Photonic Matrix Computing With Modulated Photodetector Weights
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Neural network computations in electronic systems face bottlenecks due to heat generation, power consumption, and bandwidth limitations.
Innovation Solution
Implementing photonic computing methods that modulate optical signals with input and weight information to perform matrix operations, using photodetectors and control circuitry to generate electrical outputs representing computation results, including activation functions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If electronic systems use transistors and memory units to execute neural network operations, then computation can be performed, but heat generation and power consumption increase
Solution Approach 1:
The patent replaces electronic computation systems with photonic computation systems. Optical modulators modulate optical carriers with input and weight data, and photodetectors detect the optical signals to perform matrix multiplication operations. This substitution of electronic systems with photonic systems reduces heat generation and power consumption while maintaining computational functionality.
2Productivity
If electronic systems use transistors and memory units to execute neural network operations, then computation can be performed, but bandwidth limitation occurs
Solution Approach 1:
The patent substitutes electronic signal processing with optical signal processing. Optical carriers can be modulated and detected to perform computations, enabling higher bandwidth operations. The photonic system uses optical modulators to encode data onto optical carriers and photodetectors to detect the modulated signals, achieving higher computational throughput and bandwidth compared to electronic systems.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Reduces heat and power consumption while enhancing computational bandwidth by performing neural network operations optically, addressing the limitations of electronic systems.
Implementation Method 1
detecting each of the second modulated optical signals by a corresponding photodetector of a plurality of photodetectors
Data Source
AI summary
Methods, circuits, devices, systems and techniques for photonic computing are provided. In one aspect, a method includes: detecting at least one modulated optical signal using a plurality of photodetectors, the at least one modulated optical signal representing one or more input elements of an input matrix of a mathematical function, the plurality of photodetectors being modulated based on weight elements of a weight matrix of the mathematical function; and generating electrical outputs based on outputs of the plurality of photodetectors using a control circuitry, the electrical outputs representing a computation result of the mathematical function corresponding to a multiplication of the input matrix and the weight matrix.