18316474. DEVICES AND METHODS EMPLOYING OPTICAL-BASED MACHINE LEARNING USING DIFFRACTIVE DEEP NEURAL NETWORKS simplified abstract (THE REGENTS OF THE UNIVERSITY OF CALIFORNIA)
Contents
- 1 DEVICES AND METHODS EMPLOYING OPTICAL-BASED MACHINE LEARNING USING DIFFRACTIVE DEEP NEURAL NETWORKS
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 DEVICES AND METHODS EMPLOYING OPTICAL-BASED MACHINE LEARNING USING DIFFRACTIVE DEEP NEURAL NETWORKS - A simplified explanation of the abstract
- 1.4 Simplified Explanation
- 1.5 Potential Applications
- 1.6 Problems Solved
- 1.7 Benefits
- 1.8 Original Abstract Submitted
DEVICES AND METHODS EMPLOYING OPTICAL-BASED MACHINE LEARNING USING DIFFRACTIVE DEEP NEURAL NETWORKS
Organization Name
THE REGENTS OF THE UNIVERSITY OF CALIFORNIA
Inventor(s)
Aydogan Ozcan of Los Angeles CA (US)
Yair Rivenson of Los Angeles CA (US)
Xing Lin of Los Angeles CA (US)
Deniz Mengu of Los Angeles CA (US)
DEVICES AND METHODS EMPLOYING OPTICAL-BASED MACHINE LEARNING USING DIFFRACTIVE DEEP NEURAL NETWORKS - A simplified explanation of the abstract
This abstract first appeared for US patent application 18316474 titled 'DEVICES AND METHODS EMPLOYING OPTICAL-BASED MACHINE LEARNING USING DIFFRACTIVE DEEP NEURAL NETWORKS
Simplified Explanation
The abstract describes a patent application for an all-optical Diffractive Deep Neural Network (DNN) architecture. This architecture uses passive diffractive or reflective substrate layers to perform various functions or tasks. The application of this architecture was successfully demonstrated through the creation of 3D-printed DNNs that learned to implement handwritten classifications and lens function at the terahertz spectrum. The all-optical deep learning framework can perform complex functions and tasks at the speed of light, enabling applications in image analysis, feature detection, object classification, and new camera designs. It can also be used in conjunction with a digital neural network back-end.
- All-optical Diffractive Deep Neural Network (DNN) architecture
- Passive diffractive or reflective substrate layers are designed using deep learning
- 3D-printed DNNs successfully implement handwritten classifications and lens function
- Performs complex functions and tasks at the speed of light
- Applications in image analysis, feature detection, object classification, and camera designs
- Can be used with a digital neural network back-end
Potential Applications
- All-optical image analysis
- Feature detection
- Object classification
- New camera designs
- Optical components that can learn to perform unique tasks using DNNs
Problems Solved
- Enables all-optical implementation of complex functions and tasks
- Speeds up processing by performing tasks at the speed of light
- Provides a framework for designing optical components that can learn to perform specific tasks
Benefits
- Faster processing speed compared to computer-based neural networks
- All-optical implementation allows for real-time analysis and classification
- Enables the development of new camera designs and optical components
Original Abstract Submitted
An all-optical Diffractive Deep Neural Network (DNN) architecture learns to implement various functions or tasks after deep learning-based design of the passive diffractive or reflective substrate layers that work collectively to perform the desired function or task. This architecture was successfully confirmed experimentally by creating 3D-printed DNNs that learned to implement handwritten classifications and lens function at the terahertz spectrum. This all-optical deep learning framework can perform, at the speed of light, various complex functions and tasks that computer-based neural networks can implement, and will find applications in all-optical image analysis, feature detection and object classification, also enabling new camera designs and optical components that can learn to perform unique tasks using DNNs. In alternative embodiments, the all-optical DNN is used as a front-end in conjunction with a trained, digital neural network back-end.
