Robust Explanatory Analysis for Structured Optimization Networks (REASON) in Learning

Hugging Face Daily Papers Papers
distribution-shift robustness inference optimization residual-error system-1-system-2

Summary

该论文提出 REASON 双循环推理框架，通过快速（System 1 式）和中速（System 2 式）循环分层解决残差误差，在分布偏移和未见扰动下以约一半计算量显著提升模型鲁棒性，并优于分级分类器和学习延迟方法。

We seek a hierarchical resolution of residual error under a composite learning objective as a principled mechanism for identifying and exploiting additional information, particularly under distribution shift and perturbation. Inspired by the System 1/System 2 paradigm, we formulate this principle as a two-loop mechanism serving a common inference goal. A fast, System 1-like loop resolves the residual addressable by the available evidence; the remaining epistemic residual triggers a medium speed, System 2-like loop that actively acquires additional evidence and measures its informational value through residual reduction. Over nested solution spaces, the medium-loop optimum is provably no worse than the fast-loop optimum. Experiments show that the resulting two-loop model consistently improves upon the fast loop at approximately half the computation of the medium loop, including under unseen corruptions, and, when provided a cleaner observation, outperforms staged classifiers and learning-to-defer methods at equal computational cost.
Original Article
View Cached Full Text

Cached at: 10/19/26, 11:31 PM

Paper page - Robust Explanatory Analysis for Structured Optimization Networks (REASON) in Learning

Source: https://huggingface.co/papers/2610.19378

Abstract

A two-loop inference framework resolves residual errors via fast and evidence-seeking loops to improve robustness under distribution shift with lower computation.

We seek a hierarchical resolution ofresidual errorunder acomposite learning objectiveas a principled mechanism for identifying and exploiting additional information, particularly underdistribution shiftand perturbation. Inspired by theSystem 1/System 2paradigm, we formulate this principle as atwo-loop mechanismserving a common inference goal. A fast, System 1-like loop resolves the residual addressable by the available evidence; the remainingepistemic residualtriggers a medium speed, System 2-like loop that actively acquires additional evidence and measures its informational value through residual reduction. Overnested solution spaces, the medium-loop optimum is provably no worse than the fast-loop optimum. Experiments show that the resulting two-loop model consistently improves upon the fast loop at approximately half the computation of the medium loop, including under unseen corruptions, and, when provided a cleaner observation, outperformsstaged classifiersandlearning-to-defermethods at equal computational cost.

View arXiv pageView PDFAdd to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.19378 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.19378 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.19378 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

REDaCT: Collapse-Free Unlearning in Large Reasoning Models

arXiv cs.LG

REDaCT提出一种面向大型推理模型的推理时无遗忘框架，通过二元泄漏检测器与纠正性嵌入优化，在彻底移除遗忘集信息的同时完全避免推理链崩溃，实现近乎完美的答案与CoT遗忘并显著保留推理能力。

Enhanced and Efficient Reasoning in Large Learning Models

arXiv cs.AI

This paper proposes a method for improving reasoning in large language models by recoding data to explicitly represent relationships, enabling efficient principled reasoning with polynomial-time learnability for relational rules, which addresses hallucinations and supports sound reasoning across multiple calls.