Imagen de Portada

A Thermodynamic Theory of Learning Part II: Critical Period Closure and Continual Learning Failure

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autor principal:	Okanohara, Daisuke
Formato:	Preprint
Publicado:	2026
Materias:	Machine Learning
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

A Thermodynamic Theory of Learning I: Irreversible Ensemble Transport and Epistemic Costs
por: Okanohara, Daisuke
Publicado: (2026)

Speed-accuracy relations for diffusion models: Wisdom from nonequilibrium thermodynamics and optimal transport
por: Ikeda, Kotaro, et al.
Publicado: (2024)

Generative Model for Constructing Reaction Path from Initial to Final States
por: Hayashi, Akihide, et al.
Publicado: (2024)

One Period to Rule Them All: Identifying Critical Learning Periods in Deep Networks
por: Fukase, Vinicius Yuiti, et al.
Publicado: (2025)

On the Occurence of Critical Learning Periods in Neural Networks
por: Pawlak, Stanisław
Publicado: (2025)

Machine Learning for Improved Density Functional Theory Thermodynamics
por: Simak, Sergei I., et al.
Publicado: (2025)

Refining Machine Learning Potentials through Thermodynamic Theory of Phase Transitions
por: Fuchs, Paul, et al.
Publicado: (2025)

Critical Learning Periods Emerge Even in Deep Linear Networks
por: Kleinman, Michael, et al.
Publicado: (2023)

Theory on Mixture-of-Experts in Continual Learning
por: Li, Hongbo, et al.
Publicado: (2024)

From Physics to Machine Learning and Back: Part II - Learning and Observational Bias in PHM
por: Fink, Olga, et al.
Publicado: (2025)

When Does Metadata Conditioning (NOT) Work for Language Model Pre-Training? A Study with Context-Free Grammars
por: Higuchi, Rei, et al.
Publicado: (2025)

Critical Patch-Aware Sparse Prompting with Decoupled Training for Continual Learning on the Edge
por: Lim, Wonseon, et al.
Publicado: (2026)

Hierarchical Inference and Closure Learning via Adaptive Surrogates for ODEs and PDEs
por: Zhang, Pengyu, et al.
Publicado: (2026)

Broad Critic Deep Actor Reinforcement Learning for Continuous Control
por: Thalagala, Shiron, et al.
Publicado: (2024)

A Statistical Theory of Regularization-Based Continual Learning
por: Zhao, Xuyang, et al.
Publicado: (2024)

Information as Structural Alignment: A Dynamical Theory of Continual Learning
por: Negulescu, Radu
Publicado: (2026)

Differentiable Thermodynamic Phase-Equilibria for Machine Learning
por: Hicham, Karim K. Ben, et al.
Publicado: (2026)

Stochastic Thermodynamics of Learning Parametric Probabilistic Models
por: Parsi, Shervin Sadat
Publicado: (2023)

Thermodynamics of Reinforcement Learning Curricula
por: Adamczyk, Jacob, et al.
Publicado: (2026)

The Machine Learning Approach to Moment Closure Relations for Plasma: A Review
por: Burles, Samuel, et al.
Publicado: (2025)

Analysis of Diagnostics (Part II): Prevalence, Linear Independence, and Unsupervised Learning
por: Patrone, Paul N., et al.
Publicado: (2024)

On the Theory of Continual Learning with Gradient Descent for Neural Networks
por: Taheri, Hossein, et al.
Publicado: (2025)

Predicting Channel Closures in the Lightning Network with Machine Learning
por: Antonelli, Simone, et al.
Publicado: (2026)

Quantile Online Learning for Semiconductor Failure Analysis
por: Zhou, Bangjian, et al.
Publicado: (2023)

Inferential Mechanics Part 1: Causal Mechanistic Theories of Machine Learning in Chemical Biology with Implications
por: Balabin, Ilya, et al.
Publicado: (2026)

Periodic Regularized Q-Learning
por: Yang, Hyukjun, et al.
Publicado: (2026)

Monte Carlo Beam Search for Actor-Critic Reinforcement Learning in Continuous Control
por: Alzorgan, Hazim, et al.
Publicado: (2025)

A Theory of Machine Learning
por: Kim, Jinsook, et al.
Publicado: (2024)

Dissecting the Failure of Invariant Learning on Graphs
por: Wang, Qixun, et al.
Publicado: (2024)

Statistical Theory of Multi-stage Newton Iteration Algorithm for Online Continual Learning
por: Lu, Xinjia, et al.
Publicado: (2025)

A Study of Continual Learning Methods for Q-Learning
por: Bagus, Benedikt, et al.
Publicado: (2022)

Machine Learning Coarse-Grained Potentials of Protein Thermodynamics
por: Majewski, Maciej, et al.
Publicado: (2022)

Neural Functions for Learning Periodic Signal
por: Cho, Woojin, et al.
Publicado: (2025)

Reinforcement Learning Closures for Underresolved Partial Differential Equations using Synthetic Data
por: Heimbach, Lothar, et al.
Publicado: (2025)

Boltzmann Classifier: A Thermodynamic-Inspired Approach to Supervised Learning
por: Amin, Muhamed, et al.
Publicado: (2025)

A Theory of Generalization in Deep Learning
por: Litman, Elon, et al.
Publicado: (2026)

Mathematics of Continual Learning
por: Peng, Liangzu, et al.
Publicado: (2025)

Continual Learning on Graphs: A Survey
por: Tian, Zonggui, et al.
Publicado: (2024)

Physics-Informed Neural Networks for Predicting Hydrogen Sorption in Geological Formations: Thermodynamically Constrained Deep Learning Integrating Classical Adsorption Theory
por: Nooraiepour, Mohammad, et al.
Publicado: (2026)

Learning from Failures: Understanding LLM Alignment through Failure-Aware Inverse RL
por: Patel, Nyal, et al.
Publicado: (2025)