Semiconductor equipment
The semiconductor device addresses power and heat issues in neural networks by using specialized circuit configurations with metal oxide transistors and capacitors, ensuring efficient and stable operations with reduced power consumption and heat.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- SEMICON ENERGY LAB CO LTD
- Filing Date
- 2026-01-26
- Publication Date
- 2026-05-26
AI Technical Summary
Artificial neural networks face challenges with increasing power consumption and heat generation due to complex calculations, especially in hierarchical networks, which affect transistor characteristics and lead to data degradation in multiplication cells.
A semiconductor device with specific circuit configurations, including transistors and capacitors, is designed to perform sum-of-products operations with reduced power consumption and minimal heat generation, utilizing metal oxide in the channel formation region and incorporating a current comparison circuit.
The semiconductor device achieves low power consumption, reduced heat generation, and minimizes the impact of environmental temperature and transistor variations, while maintaining accurate calculations by periodically rewriting weight coefficients.
Smart Images

Figure 2026086483000001_ABST