Trained Concept Learning for Digital Media Retrieval

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing digital media retrieval systems fail to meet user expectations by not effectively learning and retrieving content relevant to specific user criteria, despite using sophisticated pattern recognition systems.

Innovation Solution

A system and method that learns new trained concepts by obtaining digital media items, receiving feedback on positive and negative examples, and using machine learning representations to determine a trained concept for retrieving relevant content, enabling the system to improve its classification and retrieval capabilities over time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If sophisticated pattern recognition systems are used to retrieve digital media items, then the system can recognize common patterns (e.g., dog in a picture, genre of music), but the system fails to meet user expectations for retrieving content based on specific user criteria

Engineering Contradiction:

Improveability to retrieve content based on specific user criteriaVSAvoidprecision of concept recognition

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The system receives feedback indicating whether retrieved digital media items are relevant to the user's concept. This feedback is used to iteratively refine the trained concept representation, improving the precision of concept recognition while maintaining adaptability to user-specific criteria.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system performs preliminary actions by obtaining individual representations of digital media items and using them to initialize the trained concept before feedback is received. This allows the system to have a starting point for concept recognition that can then be refined through feedback.

Inventive Principle:

Principle #10Preliminary action

2Adaptability or versatility

If the system uses a fixed set of recognizable patterns for content retrieval, then the system can operate with established classification rules, but the system cannot learn new concepts specific to individual users

Engineering Contradiction:

Improveability to learn new user-specific conceptsVSAvoidcomplexity of learning system

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system uses a universal machine learning framework that can handle both fixed pattern recognition and new concept learning. The same trained concept representation and feedback mechanism work for both predefined patterns and user-specific concepts, reducing the need for separate complex systems.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system automatically learns new user-specific concepts by processing feedback without requiring manual reprogramming or complex configuration. The feedback-driven refinement process enables the system to self-adapt to user needs, reducing the complexity of system customization.

Inventive Principle:

Principle #25Self-service

3Measurement precision

If the system retrieves digital media items using existing classification systems, then retrieval operations can be performed quickly, but the system cannot discriminate information based on newly learned user concepts

Engineering Contradiction:

Improvediscrimination accuracy of contentVSAvoidtime for concept learning and retrieval

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary actions by obtaining individual representations of digital media items and initializing the trained concept before actual retrieval operations. This preparation work is done once and can be reused for multiple retrieval operations, reducing time loss in subsequent operations.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The feedback mechanism enables the system to quickly refine concept discrimination accuracy by learning from user responses. Each feedback iteration improves the trained concept representation, allowing the system to achieve high discrimination accuracy without requiring extensive pre-processing time.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS10331675B2Systems and methods for learning new trained concepts used to retrieve content relevant to the concepts learned

Publication Date: 2019.06.25 NEBIUS BV

AI summary

A system configured for learning new trained concepts used to retrieve content relevant to the concepts learned. The system may comprise one or more hardware processors configured by machine-readable instructions to obtain one or more digital media items. The one or more hardware processors may be further configured to obtain an indication conveying a concept to be learned from the one or more digital media items. The one or more hardware processors may be further configured to receive feedback associated with individual ones of the one or more digital media items. The one or more hardware processors may be configured to obtain individual neural network representations for the individual ones of the one or more digital media items. The one or more hardware processors may be configured to determine a trained concept based on the feedback and the neural network representations of the one or more digital media items.