Analog Multiply-Accumulate Circuit Using Controlled Capacitor Charging

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing analog systems for multiply-accumulate operations in neural networks face challenges in achieving high accuracy and efficient power consumption while performing multiply-accumulate operations.

Innovation Solution

A multiply-accumulate operation device that includes input lines, multiplication units, an accumulation unit, a charging unit, and an output unit, where pulse signals with specific pulse widths are used to generate charges, accumulate them, and output a signal based on threshold determination, with adjustable charging speed and threshold values to improve resolution and accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Use of energy by moving object

If analog multiply-accumulate operation is performed using conventional circuits, then power consumption is reduced compared to digital systems, but operation accuracy is insufficient

Engineering Contradiction:
Improvepower consumptionVSAvoidoperation accuracy
Core Design Contradiction:
Use of energy by moving objectVSMeasurement precision

Solution Approach 1:

The patent changes the parameter of charge accumulation by introducing a controlled charging phase where a constant current source charges the capacitor at a predetermined speed after multiplication charges are accumulated. This parameter change in the charging process enables more precise voltage representation of the multiply-accumulate result, thereby improving operation accuracy while maintaining the low power consumption advantage of analog systems

Inventive Principle:
Principle #35Parameter changes

2Device complexity

If charge accumulation is performed without controlled charging, then circuit complexity is reduced, but resolution to express operation results is insufficient

Engineering Contradiction:
Improvecircuit complexityVSAvoidresolution to express results
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent applies preliminary action by performing a controlled charging operation at a predetermined speed before the threshold determination step. This preliminary charging action ensures that the capacitor voltage accurately reflects the accumulate state, providing sufficient resolution for the subsequent threshold comparison to determine the multiply-accumulate result, thereby improving result expression resolution without significantly increasing circuit complexity

Inventive Principle:
Principle #10Preliminary action

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

This approach enhances the accuracy of multiply-accumulate operations in analog systems by optimizing the charging speed and threshold values, allowing for improved expression and calculation of results.

Implementation Method 1

charges corresponding to the product of the corresponding electrical signal and the weight are output, and the output charges are appropriately accumulated in a capacitor

Methodology Applied
Scientific EffectCapacitance: Capacitance

Implementation Method 2

The charging unit charges, at a charging speed associated with an accumulation state of the accumulation unit, the accumulation unit in which the sum of the charges corresponding to the multiplication values have been accumulated

Methodology Applied
Scientific EffectElectrical charging: Conduction (electrical)

Implementation Method 3

The output unit outputs a multiply-accumulate signal representing a sum of the multiplication values by executing threshold determination using a threshold value associated with the accumulation state of the accumulation unit on a voltage held by the accumulation unit

Methodology Applied
Scientific EffectThreshold determination:

Data Source

PatentEP3822843B1Multiply-accumulate device, multiply-accumulate circuit, multiply-accumulate system, and multiply-accumulate method
Publication Date: 2024.10.16 SONY GROUP CORP
  • EP3822843B1 patent drawingFigure 1~2
  • EP3822843B1 patent drawingFigure 3~4
  • EP3822843B1 patent drawingFigure 5

AI summary

This multiply-accumulate operation device executes a multiply-accumulate operation with an analog circuit, and includes: a plurality of input lines; a plurality of multiplication units; an accumulation unit; a charging unit; and an output unit. Pulse signals having pulse widths corresponding to input values are input to the plurality of input lines. The plurality of multiplication units generates, on the basis of the pulse signals input to each of the plurality of input lines, charges corresponding to multiplication values obtained by multiplying the input values by weight values. The accumulation unit accumulates a sum of the charges corresponding to the multiplication values generated by each of the plurality of multiplication units. The charging unit charges, at a charging speed associated with an accumulation state of the accumulation unit, the accumulation unit in which the sum of the charges corresponding to the multiplication values have been accumulated. The output unit outputs a multiply-accumulate signal representing a sum of the multiplication values by executing threshold determination using a threshold value associated with the accumulation state of the accumulation unit on a voltage held by the accumulation unit after the charging by the charging unit is started.