Neural Network Learning Rate Scaling for Quantization Noise

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing training algorithms for deep neural networks using lower-precision quantized formats introduce quantization noise that can lead to inaccurate gradient directions and reduced accuracy, particularly when quantization errors are not uniformly distributed across network elements.

Innovation Solution

A method is introduced to compensate for quantization noise by adjusting hyper-parameters, such as the learning rate, based on noise-to-signal metrics, which are calculated to mitigate the aggregation of quantization noise during back-propagation, allowing for efficient training with lower-precision formats while maintaining accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If lower-precision quantized formats are used for training deep neural networks, then computational efficiency and memory usage are improved, but quantization noise increases leading to inaccurate gradient directions and reduced training accuracy

Engineering Contradiction:
Improvecomputational efficiencyVSAvoidtraining accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent dynamically adjusts the learning rate based on the quantization bit-width and noise characteristics. By changing the learning rate parameter in response to quantization precision changes, the system compensates for increased noise in lower-precision formats, allowing efficient training with reduced precision while maintaining convergence and accuracy

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent implements a feedback mechanism where the quantization noise level is continuously monitored and used to adjust the learning rate. The system calculates noise metrics based on the quantization configuration and feeds this information back to modify training parameters, creating a closed-loop control that adapts to quantization effects in real-time

Inventive Principle:
Principle #23Feedback

2Ease of operation

If uniform quantization is applied to all network elements, then implementation simplicity is improved, but elements with different sensitivity requirements cannot be optimized individually leading to suboptimal accuracy

Engineering Contradiction:
Improveimplementation simplicityVSAvoidmodel accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent applies different learning rate scaling factors to different elements of the neural network based on their individual noise sensitivity and quantization characteristics. Instead of uniform treatment, each weight or parameter can receive customized learning rate adjustments, allowing elements with different sensitivity requirements to be optimized individually while maintaining manageable complexity through automated per-element analysis

Inventive Principle:
Principle #3Local quality

3Device complexity

If quantization noise is not compensated, then training process simplicity is improved, but gradient direction accuracy deteriorates leading to slower convergence and increased training epochs

Engineering Contradiction:
Improvetraining process complexityVSAvoidtraining time
Core Design Contradiction:
Device complexityVSLoss of time

Solution Approach 1:

The patent performs preliminary calculation of quantization noise metrics and determines appropriate learning rate scaling factors before the actual training process begins. By pre-computing the noise compensation parameters based on the quantization configuration, the system avoids adding complexity during training while ensuring gradient direction accuracy from the start, leading to faster convergence without increasing training process complexity

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3899801B1Scaled learning for training deep neural networks
Publication Date: 2025.07.09 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP3899801B1 patent drawingFigure 1
  • EP3899801B1 patent drawingFigure 2
  • EP3899801B1 patent drawingFigure 3

AI summary

Methods and apparatus are disclosed for adjusting hyper-parameters of a neural network to compensate for noise, such as noise introduced via quantization of one or more parameters of the neural network. In some examples, the adjustment can include scaling the hyper-parameter based on at least one metric representing noise present in the neural network. The at least one metric can include a noise-to-signal ratio for weights of the neural network, such as edge weights and activation weights. In a quantized neural network, a learning rate hyper-parameter used to compute a gradient update for a layer during back propagation can be scaled based on the at least one metric. In some examples, the same scaled learning rate can be used when computing gradient updates for other layers.