New Deep Neural Networks Formulated Using Non-Archimedean Analysis

A new class of deep neural networks (DNNs) has been developed, utilizing multilayered tree-like architectures. These architectures are encoded using numbers from the ring of integers of non-Archimedean local fields, which possess an inherent hierarchical structure. The proposed DNNs are demonstrated to be robust universal approximators for real-valued functions defined on these rings, as well as for square-integrable functions on the unit interval. AI

RANK_REASON Academic paper published on arXiv detailing a new formulation for deep neural networks. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• Deep Neural Networks
• W. A. Zúñiga-Galindo

• paper
• other

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