Knowledge Graph Sub-Graphs for Explainable Recommendations

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing recommendation systems lack transparency and interpretability, leading to poor user understanding of recommendation results, which affects their effectiveness.

Innovation Solution

A knowledge graph-based approach that generates sub-graphs representing user interactions, providing structured explanations for recommendation results, enhancing transparency and accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional recommendation algorithms are used, then recommendation speed and coverage are improved, but transparency and interpretability deteriorate

Engineering Contradiction:
Improverecommendation speedVSAvoidinterpretability
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent introduces knowledge graphs as an intermediary between traditional recommendation algorithms and users. The knowledge graph structures recommendation reasoning processes with explicit entities, relationships, and paths, serving as a mediator that preserves algorithmic efficiency while enhancing interpretability through visualizable reasoning trails.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces the black-box mechanical recommendation system with a knowledge graph-based semantic reasoning system. This substitution transforms opaque algorithmic operations into interpretable logical inference processes based on explicit knowledge representations, maintaining productivity while resolving the interpretability loss.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If complex recommendation algorithms are deployed, then recommendation accuracy is improved, but system complexity increases

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

Solution Approach 1:

The patent segments the complex recommendation system into modular components: knowledge graph construction module, sub-graph extraction module, and recommendation generation module. Each component handles specific tasks with clear interfaces, reducing overall system complexity while maintaining recommendation accuracy through specialized processing at each stage.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the fundamental parameter of recommendation from direct algorithmic scoring to knowledge graph-based semantic matching. This parameter transformation shifts the system from complex statistical modeling to structured logical reasoning, improving accuracy through explicit knowledge representation while simplifying the underlying computational mechanisms.

Inventive Principle:
Principle #35Parameter changes

3Adaptability or versatility

If more data is processed, then recommendation personalization is improved, but information overload for users increases

Engineering Contradiction:
ImprovepersonalizationVSAvoidinformation overload
Core Design Contradiction:
Adaptability or versatilityVSObject-affected harmful factors

Solution Approach 1:

The patent extracts only the most relevant information for personalized recommendations by querying the knowledge graph for specific user-related sub-graphs. This extraction process filters out unnecessary data while preserving essential personalization signals, delivering tailored recommendations without overwhelming users with excessive information.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent transitions from processing vast amounts of raw user data in traditional dimensions to representing user preferences and item relationships in the semantic dimension of the knowledge graph. This dimensional transformation enables effective personalization through structured semantic matching while naturally limiting information overload through the graph's inherent hierarchical organization.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP4679292A1Methods, apparatuses, devices, and computer program products for recommendation
Publication Date: 2026.01.14 ROBERT BOSCH GMBH
  • EP4679292A1 patent drawingFigure 1
  • EP4679292A1 patent drawingFigure 2
  • EP4679292A1 patent drawingFigure 3

AI summary

Examples of the present disclosure relate to methods, apparatuses, devices, and computer program products for recommendation. The method comprises obtaining a knowledge graph comprising a plurality of user nodes and a plurality of object nodes. The method further comprises generating one or more sub-graphs based on the knowledge graph, wherein the sub-graph comprises a user node corresponding to the first user and related object nodes among a plurality of user nodes. In addition, the method further comprises determining a recommendation result for the first user based on the one or more sub-graphs and the plurality of object nodes. The solution provided by the examples of the present disclosure enables the generation of sub-graphs representing the recommendation explanation when generating recommendation results, which not only improves the effectiveness of the recommendation, but also enhances the transparency and interpretability of the recommendation system, allowing users to better understand and accept the recommendation results and thereby improving the user experience of the recommendation system.