Unified Image Comparison Interface for Entity Recommendation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Users face difficulties in efficiently selecting and comparing images of entities like hotels and restaurants due to the need to visit multiple webpages and the lack of a unified interface for image comparison, leading to inconvenient browsing experiences.

Innovation Solution

A system utilizing an image engine that tags images with machine learning, selects images based on user preference models, and displays them in a single interface, allowing users to provide interest indicators for personalized recommendations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If users visit multiple webpages to view images of different entities, then they can see more detailed information about each entity, but the browsing time and operational complexity increase significantly

Engineering Contradiction:
Improveimage information completenessVSAvoidbrowsing time
Core Design Contradiction:
Loss of informationVSLoss of time

Solution Approach 1:

The patent merges images from multiple entities into a single unified interface, allowing users to view and compare images of different hotels, restaurants, or other entities without navigating to separate webpages. This consolidation directly reduces browsing time while maintaining access to comprehensive image information across all entities.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent introduces a new dimensional organization by grouping images according to common features (such as pool images, room images, or dish images) rather than organizing them by entity. This dimensional shift allows users to compare specific features across multiple entities simultaneously in one view, eliminating the need to visit multiple webpages.

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

2Loss of information

If users visit multiple webpages to compare images across entities, then they can access comprehensive image sets, but the ease of operation deteriorates due to multiple clicks and navigation

Engineering Contradiction:
Improveimage availabilityVSAvoidbrowsing convenience
Core Design Contradiction:
Loss of informationVSEase of operation

Solution Approach 1:

The system combines image displays from multiple entities into a single interface, eliminating the need for users to navigate between webpages. This merging maintains comprehensive image availability while dramatically improving ease of operation by allowing users to view and interact with all images in one location.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent segments images by their content features (such as separating pool images, room images, and dining images) rather than by entity. This segmentation allows users to easily find and compare specific feature images across entities without dealing with the complexity of multiple webpages, improving both information access and operational ease.

Inventive Principle:
Principle #1Segmentation

3Device complexity

If traditional search interfaces show only basic entity information, then the interface remains simple, but the ability to provide personalized recommendations based on user interests is limited

Engineering Contradiction:
Improveinterface simplicityVSAvoidpersonalization capability
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent implements feedback mechanisms where user interactions with images (such as selections, clicks, or time spent viewing) are captured and used to update user preference models. This feedback loop enables the system to progressively improve personalized recommendations while maintaining interface simplicity, as the personalization occurs automatically in the background.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system performs automatic image selection and presentation based on user preferences without requiring complex user configuration. The preference models and image selection algorithms operate autonomously, adapting to user interests through observed behavior rather than through complex setup procedures, thus maintaining interface simplicity while enhancing adaptability.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS20210383455A1System and method for recommending entities based on interest indicators
Publication Date: 2021.12.09 MICROSOFT TECHNOLOGY LICENSING LLC
  • US20210383455A1 patent drawing
  • US20210383455A1 patent drawing
  • US20210383455A1 patent drawing

AI summary

Entities such as hotels, restaurants, resorts, houses, vehicles, and other places and things, are associated with images of each entity. The images are tagged using machine learning to identify what aspects of the associated entity are captured by each image. When a user is requested to select an entity from a set of entities, a user preference model is used to determine what tags the user is interested in. The tags are used to select images associated with the entities from the set of entities, and the selected images are displayed to the user. The user can then provide indicators that show which of the displayed images the user likes or dislikes. Based on the indicators, one or more entities from the set of entities is presented to the user. The model may also be updated based on the indicators.