5D Library › Books › 5DL-B-015
沙盘、宿命与主动推理
马尔可夫毯、变分自由能最小化、探索-利用权衡与东方因果立命观的计算神经科学
The Sandbox, Determinism and Active Inference: Markov Blankets, Variational Free-Energy Minimisation, Exploration-Exploitation Trade-offs and East Asian Causality in Computational Neuroscience
DOI10.5281/zenodo.23738794摘要
有些人觉得自己的人生像照着剧本在演：换了环境，同样的冲突与失望一再重演。本书用计算神经科学的预测加工与主动推理理论考察这一现象。大脑用先验信念预测感官输入，并设法减小预测误差；减小误差有两条路：修改信念，或者采取行动让世界符合预测。本书依次讨论：精度加权的贝叶斯更新与行为确认（默顿、斯奈德等人的研究，以及教师期望效应较小且有争议的证据）；变分自由能与期望自由能，以及一个简化的主动推理智能体如何因悲观的结果先验而反复选择同一个陷阱、又如何通过低风险条件下累积的相反经验脱离陷阱；马尔可夫毯的统计含义、它在自由能原理中的用法所受的批评，以及输入来源的改变如何使内部模型与世界脱钩并通过行动推开环境；多臂老虎机中的贪婪、ε-贪婪、UCB1 与汤普森采样，模拟退火与信息增益；压力与过度利用、心理安全和严肃游戏的证据，以及一个温度参数模型；洛伦茨系统的李雅普诺夫指数与可预测期限，以及相容论。
八章与袁黄《了凡四训·立命之学》、《成唯识论》的种子与熏习、《楞严经》「若能转物」并读，并指出并读的限度；最后给出一份四步贝叶斯审计表，以及一项比较结构化沙盘与牌卡推演工作坊和时间匹配的阅读讲授课程的预注册随机对照试验方案（主要结局为视野任务中的定向探索参数，评估者与分析者盲，按小班聚类，每组 122 人完成，判定规则写成代码）。书中以标注框区分已发表的证据、可运行的 Python 计算（15 段程序，全部核对）、原典与作者的假说。本书不是心理治疗或诊断；关于「旧先验驱动重复挫折」的内容是假说与跨学科类比，绝不意味着受害者要为自己的遭遇负责。
关键词：主动推理；自由能原理；预测加工；自我实现的预言；行为确认；马尔可夫毯；探索与利用；多臂老虎机；汤普森采样；上置信界；模拟退火；定向探索；洛伦茨吸引子；李雅普诺夫指数；相容论；了凡四训；唯识
Abstract
Some people feel their life follows a script: the surroundings change, and the same conflict and disappointment play out again. This book examines the experience with predictive processing and active inference from computational neuroscience. The brain predicts its sensory input from prior beliefs and works to reduce prediction error, which it can do in two ways: by revising the belief, or by acting so that the world fits the prediction. The book covers precision-weighted Bayesian updating and behavioural confirmation (Merton, Snyder and colleagues, and the smaller, disputed evidence on teacher-expectation effects); variational and expected free energy, and how a simplified active-inference agent with a pessimistic outcome prior keeps choosing the same trap, and escapes it through contrary experience accumulated under low stakes; the statistical meaning of Markov blankets, the criticism of their use in the free-energy principle, and how changing the source of input decouples the internal model from the world and, through action, pushes the environment away; greedy, epsilon-greedy, UCB1 and Thompson sampling in multi-armed bandits, simulated annealing and information gain; evidence on stress and over-exploitation, psychological safety and serious games, with a temperature-parameter model; and the Lyapunov exponent and predictability horizon of the Lorenz system, with compatibilism.
The eight chapters are read beside Yuan Huang's Four Lessons of Liaofan, the seeds and perfuming of the Cheng Weishi Lun and the Shurangama Sutra's "if one can turn things", noting the limits of such a reading, and end with a four-step Bayesian audit and a pre-registered randomised controlled trial comparing a structured sandbox and card-simulation workshop with a time-matched reading-and-lecture course (primary outcome the directed-exploration parameter of the Horizon task; assessors and analysts blind; analysis clustered by group; 122 completers per arm; decision rule written as code). Marked boxes separate published evidence, runnable Python calculations (15 programs, all checked), classical sources and the author's hypotheses. This book is not psychotherapy or diagnosis; its account of old priors driving repeated setbacks is a hypothesis and a cross-disciplinary analogy and never implies that victims are responsible for what happened to them.
