Files
myWiki/concepts/koopman-autoencoder.md

35 lines
1.0 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

---
title: "Koopman Autoencoder (KAE)"
created: 2026-05-11
updated: 2026-05-11
type: concept
tags: [deep-learning, autoencoder, dynamical-systems]
sources: [[liu-koopa-2023]]
---
# Koopman Autoencoder (KAE)
## 定义
Koopman 自编码器是将 [[koopman-theory|Koopman 理论]] 与深度学习自编码器结合的框架。用编码器学习测量函数 gKoopman 嵌入),线性层实现 Koopman 算子,解码器重建状态。
## 架构
```
x_t → [Encoder] → g(x_t) → [K·] → g(x_{t+1}) → [Decoder] → x_{t+1}
```
- **编码器** = 测量函数:将状态映射到测量空间
- **线性中间层** = Koopman 算子 K
- **解码器** = 逆测量函数
## Koopa 对 KAE 的改进
传统 KAE 需要**重构损失绑定**(编码器-解码器 + 前向预测双重目标Koopa 通过深度残差结构将其分解,实现端到端预测优化,消除了绑定的训练困难。
## 相关概念
- [[koopman-theory|Koopman 理论]]
- [[dynamic-mode-decomposition|DMD]]
- [[liu-koopa-2023|Koopa]]