Vision-Language Transformer Pre-Training With Misalignment Distillation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing vision-language transformer pre-training methods face challenges with excessive data processing overhead and misalignment issues due to image augmentation, leading to deteriorated performance.

Innovation Solution

A method that intentionally induces misalignment between augmented images and text by random enlargement or masking, utilizing misalignment as useful information through a contrast image-text pre-training method (MCD) with knowledge distillation, aligning feature vector representations using alignment matrices and momentum-based encoders.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If image augmentation (random enlargement, cropping, masking) is applied to secure large-scale pre-training data, then the quantity of pre-training data increases, but misalignment between augmented images and text occurs leading to deteriorated model performance

Engineering Contradiction:
Improvequantity of pre-training dataVSAvoidalignment between image and text
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The patent converts the harmful misalignment effect into a beneficial learning signal by introducing a mask token that represents the augmented (misaligned) region. The model learns to predict this mask token, thereby learning to handle misalignment rather than being degraded by it. This transforms the data quality issue into a useful training objective that improves model robustness.

Inventive Principle:
Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The mask token acts as an intermediary element between the augmented image and the original text. Instead of directly comparing misaligned image regions with text (which causes performance degradation), the mask token mediates this relationship by providing a structured representation of the augmentation operation, allowing the model to learn appropriate associations.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If conventional image augmentation is used to expand dataset size, then data diversity increases, but data processing overhead becomes excessive

Engineering Contradiction:
Improvedata diversityVSAvoiddata processing efficiency
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent performs preliminary recording of augmentation operations (translation vectors, mask parameters) during data generation. This pre-recording of transformation metadata allows the model to efficiently process augmented data without requiring re-computation or complex tracking of augmentation steps during training, thereby reducing processing overhead while maintaining data diversity.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If external modules (object detector, summary extractor) are introduced to correct misalignment, then alignment accuracy improves, but device complexity and processing burden increase

Engineering Contradiction:
Improvealignment accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent enables the vision-language transformer to handle misalignment internally through the mask token mechanism and augmented image processing, without requiring external correction modules. The model learns to self-correct for misalignment by predicting mask tokens and processing augmented images directly, eliminating the need for separate object detectors or summary extractors.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4664418A9Method for pre-training vision-language transformer, and artificial intelligence system comprising vision-language transformer pre-trained through same method
Publication Date: 2026.02.11 LG MANAGEMENT DEV INST CO LTD
  • EP4664418A9 patent drawingFigure 1
  • EP4664418A9 patent drawingFigure 2
  • EP4664418A9 patent drawingFigure 3

AI summary

The present disclosure relates to a method and a system for promptly training a simplified vision-language transformer, in which large uncurated datasets are augmented through image enlargement and/or masking, etc. and vision-language transformers are pre-trained by reflecting, through a knowledge distillation framework, misaligned information between an augmented image and text upon the augmentation, thereby reducing data processing overhead.