相同行为、不同机制：有限实验能对整合信息证明什么
Same Behavior, Different Mechanism: What Finite Experiments Can Certify About Integrated Information
AI 导读
一项 arXiv 研究分析有限干预实验何时能识别、连续估计并证明机制级量，以整合信息理论的 φ 为例在 4、5 单元随机网络中给出精确答案。行为相同的前馈与循环网络 φ 分别为 0 和 0.28 比特，但内部记录在 4-53 次试验内即可区分；部分记录带来 0.30-1.02 比特的精度下限，且约 1% 的代理模型会误报 φ 超过 20%。研究表明可测性是可计算的设计问题。
正文 · 原文
Abstract:Interpretability, AI auditing and assessments of machine consciousness increasingly rest on mechanism-level quantities: functionals of a system's causal transition law that describe how it computes, not what it outputs. Experiments see that law only through finite, partial and noisy interventions. We ask when such a quantity is identifiable from an experiment's access, continuous in the mechanism and certifiable from finite data, and answer exactly for a deliberately demanding case - the integrated information $\varphi$ of integrated information theory - in stochastic networks of four and five units with known ground truth. Behavior can be silent: a recurrent and a feed-forward network with identical input-output laws at their response delay carry 0.28 and 0 bits, yet internal recordings separate them within 4-53 trials. Partial access can be just as blind: in our examples, units left out of a one-step recording impose precision floors of 0.30-1.02 bits that no sample size lowers, and in all-but-self parity networks, recordings of any length see several hidden units only through the product of their fidelities - a symmetry that one intervention broke in exact checks. The measure has sharp edges: because $\varphi$ selects a partition, and the cause and effect it specifies, by one criterion and reports another, infinitesimal changes of mechanism that cross an edge change it by 25-100% at 99% of the edges we found in random networks, and about 1% of surrogate models that fit interventional data almost perfectly misreport it by more than 20%. One-sided claims survive: the same data certify a lower bound on both sides of an edge, and we prove $\varphi>1.004$ bits throughout a 64-dimensional region of mechanisms and certify it from about $10^7$ controlled transitions with false-certificate probability at most $\alpha$. Measurability is thus a design question with computable answers.
| Comments: | 30 pages, 7 figures. Code available as ancillary files |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Neurons and Cognition (q-bio.NC) |
| Cite as: | arXiv:2611.08469 [cs.LG] |
| (or arXiv:2611.08469v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2611.08469 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Linqian Wu [view email]
[v1]
Thu, 5 Nov 2026 18:55:31 UTC (704 KB)
来源：arXiv cs.AI · arxiv.org