Machine learningRecurrent / reservoir
回声状态网络
回声状态网络（Echo State Network, ESN）是Herbert Jaeger和Harald Haas于2004年提出的一种循环神经网络。它利用一个大型、随机连接、固定的循环层（即“储备池”）将输入信号投射到高维非线性空间中。只有线性输出权重需要训练，通常通过岭回归进行，这使得ESN在计算上成本低廉，但在时间序列和混沌时间序列建模任务中具有很强的表达能力。
阅读完整方法
仅限会员
登录使用免费账户登录即可阅读本节。
Method map
The neighbourhood of related methods — select a node to explore.
来源
- Jaeger, H., & Haas, H. (2004). Harnessing nonlinearity: Predicting chaotic systems and saving energy in wireless communication. Science, 304(5667), 78–80. DOI: 10.1126/science.1091277 ↗
如何引用本页
ScholarGate. (2026, June 2). Echo State Network (Reservoir Computing). ScholarGate. https://scholargate.app/zh/deep-learning/echo-state-network
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
Compare side by side →