mlsanityData to 14 Oct 2026

Implicit Target Shift in Online Learning: Characterization and Correction

arXiv 2605.07886 ↗First posted 8 May 2026stat.ML cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Online learning from a stream of data is a defining feature of intelligence, yet modern machine learning systems often struggle in this setting, especially under distributional shift. To understand its basic properties, we study the relationship between online and offline learning in the context of kernel regression by deriving a closed-form expression for the function learned by online kernel regression. We reveal that online kernel regression is equivalent to offline regression with shifted, inaccurate target outputs. Conversely, we show that by compensating for this implicit target shift in the teaching signal through target correction, online kernel-based learning can provably learn the same predictor as its offline counterpart. We derive both a closed-form expression for this target correction and an iterative form that can be applied sequentially. Applying this framework to continual image classification tasks on domain-incremental Split CIFAR-10 and CORe50, we show that online stochastic gradient descent with iteratively corrected targets outperforms learning with the true targets. This work therefore provides a basic framework for analyzing and improving online learning in non-stationary environments from the perspective of implicit target shift.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 36 of its 72 references that are in the map. First posted on arXiv 8 May 2026.

Builds on

36 of 72 references.

References · 36

  1. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  2. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,475
  3. 2016Overcoming catastrophic forgetting in neural networkscited by 4,787
  4. 2018Neural Tangent Kernel: Convergence and Generalization in Neural Networkscited by 2,621
  5. 2016iCaRL: Incremental Classifier and Representation Learningcited by 2,333
  6. 2017Gradient Episodic Memory for Continual Learningcited by 1,898
  7. 2017Understanding Black-box Predictions via Influence Functionscited by 1,741
  8. 2016On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minimacited by 1,352
  9. 2019When Does Label Smoothing Help?cited by 988
  10. 2020Dark Experience for General Continual Learning: a Strong, Simple Baselinecited by 911
  11. 2019Wide Neural Networks of Any Depth Evolve as Linear Models Under Gradient Descentcited by 738
  12. 2018Overcoming catastrophic forgetting with hard attention to the taskcited by 681
24 more
  1. 2023An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-tuningcited by 639
  2. 2019On Exact Computation with an Infinitely Wide Neural Netcited by 573
  3. 2019Gradient based sample selection for online continual learningcited by 571
  4. 2015Deep Kernel Learningcited by 445
  5. 2017CORe50: a New Dataset and Benchmark for Continuous Object Recognitioncited by 289
  6. 2021Online Continual Learning in Image Classification: An Empirical Surveycited by 245
  7. 2023Grounding Large Language Models in Interactive Environments with Online Reinforcement Learningcited by 195
  8. 2020Deep learning versus kernel learning: an empirical study of loss landscape geometry and the time evolution of the Neural Tangent Kernelcited by 187
  9. 2013Non-stationary Stochastic Optimizationcited by 175
  10. 2019Neural Tangents: Fast and Easy Infinite Neural Networks in Pythoncited by 149
  11. 2020Implicit Gradient Regularizationcited by 146
  12. 2020On the linearity of large non-linear models: when and why the tangent kernel is constantcited by 129
  13. 2022Class-Incremental Continual Learning into the eXtended DER-versecited by 124
  14. 2021Continual World: A Robotic Benchmark For Continual Reinforcement Learningcited by 116
  15. 2020A Theoretical Analysis of Catastrophic Forgetting through the NTK Overlap Matrixcited by 103
  16. 2015Gaussian Processes for Data-Efficient Learning in Robotics and Controlcited by 98
  17. 2021Continual Learning in the Teacher-Student Setup: Impact of Task Similaritycited by 88
  18. 2022How catastrophic can catastrophic forgetting be in linear regression?cited by 81
  19. 2023Theory on Forgetting and Generalization of Continual Learningcited by 72
  20. 2020Generalisation Guarantees for Continual Learning with Orthogonal Gradient Descentcited by 68
  21. 2018An Overview of Machine Teachingcited by 63
  22. 2014Scalable Kernel Methods via Doubly Stochastic Gradientscited by 36
  23. 2023The Ideal Continual Learner: An Agent That Never Forgetscited by 35
  24. 2025Learning Curves of Stochastic Gradient Descent in Kernel Regressioncited by 4

Built on by

Not cited yet.