Everywhere Learning: Artificial Intelligence with Pointwise Constraints
Not on the map. Why?
Abstract (arXiv)
Everywhere learning is a new paradigm whereby Artificial Intelligence (AI) systems are trained to satisfy loss constraints with probability one over the data distribution. This is in contrast to the standard paradigm of training AI systems to minimize average losses. We develop an approximate duality theory to substantiate a generalization analysis that establishes the proximity between solutions of empirical and statistical everywhere learning problems. Our results show that dual variables reweigh the data distribution towards points in which loss constraints are more difficult to satisfy and that generalization is controlled by the mismatch between the concentration of mass of the data distribution and the concentration of mass on points where constraints are more difficult to satisfy. We further show that we can control generalization with a sparse L1 penalty on constraint relaxations. We illustrate the merits of everywhere learning with an experiment in agentic classification for language model tasks.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 24 of its 50 references that are in the map. First posted on arXiv 1 Jun 2026.
Builds on
24 of 50 references.
References · 24
- 2021Evaluating Large Language Models Trained on Codecited by 8,320
- 2021Program Synthesis with Large Language Modelscited by 3,637
- 2017A Survey on Multi-Task Learningcited by 840
- 2018Multi-Task Learning as Multi-Objective Optimizationcited by 821
- 2024AFlow: Automating Agentic Workflow Generationcited by 432
- 2020Multi-Task Learning with Deep Neural Networks: A Surveycited by 375
- 2018Fairness Without Demographics in Repeated Loss Minimizationcited by 364
- 2020Robust Learning Through Cross-Task Consistencycited by 105
- 2020Probably Approximately Correct Constrained Learningcited by 47
- 2020Train simultaneously, generalize better: Stability of gradient-based minimax learnerscited by 44
- 2021Constrained Learning with Non-Convex Lossescited by 42
- 2020Generalization Bounds for Stochastic Saddle Point Problemscited by 28
12 more
- 2025GNNs as Predictors of Agentic Workflow Performancescited by 17
- 2021Multi-Objective Learning to Predict Pareto Fronts Using Hypervolume Maximizationcited by 16
- 2022What is a Good Metric to Study Generalization of Minimax Learners?cited by 16
- 2014Robust Optimization using Machine Learning for Uncertainty Setscited by 14
- 2025Composition and Alignment of Diffusion Models using Constrained Learningcited by 10
- 2024Near-Optimal Solutions of Constrained Learning Problemscited by 9
- 2025Alignment of large language models with constrained learningcited by 8
- 2025Feasible Learningcited by 5
- 2018Functional Nonlinear Sparse Modelscited by 4
- 2022Strong Duality in Risk-Constrained Nonconvex Functional Programmingcited by 4
- 2025AL-CoLe: Augmented Lagrangian for Constrained Learningcited by 2
- 2025Learning Optimal Power Flow with Pointwise Constraintscited by 1
Built on by
Citing papers per year: 2026: 1.
1 paper.
Similar, not cited
Close in wording, no citation either way
- 2026Do World Models Learn Global Understanding?
- 2026Adaptivity Under Realizability Constraints: Comparing In-Context and Agentic Learning
- 2025On the creation of narrow AI: hierarchy and nonlocality of neural network skills
- 2026Generalization in LLM Problem Solving: The Case of the Shortest Path
- 2025Bound by semanticity: universal laws governing the generalization-identification tradeoff
- 2025Task Generalization With AutoRegressive Compositional Structure: Can Learning From D Tasks Generalize to DT Tasks?
- 2025The World Is Bigger! A Computationally-Embedded Perspective on the Big World Hypothesis
- 2024REGENT: A Retrieval-Augmented Generalist Agent That Can Act In-Context in New Environments