Data Processing Model Adaptation via Concept Drift Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Data processing models face performance degradation due to concept drift in dynamic environments like finance and transportation, where data distribution changes over time, affecting their accuracy.

Innovation Solution

A method and device that detect concept drift by reading a test data stream, acquiring outlier detection parameters, and retraining the data processing model when drift is detected, using updated parameters to continuously monitor and adapt to changes in the data stream.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a data processing model is trained on known data and then processes new unknown data, then the model can perform data classification and prediction, but the model accuracy deteriorates when concept drift occurs due to changing data distribution over time

Engineering Contradiction:
Improvemodel adaptability to changing data distributionVSAvoidmodel accuracy
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent implements a feedback mechanism by continuously monitoring data streams and detecting concept drift, then triggering model retraining based on detected drift. The system uses a detection module to monitor data distribution changes and provides feedback to the training module, which retrains the model when drift is detected, creating a closed-loop system that adapts to changing conditions while maintaining accuracy.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent transforms the static trained model into a dynamic system that can adapt to changing conditions. The model transitions from a fixed state after initial training to a dynamic state where it can be continuously retrained based on detected concept drift. This dynamic approach allows the model to maintain reliability despite changing data distributions over time.

Inventive Principle:
Principle #15Dynamics

2Reliability

If the data processing model is continuously retrained to adapt to concept drift, then model accuracy is maintained, but system complexity increases due to additional detection and retraining mechanisms

Engineering Contradiction:
Improvemodel accuracyVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent extracts the concept drift detection functionality into a separate, independent detection module that operates independently from the main data processing model. This modular approach allows the detection mechanism to be added without fundamentally redesigning the core model architecture. The detection module can be selectively activated based on drift detection, reducing the need for continuous complex operations.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent performs preliminary action by pre-training the model on known data before deployment. This initial training phase prepares the model for future operations, and the system only activates additional detection and retraining mechanisms when concept drift is detected. This approach avoids the complexity of continuous monitoring and retraining from the outset, adding complexity only when necessary.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4170561A1Method and device for improving performance of data processing model, storage medium and electronic device
Publication Date: 2023.04.26 JINGDONG CITY BEIJING DIGITS TECH CO LTD
  • EP4170561A1 patent drawingFigure 1
  • EP4170561A1 patent drawingFigure 2~3
  • EP4170561A1 patent drawingFigure 4

AI summary

A method and a device for improving performance of a data processing model, a storage medium and an electronic device are provided. A piece of data in a determined test data read currently is determined as target data. Outlier detection parameters in a detection module are acquired. Detection of concept drift is performed on the data processing model based on the target data and the outlier detection parameters. A detection module is triggered to update each of the outlier detection parameters and the data processing model is retrained when concept drift is successfully detected. After the data processing model is already trained, a piece of data to be read next is determined as the target data, the updated outlier detection parameters in the detection module are acquired, and the detection is resumed until all the pieces of data in the test data stream are read.