神经科学与心理学2024nature.com
在可塑性之前推断神经活动：超越反向传播的学习
Inferring neural activity before plasticity for learning beyond backpropagation
专业 · 约 90 分钟读完前置知识：神经网络与反向传播、信用分配问题、计算神经科学基础
导读
这篇论文提出一种叫“前瞻配置”（prospective configuration）的学习原理，试图解释大脑如何解决信用分配问题，作为反向传播（backpropagation）的替代方案。其核心做法是先推断出能解释误差的神经活动（激活值），再据此更新权重，从而避免反向传播中常见的“灾难性干扰”（新学习破坏旧关联）。作者展示该方法学习效率更高、只需局部计算，并能解释多种学习任务中的实验数据。
HN 关注点：HN 读者长期关注“神经网络是否像大脑”这一话题，这篇 Nature 论文给出了大脑学习机制不同于反向传播的具体理论，引发对 AI 与生物大脑类比的热议。
HN 评论精选
这几条评论里：支持 1 质疑 1 补充 2
补充概述论文核心机制：先推断激活再更新权重，避免灾难性干扰
tl;dr：论文提出“前瞻配置”原理来解释大脑如何做信用分配和学习，区别于反向传播。反向传播会导致“灾难性干扰”——学新东西会破坏旧关联，这与观察到的生物过程不符。据我理解，前瞻配置通过求解“激活值应该是什么才能解释误差”来学习，再据此更新权重，从而避免破坏旧关联。他们还展示了该原理如何解释观察到的生物过程。很酷，可惜找不到代码。
isaacimagine查看英文原文
Wait, my brain doesn't do backprop over a pile of linear algebra after having the internet rammed through it? No way that's crazy /s tl;dr: paper proposes a principle called 'prospective configuration' to explain how the brain does credit assignment and learns, as opposed to backprop. Backprop can lead to 'catastrophic interference' where learning new things abalates old associations, which doesn't match observed biological processes. From what I can tell, prosp. config learns by solving what the activations should have been to explain the error, and then updates the weights in accordance, which apparently somehow avoids abalating old associations. They then show how prosp. config explains observed biological processes. Cool stuff, wish I could find the code. There's some supplemental notes: https://static-content.springer.com/esm/art%3A10.1038%2Fs41593-023-01514-1/MediaObjects/41593_2023_1514_MOESM1_ESM.pdf
质疑指出大脑与 ANN 的差异早有共识，标题有误导性
很久以来就很清楚（例如 Marvin Minsky 的早期研究）：1. ANN 和大脑都需要解决信用分配问题；2. 反向传播对 ANN 有效，但大脑很可能不是这样解决的。这篇论文很有意思，但它更多是关于大脑如何解决信用分配问题的新理论。HN 标题让人以为大脑与 ANN 的差异是首次被发现，有误导性。
lukeinator42查看英文原文
It has been clear for a long time (e.g. Marvin Minsky's early research) that: 1. both ANNs and the brain need to solve the credit assignment problem 2. backprop works well for ANNs but probably isn't how the problem is solved in the brain This paper is really interesting, but is more a novel theory about how the brain solves the credit assignment problem. The HN title makes it sound like differences between the brain and ANNs were previously unknown and is misleading IMO.
补充建议用“模型训练”等术语区分 AI 与大脑，避免误导性类比
有人对标题或研究本身感到意外，这可能是在回应大量声称“神经网络像大脑”的文章——甚至使用神经元、突触这类术语。这些说法广泛传播后，人们开始在其上构建让 AI 更像人类的理论。我们中许多人对这种说法感到厌倦，于是用不同的术语来区分：我称 AI 的东西为模型，说“模型训练”而不是“学习”，说“在数据中发现并利用模式”。连外行也能理解这些术语，不会误以为它们和大脑一样。
nickpsecurity查看英文原文
Some are surprised that anyone would make this point, either the title or the research. It might be a response to the many, many claims in articles that neural networks work like the brain. Even using terms like neurons and synapses. With those claims getting widespread, people also start building theories on top of them that make AI’s more like humans. Then, we won’t need humans or they’ll be extinct or something. Many of us whom are tired of that are both countering it and just using different terms for each where possible. So, I’m calling the AI’s models, saying model training instead of learning, and finding and acting on patterns in data. Even laypeople seem to understand these terms with less confusion about them being just like brains.
