预测瓶颈不会发现因果结构(但它们实际上做了什么)
Prediction Bottlenecks Don't Discover Causal Structure (But Here's What They Actually Do)
AI总结:
研究通过实验证明,预测模型中的瓶颈无法发现因果结构,但在特定条件下仍表现出一定的干预效果,主要贡献是提出了可复用的验证基准。
摘要:
一个仅用于下一步预测的Mamba状态空间模型似乎通过简单的读出$S = |W_{out} W_{in}|$恢复了格兰杰因果结构,早期实验表明该现象在不同架构中普遍,并在$p < 10^{-5}$时受益于干预数据。我们包装了用于测试该主张的协议——标准化合成生成器(VAR/洛伦兹/CauseMe式)、三种干预语义($do(X=c)$、软噪声、随机强迫)、三个真实数据集上的边来源卡片,以及大小匹配的对照组——作为可重用的验证基准,并在五个阶段中检验该主张。方法层面的主张未能通过:(i)简单的线性瓶颈同样表现良好或更优;(ii)在合成CauseMe式基准和洛伦兹96(唯一具有明确地面真实性的现实基准)上,调优的Lasso在瓶颈之上;经典PCMCI和格兰杰领先紧邻的集群中,瓶颈落后;(iii)头条干预优势约为60%的样本量混杂因素,残差在标准$do(X=c)$干预下消失,仅在非标准随机强迫方案下存活;(iv)即使该残差再现,其效果在经典二元格兰杰中重现,效果更具普遍性。所剩的是狭窄的特征化结果;基准是持久的产物,上述每个阶段都是其对照组之一。
英文摘要:
A Mamba state-space model trained only for next-step prediction appears to recover Granger-causal structure through a simple readout $S = |W_{out} W_{in}|$, with early experiments suggesting the phenomenon generalized across architectures and benefited from interventional data at $p < 10^{-5}$. We package the protocol used to test that claim -- standardized synthetic generators (VAR/Lorenz/CauseMe-style), three intervention semantics ($do(X=c)$, soft-noise, random-forcing), edge-provenance cards on three real datasets, and size-matched control arms -- as a reusable falsification benchmark, and walk the claim through it in five stages. The method-level claim does not survive: (i) a plain linear bottleneck does as well or better; (ii) tuned Lasso beats the bottleneck on synthetic CauseMe-style benchmarks, and on Lorenz-96 (the only real benchmark with unambiguous ground truth) classical PCMCI and Granger lead a tight cluster in which the bottleneck trails; (iii) the headline intervention advantage is roughly 60% a sample-size confound, and the residual disappears under standard $do(X=c)$ interventions, surviving only under a non-standard random-forcing scheme; (iv) even that residual reproduces, with a larger effect, in classical bivariate Granger -- the effect is method-agnostic. What survives is a narrow characterization result; the benchmark is the lasting artifact, and each stage above is one of its control arms.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。