Dynamics of Reservoir Computing at the Edge of Stability

详细信息    查看全文

• 关键词：Reservoir computing ; Echo state network ; Bifurcation phenomena ; Center manifold theory ; Physical reservoir
• 刊名：Lecture Notes in Computer Science
• 出版年：2016
• 出版时间：2016
• 年：2016
• 卷：9947
• 期：1
• 页码：205-212
• 全文大小：819 KB
• 参考文献：1.Caluwaerts, K., et al.: The spectral radius remains a valid indicator of the echo state property for large reservoirs. In: IJCNN pp. 1–6 (2013)
  2.Guckenheimer, J., Holmes, P.: Nonlinear Oscillations, Dynamical Systems, and Bifurcations of Vector Fields. Springer, Heidelberg (1983)CrossRef MATH
  3.Lukoševičius, M., Jaeger, H.: Reservoir computing approaches to recurrent neural network training. Comput. Sci. Rev. 3(3), 127–149 (2009)CrossRef MATH
  4.Mori, H., Kuramoto, Y.: Dissipative Structures and Chaos. Springer, Heidelberg (2011)MATH
  5.Pieroux, D., Mandel, P.: Bifurcation diagram of a complex delay-differential equation with cubic nonlinearity. Phys. Rev. E 67, 056213 (2003)MathSciNet CrossRef
  6.Vandoorne, K., et al.: Toward optical signal processing using photonics reservoir computing. Opt. Express 16(15), 11182–11192 (2008)CrossRef
  7.Yamane, T., Katayama, Y., Nakane, R., Tanaka, G., Nakano, D.: Wave-based reservoir computing by synchronization of coupled oscillators. In: Arik, S., Huang, T., Lai, W.K., Liu, Q. (eds.) ICONIP 2015. LNCS, vol. 9491, pp. 198–205. Springer, Heidelberg (2015). doi:10.​1007/​978-3-319-26555-1_​23 CrossRef
  
• 作者单位：Toshiyuki Yamane (19)
  Seiji Takeda (19)
  Daiju Nakano (19)
  Gouhei Tanaka (20)
  Ryosho Nakane (20)
  Shigeru Nakagawa (19)
  Akira Hirose (20)
  
  19. IBM Research - Tokyo, Kawasaki, Kanagawa, 212-0032, Japan
  20. Graduate School of Engineering, The University of Tokyo, Tokyo, 113-8656, Japan
  
• 丛书名：Neural Information Processing
• ISBN：978-3-319-46687-3
• 刊物类别：Computer Science
• 刊物主题：Artificial Intelligence and Robotics
  Computer Communication Networks
  Software Engineering
  Data Encryption
  Database Management
  Computation by Abstract Devices
  Algorithm Analysis and Problem Complexity
  
• 出版者：Springer Berlin / Heidelberg
• ISSN：1611-3349
• 卷排序：9947

文摘

We investigate reservoir computing systems whose dynamics are at critical bifurcation points based on center manifold theorem. We take echo state networks as an example and show that the center manifold defines mapping of the input dynamics to higher dimensional space. We also show that the mapping by center manifolds can contribute to recognition of attractors of input dynamics. The implications for realization of reservoir computing as real physical systems are also discussed.