让模型自己判断何时该多算、停算或放弃，提升安全预测能力。
SCI: A Metacognitive Control for Signal Dynamics
- 在现有模型外加轻量控制层，用熵值动态调整推理步数。
- 错误分类样本推理次数最多达正确样本的4倍，异常检测准确率最高0.86。
- 适合医疗、工业等对安全敏感的场景，可作为故障预警信号。
现代深度学习系统通常以开环方式运行：单次输入即输出，不调节每个样本的计算量或解释力度。在高安全要求场景下，这种设计脆弱：简单和模糊输入获得相同处理，不确定性只能事后从原始概率中读取。本文提出外科认知解释器（SCI），一种轻量级闭环元认知控制层，可包裹已有随机模型，将预测转为迭代过程。SCI监控一个标量可解释状态SP(t)，此处以归一化熵为基础的置信度信号实现，并自适应决定停止、继续采样或放弃。目标并非直接提高准确率，而是调控解释误差ΔSP，暴露能追踪模型可能失效的安全信号。在三个领域验证：视觉（MNIST数字）、医学时间序列（MIT-BIH心律失常）和工业状态监测（滚动轴承）。所有情况下，误分类样本分配的推理步数均高于正确样本（MNIST与轴承达3-4倍，MIT-BIH达1.4倍）。由此产生的ΔSP可作为有效安全信号检测误分类（MNIST AUROC 0.63，MIT-BIH 0.70，轴承 0.86）。代码与可复现性：https://github.com/vishal-1344/sci
原文摘要 · Abstract (English)
Modern deep learning systems are typically deployed as open-loop function approximators: they map inputs to outputs in a single pass, without regulating how much computation or explanatory effort is spent on a given case. In safety-critical settings, this is brittle: easy and ambiguous inputs receive identical processing, and uncertainty is only read off retrospectively from raw probabilities. We introduce the Surgical Cognitive Interpreter (SCI), a lightweight closed-loop metacognitive control layer that wraps an existing stochastic model and turns prediction into an iterative process. SCI monitors a scalar interpretive state SP(t), here instantiated as a normalized entropy-based confidence signal, and adaptively decides whether to stop, continue sampling, or abstain. The goal is not to improve accuracy per se, but to regulate interpretive error ΔSP and expose a safety signal that tracks when the underlying model is likely to fail. We instantiate SCI around Monte Carlo dropout classifiers in three domains: vision (MNIST digits), medical time series (MIT-BIH arrhythmia), and industrial condition monitoring (rolling-element bearings). In all cases, the controller allocates more inference steps to misclassified inputs than to correct ones (up to about 3-4x on MNIST and bearings, and 1.4x on MIT-BIH). The resulting ΔSP acts as a usable safety signal for detecting misclassifications (AUROC 0.63 on MNIST, 0.70 on MIT-BIH, 0.86 on bearings). Code and reproducibility: https://github.com/vishal-1344/sci
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。