“Over my head, but I'd like to get Don Hoffman and Chetan Prakash to review this paper. Link discussing their work if you are not familiar, which no doubt I found here on HN. https://www.frontiersin.org/articles/10.3389/fpsyg.2014.00577/full”
The world as a neural network
We discuss a possibility that the entire universe on its most fundamental level is a neural network. We identify two different types of dynamical degrees of freedom: "trainable" variables (e.g. bias vector or weight matrix) and "hidden" variables (e.g. state vector of neurons). We first consider stochastic evolution of the trainable variables to argue that near equilibrium their dynamics is well approximated by Madelung equations (with free energy representing the phase) and further away from the equilibrium by Hamilton-Jacobi equations (with free energy representing the Hamilton's principal function). This shows that the trainable variables can indeed exhibit classical and quantum behaviors with the state vector of neurons representing the hidden variables. We then study stochastic evolution of the hidden variables by considering $D$ non-interacting subsystems with average state vectors, $\bar{\bf x}^{1}$, ..., $\bar{\bf x}^{D}$ and an overall average state vector $\bar{\bf x}^{0}$. In the limit when the weight matrix is a permutation matrix, the dynamics of $\bar{\bf x}^μ$ can be described in terms of relativistic strings in an emergent $D+1$ dimensional Minkowski space-time. If the subsystems are minimally interacting, with interactions described by a metric tensor, then the emergent space-time becomes curved. We argue that the entropy production in such a system is a local function of the metric tensor which should be determined by the symmetries of the Onsager tensor. It turns out that a very simple and highly symmetric Onsager tensor leads to the entropy production described by the Einstein-Hilbert term. This shows that the learning dynamics of a neural network can indeed exhibit approximate behaviors described by both quantum mechanics and general relativity. We also discuss a possibility that the two descriptions are holographic duals of each other.
候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。
HN 书评
没有评论直接链到它;下面是它被提交的讨论串里回复最多的评论(不计入统计)“First a watch, now a neural network. Nothing new under the sun. > All that is needed is to find a physical phenomenon which cannot be described by neural networks. But neural networks are universal function approximations... Does this offer any falsifiable predictions? I freely admit that my understanding of physics at this level is superficial - can a…”
“A model is only useful based on its predictive power versus its complexity. The math of physics is about providing the simplest model possible to explain what we observe in reality. A neural network or any universal approximating function should be able to describe the entirety of reality with enough parameters. Lots and lots of different models can descr…”
- The World as a Neural Network ↗kipply · 2020-08-05 · 3 分
- The world as a neural network ↗hhs · 2020-08-06 · 1 分
- The World as a Neural Network ↗hhs · 2020-08-08 · 3 分
- The World as a Neural Network ↗godelmachine · 2020-08-08 · 3 分
- The World as a Neural Network ↗MindGods · 2020-09-11 · 24 分