Early Stopping

Follow this topic Ask about this topic

Early stopping studies how to choose when to end iterative learning or computation: continuing can overfit or waste resources, while stopping too soon can harm accuracy. For training, theory treats stopped gradient descent as implicit spectral regularization, while validation losses, discrepancy tests, complexity bounds, and sequential statistics help select checkpoints or budgets. At inference, confidence, answer stability, hidden-state probes, and semantic redundancy can help reasoning models, agents, and generation systems exit early; related rules also allocate evaluation and optimization effort. A central tension is detecting diminishing returns reliably without stopping prematurely or spending more on probes and validation than the saved computation is worth; results depend on the data, task, and serving setup.

Latest papers

46 papers7 papers in the last four weeks, up 17% on the four weeks before. 0.0% of all new papers.

When is early stopping beneficial under label noise? A phase diagram for kernel gradient flow
Oct 24, 2026Fabrizio Boncoraglio, Lenka ZdeborováEarly Stopping Learning with Noisy Labels
Statistical Physics of Computation Laboratory, École Polytechnique Fédérale de Lausanne (EPFL) CH-1015 Lausanne
Same Beliefs, Different Policies: Learning Reusable Representations for Early Stopping in LLM Reasoning
Oct 21, 2026Jingwei Ni, Dmitry Abulkhanov, Vladislav Smirnov +5Early Stopping for Language Model Reasoning Early Stopping
ETH Zürich · MBZUAI · University of Zürich +1
Answer Delivery Under a Shared Token Budget: An Audited Study of Stopping and Finalization
Oct 19, 2026Yupeng Chang, Linyuan Gao, Zhiyuan Li +2Efficient Language Model Inference Early Stopping
School of Artificial Intelligence, Jilin University · Engineering Research Center of Knowledge-Driven Human-Machine Intelligence, MOE, China · International Center of Future Science, Jilin University
StopLoss: Teaching Agents to Predict Failure and Stop Early
Oct 18, 2026Jiarong Liang, Xueguang Ma, Dongfu Jiang +3Early Stopping AI Agent Monitoring
University of Waterloo
Theory-Guided Instance-Dependent Early Stopping via Second-Order Loss Dynamics
Oct 17, 2026Ahad Bin Islam Shoeb, Jamal Uddin Tanvin, Kamrul Hasan +2Early Stopping Neural Network Optimization
University of Dhaka · Military Institute of Science and Technology · Tennessee State University +1
Local Epochs as Optimal Stopping: Theory and Evidence in Federated Learning
Oct 15, 2026Ke Xiao, Qiyuan Wang, Christos AnagnostopoulosFederated Learning Early Stopping
School of Computing Science, University of Glasgow
Revisiting Tabular Deep Regression: Objectives and the Limits of Global Early Stopping
Oct 9, 2026Danil Rudenko, Georgii Bulgakov, Nikolay Kartashev +3Tabular Regression Quantile Regression
Yandex Research · BRAInLab · HSE University
When Is Enough Enough in Self-Evolving LLM Systems?
Oct 3, 2026Enoch Yin, Bin Liu, Zhengling QiEarly Stopping Language Model Self-Improvement
Department of Statistics, The George Washington University · Department of Statistics and Data Science, School of Management, Fudan University · Department of Decision Sciences, School of Business, The George Washington University
The Model Knows When to Stop: Training-Free Early Stopping for Long-Context Reading
Sep 28, 2026Muath Alyobi, Mohamed Eltahir, Almoayyad Abuljdail +3Early Stopping Long-Context Language Model Evaluation
King Abdullah University of Science and Technology (KAUST), Thuwal, Saudi Arabia · Department of Computer Science, Edge Hill University, Ormskirk, England
Do We Need Complex Topology Control? Distinct-Peer Random Routing Improves Cost-Efficiency in Sparse Multi-Agent Debate
Sep 22, 2026Boxuan Wang, Zhuoyun Li, Xiaowei Huang +1Multi-Agent Debate Multi-Agent Communication Topology
School of Computer Science and Informatics, University of Liverpool, United Kingdom
A Smoothed Discrepancy Principle for Random Feature Methods and Neural Networks
Sep 17, 2026Mike Nguyen, Nicole MückeEarly Stopping Nonparametric Regression
Technical University of Braunschweig
Early-Stopping Thresholds for ES-HyperNEAT: A Data-Driven Approach from Fitness Dynamics
Sep 11, 2026Romain Claret, Arthur Gygax, Michael O'Neill +2Early Stopping Evolutionary Optimization
University of Neuchâtel, Switzerland · Information Management Institute · Computer Science Department +1
Stable Answers, Unfinished Reasoning: Why Self-Consensus Is Not a Safe Early-Exit Signal
Sep 9, 2026Yunxiang Mo, Donghao Zhao, Hejia GengEarly Stopping LLM Reliability
HKUST · University of Oxford
Quit While You're Ahead: Quit for Efficient Candidate Generation in Machine Translation Reranking
Sep 1, 2026Guangyu Chen, Boxuan Lyu, Hidetaka Kamigaito +2Early Stopping Machine Translation
Institute of Science Tokyo · Nara Institute of Science and Technology
Adaptive Regularization for Random Features: A Neighboring Early-Stopping Rule with Oracle-Rate Guarantees
Aug 26, 2026Caixing Wang, Zhibo Chen, Yue WangRandom Feature Methods Kernel Ridge Regression
School of Statistics and Data Science, Southeast University · Faculty of Business for Science and Technology, School of Management, University of Science and Technology of China
A Data-dependent Early Stopping Rule using Rademacher Complexity with L1-norm
Aug 25, 2026Duy Hoang, Bastien Berret, Olivier Bruneau +1Early Stopping Neural Network Generalization
Universit´e Paris-Saclay, CNRS, ENS Paris-Saclay, LMF Gif-sur-Yvette, 91190, France · Universit´e Paris-Saclay, Inria, CIAMS Gif-sur-Yvette, 91190, France · Universit´e Paris-Saclay, ENS Paris-Saclay, LURPA Gif-sur-Yvette, 91190, France
AV-AIVAT: 74x Cheaper Agent Evaluation with Certified Anytime-Valid Stopping in Imperfect-Information Games
Aug 6, 2026Boning Li, Yu Chen, Longbo HuangEarly Stopping Imperfect-Information Games
IIIS, Tsinghua University
Minimax Optimal Early-Stopped Gradient Descent for Gaussian Mixture Classification
Aug 6, 2026Alex Buna, Shirley Xiaoqi Liu, Patrick RebeschiniEarly Stopping Gradient Descent
Department of Statistics, University of Oxford
Beyond Negative-Ridge Endpoints: Mixed-Sign Spectral Regularization via Negative-Shifted Gradient Descent
Jul 24, 2026Peng ZhaoSpectral Regularization Early Stopping
Department of Applied Economics and Statistics · University of Delaware
Stop Guessing When to Stop Testing: Efficient Model Evaluation with Just Enough Data
Jul 9, 2026Ofir Arviv, Kristjan Greenewald, Yotam Perlitz +3Early Stopping Optimal Stopping
IBM Research
An interpretable Good--Turing restart criterion for k-means++
Jul 9, 2026Renato Cordeiro de AmorimEarly Stopping Clustering
School of Computer Science and Electronic Engineering, University of Essex, Colchester, UK.
Doomed from the Start: Early Abort of LLM Agent Episodes via a Recall-Controlled Probe Cascade
Jul 7, 2026Kai Ruan, Zihe Huang, Ziqi Zhou +4LLM Inference Efficiency Early Stopping
Gaoling School of Artificial Intelligence, Renmin University of China · Institute of Computing Technology, Chinese Academy of Sciences · Duke University +3
AdaStop: Cost-Aware Early Stopping for DNN Test Selection
Jul 6, 2026Bonan Shen, Wei-Jung Huang, Xin Liu +2Early Stopping DNN Testing
Independent Researcher
When Does Learning to Stop Help? A Cost-Aware Study of Early Exits in Reasoning Models
Jun 29, 2026Zhe Dong, Fang Qin, Manish ShahEfficient Language Model Reasoning Cost-Aware Inference
University of Maine at Presque Isle · Stanford University · Independent Researcher
Stabilizing Extrapolation in Looped Transformers via Learned Stochastic Stopping
Jun 29, 2026Hsun-Yu Kuo, El Mahdi Chayti, Patrik Reizinger +2Early Stopping OOD Generalization
EPFL, Lausanne, Switzerland · Max Planck Institute for Intelligent Systems & ELLIS Institute, Tübingen, Germany
Halt Fast! Early Stopping for Certified Robustness
Jun 26, 2026Andrew C. Cullen, Paul Montague, Benjamin I. P. RubinsteinEfficient Neural Network Inference Early Stopping
University of Melbourne · DST Group, Adelaide
Semantic Early-Stopping for Iterative LLM Agent Loops
Jun 25, 2026Sahil ShrivastavaEarly Stopping LLM Agent Evaluation
Semantic Halting Problem (SHP) Project
The Geometry of Saturation: Effective Rank Predicts When Labels Stop Helping in Few-Shot Classification
Jun 12, 2026Arnav GuptaEarly Stopping Few-Shot Learning
Independent Researcher, Nepal
Quantitative Performance Analysis of Stopping Criteria for CMA-ES
Jun 8, 2026Ryoji TanabeEarly Stopping Evolutionary Optimization
Yokohama National University, Yokohama, Japan
EvalStop: Using World Feedback to Detect and Correct Reward Overoptimization in Multi-Tenant RLHF Platforms
Jun 2, 2026Guilin Zhang, Chuanyi Sun, Kai Zhao +3Reward Over-Optimization Early Stopping
Workday AI Research · The George Washington University, Washington, DC, USA.