支持引用论文原文，强调其解决可塑性-稳定性困境并优于反向传播
好论文。换句话说：我们的大脑不只是线性代数运算的集合……论文证明这种神经活动的重新配置是实现优于反向传播的学习性能的关键，并能解释多种学习任务的实验数据。前瞻配置还提供了一系列与反向传播不同的实验预测。它通过减少干扰实现比反向传播更高效的学习，只需局部计算和可塑性，并解决了长期存在的问题：大脑如何解决“可塑性-稳定性”困境——比如尽管初级视觉皮层在学习中不断调整表征，我们仍能理解一生中学到的视觉刺激的含义。
belter查看英文原文
Great paper. In other words: Our brain is more than just a set of linear algebra operations... "...We demonstrate that this reconfiguration of neural activity is the key to achieving learning performance superior to that of backpropagation and to explaining experimental data from diverse learning tasks. Prospective configuration further offers a range of experimental predictions distinct from those of backpropagation (Supplementary Figs. 11 and 12). Together, we have demonstrated that prospective configuration enables more efficient learning than backpropagation by reducing interference, demonstrates superior performance in situations faced by biological organisms, requires only local computation and plasticity and matches experimental data across a wide range of tasks... ...Our theory addresses a long-standing question of how the brain solves the plasticity-stability dilemma, for example, how it is possible that, despite adjustment of representation in the primary visual cortex during learning, we can still understand the meaning of visual stimuli we learned over our lifetime..."
相关资源
从这篇论文所有 HN 讨论帖的全部评论里整理出来的链接。
历次讨论
- 2024Hacker News 讨论帖17 分 · 1 评
- 2024Hacker News 讨论帖145 分 · 48 评
同领域还值得读
神经科学家能理解微处理器吗？（2017）
Could a Neuroscientist Understand a Microprocessor? (2017)
提出一个思想实验：把神经科学常用的分析方法（损伤实验、电生理、功能定位等）应用到一个我们完全了解其原理的工程系统——6502 微处理器（运行游戏）上，检验这些方法能否真正“理解”…
神经科学与心理学2017 · 411 分 · 135 评进阶 · 约 60 分钟编辑精选大脑衰老呈现非线性转变，提示中年“关键窗口”
Brain aging shows nonlinear transitions, suggesting a midlife "critical window"
这篇 PNAS 论文研究大脑衰老的非线性变化，发现大脑老化并非匀速进行，而是在中年（约 40-60 岁）存在一个关键的"临界窗口期"，此时干预可能有效。
神经科学与心理学2025 · 295 分 · 159 评进阶 · 约 75 分钟编辑精选社交网络中大规模情绪传染的证据（2014）
Evidence of massive-scale emotional contagion through social networks (2014)
这篇 2014 年发表于 PNAS 的论文研究了 Facebook 上约 70 万用户的信息流中，减少正面或负面内容是否会传染性地影响用户自己发布内容的情绪。
神经科学与心理学2020 · 285 分 · 72 评入门 · 约 45 分钟编辑精选裸盖菇素使人类大脑去同步化
Psilocybin desynchronizes the human brain
这篇 Nature 研究用个体化的 fMRI 精密成像追踪受试者服用裸盖菇素（psilocybin）前后的脑功能连接变化，发现药物会造成大范围的脑网络「去同步化」，尤其是默认模式…
神经科学与心理学2024 · 273 分 · 259 评进阶 · 约 60 分钟编辑精选对智力的赞扬会削弱儿童的动机与表现（1998）
Praise for Intelligence Can Undermine Children's Motivation (1998)
这篇 1998 年的经典论文（Dweck 等人）通过实验研究表扬方式对儿童动机的影响：夸孩子“聪明”（表扬智力）会让他们回避挑战、害怕失败，而表扬努力则更能培养坚持性。
神经科学与心理学2016 · 253 分 · 155 评入门 · 约 45 分钟编辑精选