Associative Memory of Analog Networks without the Energy Concept Using Biased Patterns and Positive-Valued Transfer Functions

スポンサーリンク

概要

• 論文の詳細を見る

• Associative memory of analog netmral networks for which the concept of energy functions islost is sttmdied by means of the SCSNA (Self-Consistent Signal-to-Noise Analysis). Asymrnetric synaptic couplings with biased random patterns are assurned together with positive-valuedtransfer functions which allow nonrnonotonicity resulting from an appropriate cut off of outputactivity, Phase diagrams are given in terms of several pararneters of the networks, showing theoccurrence of enhancement of the storage capacity due to nonmonotonic transfer functions. Su-per retrieval states ensuring errorless memory retrieval under the loading of extensively znanypatterns are allowed to rernain in existence in the presence of asymmetric synaptic couplingswith biased patterns. A sample-dependent cornponent in the local fields of neurons, which arisesfrom the assumed asyrnrnetric couplings, is discussed and shown to become of no effect in thesuper retrieval phase.

• 社団法人日本物理学会の論文
• 1997-05-15

著者

• Yoshioka Masahiko

  Department of Anatomy, National Defense Medical College

• Shiino Masatoshi

  Department of Applied Physics

• Yoshioka Masahiko

  Department Of Anatomy National Defense Medical College

• Yoshioka Masahiko

  Department Of Applied Physics Tokyo Institute Of Technology

• SHINO Masatoshi

  Department of Applied Physics,Tokyo Institute of Technology

• Shino Masatoshi

  Department Of Applied Physics Tokyo Institute Of Technology

関連論文

• Transdermal Delivery of Photosensitizer by the Laser-Induced Stress Wave in Combination with Skin Heating : Cross-Disciplinary Areas
• Nonlinear Master Equation Approach to Asymmertrica Neural Networks of the Hopfield-Hemmen Type
• Associative Memory of Analog Networks without the Energy Concept Using Biased Patterns and Positive-Valued Transfer Functions
• Noise Spectrum of a Nonlinear Oscillator Driven by Periodic and Random Forces
• Chaotic Image Retrieval in Markovian Asymmetric Neural Networks with Sign-Constrained Synaptic Couplings
• Statistical mechanics of attractor neural networks and self-consistent signal-to-noise analysis : analog neural networks with nonmonotonic transfer functions and enhancement of the storage capacity(New Developments in Statistical Physics Similarities in D
• Critical Behaviour of Spin Relaxation in EuS
• H-Theorem with Generalized Relative Entropies and the Tsallis Statistics

スポンサーリンク