arXiv：cs.LG· Ignacio Boero, Ignacio Hounie, Luiz F. O. Chamon, Alejandro Ribeiro·· 5 小时前AI 评分30
Everywhere Learning：带逐点约束的 AI 训练新范式
Everywhere Learning: Artificial Intelligence with Pointwise Constraints
AI 导读
研究者提出"Everywhere Learning"新范式，让 AI 系统在数据分布上以概率一满足损失约束，而非传统的最小化平均损失。他们建立近似对偶理论，证明对偶变量会将数据分布向约束更难满足的点重新加权，且泛化由数据分布质量集中与难满足点质量集中之间的错配所控制，并可用稀疏 L1 惩罚控制泛化。该工作在语言模型任务的智能体分类实验中验证了其优势。
正文
Abstract: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.
| Subjects: | Machine Learning (cs.LG); Signal Processing (eess.SP) |
| Cite as: | arXiv:2606.01557 [cs.LG] |
| (or arXiv:2606.01557v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2606.01557 arXiv-issued DOI via DataCite |
Submission history
From: Ignacio Boero [view email]
[v1]
Mon, 1 Jun 2026 02:02:22 UTC (1,042 KB)
[v2]
Fri, 2 Oct 2026 14:52:15 UTC (848 KB)
来源：arXiv：cs.LG · arxiv.org