18094351. ARTIFICIAL NEURAL NETWORK CIRCUIT simplified abstract (SAMSUNG ELECTRONICS CO., LTD.)
Contents
ARTIFICIAL NEURAL NETWORK CIRCUIT
Organization Name
Inventor(s)
Jae-Joon Kim of Pohang-si (KR)
Hyungjun Kim of Pohang-si (KR)
ARTIFICIAL NEURAL NETWORK CIRCUIT - A simplified explanation of the abstract
This abstract first appeared for US patent application 18094351 titled 'ARTIFICIAL NEURAL NETWORK CIRCUIT
Simplified Explanation
The abstract describes an artificial neural network circuit that includes memory cells for storing weight data and thresholds, as well as pass transistors for transferring the stored values. The circuit also includes a column for connecting the memory cells and a sense amplifier for processing the output signal.
- The circuit includes unit weight memory cells and unit threshold memory cells.
- The weight memory devices store weight data, while the threshold memory device stores a threshold value.
- The weight-threshold column connects the unit weight memory cells and unit threshold memory cells.
- The sense amplifier receives the output signal of the weight-threshold column and a reference voltage.
Potential applications of this technology:
- Artificial intelligence systems
- Machine learning algorithms
- Pattern recognition systems
- Data analysis and processing
Problems solved by this technology:
- Efficient storage and retrieval of weight data and thresholds in neural networks
- Accurate processing of output signals in neural network circuits
Benefits of this technology:
- Improved performance and accuracy in artificial neural networks
- Enhanced efficiency in memory utilization
- Faster processing and response times in neural network circuits
Original Abstract Submitted
Provided is an artificial neural network circuit including unit weight memory cells including weight memory devices configured to store weight data and weight pass transistors, unit threshold memory cells including a threshold memory device programmed to store a threshold and a threshold pass transistor, a weight-threshold column in which the plurality of unit weight memory cells and the plurality of unit threshold memory cells are connected, and a sense amplifier configured to receive an output signal of the weight-threshold column as an input and receive a reference voltage as another input.
