RRAM-based RBF Neuron Hardware Achieves 89.1% MNIST Accuracy

Researchers have developed a novel hardware implementation for a Radial Basis Function (RBF) neuron using Metal-Oxide Resistive RAM (RRAM) technology. This design, based on a custom Template piXeL (TXL) cell, acts as an efficient substrate for metric-based classification and online adaptation on resource-constrained edge devices. Simulations show the RRAM-based RBF classifier achieving 89.1% accuracy on the MNIST dataset with low energy consumption. AI

IMPACT This RRAM-based hardware could enable more efficient and lower-power AI inference on edge devices.

RANK_REASON The cluster contains an academic paper detailing a novel hardware implementation for machine learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Analogue Content Addressable Memory
• arXiv
• Georgios Papandroulidakis
• Metal-Oxide Resistive RAM
• MNIST database
• radial basis function
• resistive random-access memory
• Template piXeL

• paper
• infra

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