Multi-Modal Multi-Task AI Processing System

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current AI technology is limited in processing complex tasks due to its reliance on unimodal data, resulting in weak generalization ability and difficulty in applying AI models to various application scenarios.

Innovation Solution

A system and method for multi-modal multi-task processing that includes a task representation component to define tasks in a unified format, a data conversion component to determine encoding sequences, and a data processing component to process tasks across different modalities, enabling the processing of multiple tasks simultaneously and improving generalization ability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If AI models are trained based on unimodal data, then the model structure is simple, but the generalization ability is weak and difficulty in applying to various complex application scenarios

Engineering Contradiction:
Improvegeneralization abilityVSAvoidmodel complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent introduces a unified task representation framework that can represent multiple types of tasks (image processing, text processing, audio processing, etc.) using a common data structure and processing mechanism. This allows a single AI model to handle diverse tasks across different modalities, achieving multi-functionality without requiring separate specialized models for each task type.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent changes the input data format from traditional unimodal data to a unified task representation format that includes task description information, task input information, and task output information. By transforming the data parameters into this standardized format, the system enables the same model to process different modalities and task types effectively.

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If AI technology deals with simple tasks of single tasks, small tasks or similar tasks, then the task processing is straightforward, but the application field is limited

Engineering Contradiction:
Improveapplication fieldVSAvoidtask processing complexity
Core Design Contradiction:
Adaptability or versatilityVSEase of operation

Solution Approach 1:

The patent segments the task processing into distinct components: task representation (defining what needs to be processed), data conversion (converting to appropriate format), and processing (executing the task). This segmentation allows complex tasks to be broken down into manageable steps while maintaining a consistent processing framework.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a data conversion component as an intermediary between the unified task representation and the actual processing. This intermediary layer translates the standardized task representation into modality-specific formats, enabling the system to handle diverse tasks without increasing the complexity of the core processing logic.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Stability of the object's composition

If a unified task representation framework is used to describe tasks in different modalities, then the task processing consistency is improved, but the data conversion complexity increases

Engineering Contradiction:
Improvetask processing consistencyVSAvoiddata conversion complexity
Core Design Contradiction:
Stability of the object's compositionVSDevice complexity

Solution Approach 1:

The patent applies local quality by designing modality-specific data conversion modules that handle different types of data (images, text, audio) with appropriate conversion techniques. Each conversion module is specialized for its specific modality, ensuring accurate and efficient conversion while maintaining the overall unified framework.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4300366A1Method, apparatus, and system for multi-modal multi-task processing
Publication Date: 2024.01.03 ALIBABA DAMO (HANGZHOU) TECH CO LTD
  • EP4300366A1 patent drawingFigure 1
  • EP4300366A1 patent drawingFigure 2
  • EP4300366A1 patent drawingFigure 3~4

AI summary

A system for multi-modal multi-task processing includes a task representation component configured to determine a task representation element corresponding to a task representation framework that is used to define a content format for describing a to-be-processed task, and the task representation element including an element used to define task description information, an element used to define task input information, and an element used to define task output information; and based on the task representation element, acquire task description information, task input information, and task output information corresponding to each of to-be-processed tasks in different modalities; a data conversion component configured to determine an encoding sequence corresponding to each of the to-be-processed tasks; and a data processing component configured to process each of the to-be-processed tasks based on the encoding sequence corresponding to each of the to-be-processed tasks to obtain a task processing result corresponding to each of the to-be-processed tasks.