New research paper details fixed points in hysteresis neural networks

A research paper published on arXiv explores the concept of multiple fixed points within discrete-time hysteresis neural networks. The study details how the network's binary hysteresis neurons, influenced by a threshold parameter, can lead to various stable fixed points. Researchers introduced entropy as a metric to evaluate the distribution of basin of attraction sizes, which represent the initial states that converge to a specific fixed point. The paper also demonstrates how this parameter can control entropy, potentially maximizing it to achieve a more uniform distribution, using a binary data classification problem as a concrete example. AI

IMPACT This research contributes to the theoretical understanding of neural network dynamics, potentially informing future model architectures.

RANK_REASON The cluster contains a single academic paper published on arXiv detailing theoretical research in neural networks. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.NE (Neural & Evolutionary) →

• alphaXiv
• arXiv
• arXivLabs
• Bank of America
• Basins of Attraction to Multiple Fixed Points in Discrete-time Hysteresis Neural Networks
• binary hysteresis neurons
• CatalyzeX Code Finder for Papers
• Connected Papers
• CORE Recommender
• DagsHub
• Discrete-time hysteresis neural network
• Gotit.pub
• Hugging Face
• Influence Flower
• Litmaps
• ScienceCast
• scite Smart Citations

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →