Discover, Connect & Collaborate at TECHINNOVATION 2021

Visual Mode and Algorithms for Eliciting Preferences

Technology Overview

When dealing with a large collection of products or points-of-interests, users often benefit from recommender systems. For instance, when choosing a restaurant, customers need to view long lists of restaurants to even sense what is relevant to them. To provide accurate recommendations, the systems need to model user preferences based on observational data. This technology is a know-how for efficiently eliciting user preferences via a visual interface that allows them to intuitively navigate different options, and for processing the obtained preference signals into recommendations based on machine learning. The faster we can elicit user preferences, the better the user experience with the recommender system will be. A demonstrational prototype is ready for points-of-interest such as restaurants, and the same technology could be applied to other product categories. We are seeking partners for collaboration to bring the technology into a market-ready consumer-facing application.

Technology Features, Specifications and Advantages

There are two main features of the technology.

One is a mechanism of eliciting user preferences via a visual interface, which involves identifying potentially preferred items to users, allowing users to efficiently indicate their actual preferences, recording those interactions for both for users’ own use as well as for the system to improve recommendations.

Another feature is the implementation of machine learning algorithms that improve matching of users to objects or restaurants they desire without having to go through them individually. This continually improves over time as users interact with the system, and as more users enter the system.

Potential Applications

The primary application area is to aid recommendations where customers interact with the system via a visual interface, e.g., a mobile app or a web application.

  • There is a ready demonstrational prototype of a visual application for browsing and recommending restaurants as well as food establishments.
  • The same technology applies to other domains where purchases are informed by users’ subjective ‘tastes’, such as fashion, books, etc.

Customer Benefit

There are several benefits to consumers.

  • Efficient selection of preferred items at reduced time
  • Appealing visual interface for indicating personal preferences
  • More targeted segments for advertisers
  • Social and economic benefits from increasing awareness of different providers
Contact Person

Francis Lim


Singapore Management University - School of Computing and Information Systems

Technology Category

  • Infocomm
  • Artificial Intelligence, Enterprise & Productivity, Human-Computer Interaction, Interactive Digital Media & Multimedia, Video/Image Analysis

Technology Readiness Level


visual, recommendation, food, image recognition, algorithm, preference, understanding users, machine learning, image analysis, relevance, relevant