17991373. SCALABLE NEURAL NETWORK PROCESSING ENGINE simplified abstract (Apple Inc.)
Contents
SCALABLE NEURAL NETWORK PROCESSING ENGINE
Organization Name
Inventor(s)
Erik Norden of San Jose CA (US)
Sung Hee Park of Cupertino CA (US)
Jaewon Shin of Los Altos CA (US)
Christopher L. Mills of Saratoga CA (US)
Seungjin Lee of Los Gatos CA (US)
Fernando A. Mujica of Los Altos CA (US)
SCALABLE NEURAL NETWORK PROCESSING ENGINE - A simplified explanation of the abstract
This abstract first appeared for US patent application 17991373 titled 'SCALABLE NEURAL NETWORK PROCESSING ENGINE
Simplified Explanation
The abstract describes a neural processor circuit that can instantiate one or more neural networks. It includes a data buffer, a memory external to the circuit, and multiple neural engine circuits. These neural engine circuits generate output data using input data and kernel coefficients to execute tasks that instantiate the neural networks. The circuit can selectively activate or deactivate the neural engine circuits based on configuration data of the tasks. Multiple neural processor circuits can be included in an electronic device and selectively activated or deactivated to execute the tasks.
- Neural processor circuit with scalable architecture for instantiating neural networks
- Includes data buffer and external memory
- Multiple neural engine circuits generate output data using input data and kernel coefficients
- Neural engine circuits can be selectively activated or deactivated based on task configuration data
- Multiple neural processor circuits can be included in an electronic device and selectively activated or deactivated
Potential Applications
- Artificial intelligence and machine learning applications
- Image and speech recognition systems
- Natural language processing
- Autonomous vehicles and robotics
- Medical diagnostics and analysis
Problems Solved
- Scalability and flexibility in instantiating neural networks
- Efficient utilization of resources in neural processing
- Adaptability to different tasks and configurations
Benefits
- Faster and more efficient execution of neural networks
- Improved performance and accuracy in AI applications
- Cost-effective implementation of neural processing
- Enhanced capabilities in various industries and domains
Original Abstract Submitted
Embodiments relate to a neural processor circuit with scalable architecture for instantiating one or more neural networks. The neural processor circuit includes a data buffer coupled to a memory external to the neural processor circuit, and a plurality of neural engine circuits. To execute tasks that instantiate the neural networks, each neural engine circuit generates output data using input data and kernel coefficients. A neural processor circuit may include multiple neural engine circuits that are selectively activated or deactivated according to configuration data of the tasks. Furthermore, an electronic device may include multiple neural processor circuits that are selectively activated or deactivated to execute the tasks.