Oct 24, 2026J/K move · Enter open · S save

When is early stopping beneficial under label noise? A phase diagram for kernel gradient flow

Fabrizio Boncoraglio, Lenka Zdeborová

Statistical Physics of Computation Laboratory, École Polytechnique Fédérale de Lausanne (EPFL) CH-1015 Lausanne

Early Stopping Learning with Noisy Labels Learning Curves

Open paper AskPDF arXiv

Early stopping is commonly used to prevent fitting label noise from degrading generalization. Does this mean that, when a positive fraction of the training labels is intrinsically unpredictable, the best test error must occur before interpolation? Experiments on a CIFAR10 classification task with randomly labeled examples motivate the opposite possibility: although fixed-feature and shallow models exhibit late-time deterioration of the test error, the need for early stopping essentially disappears in larger, deeper convolutional networks even after the random labels have all been classified correctly. We determine when late-time deterioration disappears entirely in a solvable setting. For kernel gradient flow on a synthetic noisy classification task, we derive sharp high-dimensional learning curves and obtain a phase diagram controlled by the balance between the linear kernel component that learns the shared relation between inputs and labels and the higher-order components that enable fitting the label noise. In one regime, generalization is optimized at finite time and early stopping is necessary. In another, generalization improves until interpolation, but the endpoint is suboptimal across kernels. At the boundary, the interpolating endpoint attains the lowest generalization error across the considered kernel family and all training times. Thus, even with extensive irreducible label noise, interpolation can outperform early stopping. The neural-network experiments leave open how learned representations acquire an analogous favorable geometry.