Neural Network Cell Array Compensation for ADC Voltage Drop
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Artificial neural networks using analog-to-digital converters face significant voltage drops due to high output currents, leading to decreased computational accuracy.
Innovation Solution
A neural network device incorporating a digital-to-analog converter, a cell array with memory cells, and an analog-to-digital converter to manage input and output voltages, along with dummy conductances to compensate for voltage differences, ensuring accurate computation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If an analog-to-digital converter is used to detect current output through analog computation, then computational speed is improved, but voltage drop increases leading to decreased accuracy
Solution Approach 1:
A sense amplifier is introduced as an intermediary component between the cell array and the analog-to-digital converter. The sense amplifier detects and amplifies the output voltage from the cell array before it reaches the converter, compensating for voltage drops and ensuring accurate detection of computational results without sacrificing speed
Solution Approach 2:
The system implements a feedback mechanism where the sense amplifier continuously monitors the output voltage and adjusts its amplification to compensate for voltage drops caused by high output currents, maintaining computational accuracy while preserving fast analog computation
2Productivity
If high output current is used for analog computation, then computational efficiency is improved, but voltage drop increases causing accuracy degradation
Solution Approach 1:
The sense amplifier serves as a mediator that allows high output currents to be used for efficient computation while it detects and compensates for the resulting voltage drops, enabling both high productivity and maintained accuracy
3Use of energy by moving object
If on-chip memory is used for CIM-based computation, then power consumption is reduced, but voltage drop management becomes more critical
Solution Approach 1:
The sense amplifier acts as a local intermediary within the on-chip memory structure, detecting and compensating for voltage drops at their source, thereby maintaining voltage stability and computational reliability while preserving the low power consumption benefits of on-chip CIM computation
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
The solution effectively mitigates voltage drops, maintaining computational accuracy and efficiency by using on-chip memory for CIM-based computation, reducing power consumption and heat generation.
Implementation Method 1
a digital-to-analog converter configured to convert a digital signal into input voltages
Implementation Method 2
the cell array is configured to output, through the plurality of bit lines, output voltages obtained by performing computation on the input voltages
Implementation Method 3
an analog-to-digital converter configured to detect the output voltages and convert the output voltages into a digital signal
Data Source
AI summary
Provided is a neural network device including a digital-to-analog converter configured to convert a digital signal into input voltages, a cell array including a plurality of memory cells that are arranged in a plurality of bit lines and a plurality of word lines and has weights of a neural network transferred thereto, wherein the cell array is configured to output, through the plurality of bit lines, output voltages obtained by performing computation on the input voltages that are input through the plurality of word lines, and an analog-to-digital converter configured to detect the output voltages and convert the output voltages into a digital signal.