Glass Characterization of Neural Networks
di: Li, Jun
Pubblicazione: (2025)

A method for quantifying the generalization capabilities of generative models for solving Ising models
di: Ma, Qunlong, et al.
Pubblicazione: (2024)

KAN: Kolmogorov-Arnold Networks
di: Liu, Ziming, et al.
Pubblicazione: (2024)

How Do Transformers "Do" Physics? Investigating the Simple Harmonic Oscillator
di: Kantamneni, Subhash, et al.
Pubblicazione: (2024)

The Persian Rug: solving toy models of superposition using large-scale symmetries
di: Cowsik, Aditya, et al.
Pubblicazione: (2024)

Towards Distributed Neural Architectures
di: Cowsik, Aditya, et al.
Pubblicazione: (2025)

Smooth Kolmogorov Arnold networks enabling structural knowledge representation
di: Samadi, Moein E., et al.
Pubblicazione: (2024)

Preisach Attention: A Hysteretic Model of Sequential Memory
di: Frydrych, Piotr
Pubblicazione: (2026)

Predictive Coding Graphs are a Superset of Feedforward Neural Networks
di: van Zwol, Björn
Pubblicazione: (2026)

High-Dimensional Learning Dynamics of Quantized Models with Straight-Through Estimator
di: Ichikawa, Yuma, et al.
Pubblicazione: (2025)

Exact Learning Dynamics of In-Context Learning in Linear Transformers and Its Application to Non-Linear Transformers
di: Mainali, Nischal, et al.
Pubblicazione: (2025)

Training Dynamics of Nonlinear Contrastive Learning Model in the High Dimensional Limit
di: Meng, Lineghuan, et al.
Pubblicazione: (2024)

Theory of Optimal Learning Rate Schedules and Scaling Laws for a Random Feature Model
di: Bordelon, Blake, et al.
Pubblicazione: (2026)

Approximation Theory for Neural Networks: Old and New
di: Mukherjee, Soumendu Sundar, et al.
Pubblicazione: (2026)

Nature-Inspired Local Propagation
di: Betti, Alessandro, et al.
Pubblicazione: (2024)

Deep Linear Network Training Dynamics from Random Initialization: Data, Width, Depth, and Hyperparameter Transfer
di: Bordelon, Blake, et al.
Pubblicazione: (2025)

Learning Linear Regression with Low-Rank Tasks in-Context
di: Takanami, Kaito, et al.
Pubblicazione: (2025)