Keywords: active inference; free energy principle; predictive processing; self-fulfilling prophecy; behavioural confirmation; Markov blanket; exploration-exploitation; multi-armed bandit; Thompson sampling; UCB1; simulated annealing; directed exploration; Lorenz attractor; Lyapunov exponent; compatibilism; Four Lessons of Liaofan; Yogacara
目录 · Contents
- 序：同一个陷阱，为什么跳进去第二次English 中文Preface: Why Fall into the Same Trap Twice
- 第一章 贝叶斯脑与自我实现的预言English 中文Chapter 1 The Bayesian Brain and the Self-Fulfilling Prophecy
- 第二章 两条路径：改变认知，还是改变世界English 中文Chapter 2 Two Routes: Change the Mind, or Change the World
- 第三章 马尔可夫毯与个体的边界English 中文Chapter 3 Markov Blankets and the Boundary of the Individual
- 第四章 跳出局部最优：探索与利用English 中文Chapter 4 Escaping the Local Optimum: Exploration and Exploitation
- 第五章 结构化沙盘与先验的重塑English 中文Chapter 5 Structured Sandboxes and the Reshaping of Priors
- 第六章 决定论、混沌与能动性English 中文Chapter 6 Determinism, Chaos and Agency
- 第七章 原典：命由我作English 中文Chapter 7 Source Texts: "Fate Is Made by Oneself"
- 第八章 实践与预注册试验English 中文Chapter 8 Practice and a Pre-Registered Trial
How to cite · 引用
Fan, Y. (2026). The Sandbox, Determinism and Active Inference: Markov Blankets, Variational Free-Energy Minimisation, Exploration-Exploitation Trade-offs and East Asian Causality in Computational Neuroscience [沙盘、宿命与主动推理] (Version 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.23738794
BibTeX
@book{fan2026active,
title = {{The Sandbox, Determinism and Active Inference}: Markov Blankets, Variational Free-Energy Minimisation, Exploration-Exploitation Trade-offs and East Asian Causality in Computational Neuroscience},
author = {Fan, Yafu},
year = {2026},
publisher = {5D Library},
edition = {Version 1.0.0},
doi = {10.5281/zenodo.23738794},
url = {https://doi.org/10.5281/zenodo.23738794},
note = {Chinese title: 沙盘、宿命与主动推理}
}
| DOI | https://doi.org/10.5281/zenodo.23738794 |
|---|---|
| DOI（全部版本 all versions） | https://doi.org/10.5281/zenodo.23738793 |
| ORCID | https://orcid.org/0009-0004-3519-1203 |
| 5D Library ID | 5DL-B-015 |
| 版本 Version | 1.0.0 · 中英双语版 Bilingual edition |
| 出版日期 Date | 2026-10-28 |
| 网址 Online | https://5dlibrary.com/books/active-inference-sandbox/ |
| SHA-256 · active-inference-sandbox_bilingual_v1.0.0.pdf | 674d899d3ba6493878aa2075c52fe5f7fcbc7c49fa3bcb8ebcaae9270d4a1d61 |
| SHA-256 · active-inference-sandbox_en_v1.0.0.pdf | dc0210ab316f1aa10a019852d1390f4a608cc98363e4c21c1e53752806fd3147 |
| SHA-256 · active-inference-sandbox_zh_v1.0.0.pdf | de8bab0dcb19be0840d679ba9f0cae66aa32d0490e9fb44cb65c0fea84230437 |
| Licence · 许可 | Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0). https://creativecommons.org/licenses/by-nc-nd/4.0/ |
Disclosure. The author uses this book as the methodological basis of his community cognition courses, 5D Ascension Cards (Dimensional Shift Card Decks) and the Yafu AI knowledge base, and so has a direct commercial interest in its subject and in the result of the Chapter 8 trial. The book promotes no course, card deck or product and makes no health or therapeutic claim; the trial is designed to be run and analysed by researchers independent of the author's business; all the code and the study protocol are published openly for anyone to check.