Tagging Model for O2O Query Term Filtering

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Online to offline (O2O) services, such as online taxi-hailing, face inefficiencies when users input queries with irrelevant information, leading to failure in recalling points of interest (POIs) and requiring repeated modifications, which is time-consuming.

Innovation Solution

A system and method for training a tagging model that processes user queries by segmenting terms, labeling relevant and irrelevant terms, and using a trained model to retrieve POIs from a database, enhancing query efficiency by automatically identifying and omitting irrelevant terms.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a user inputs a query with irrelevant information, then the query cannot recall any POI, but requiring repeated modifications is time consuming

Engineering Contradiction:

Improvequery accuracyVSAvoidtime for query modification

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system automatically identifies and removes irrelevant terms from the query without requiring user intervention. The tagging model autonomously processes the query, labels terms as relevant or irrelevant, and generates a cleaned query that can recall POIs, making the system self-correcting rather than relying on repeated user modifications

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The manual process of users repeatedly modifying queries is replaced by an automated computational system. The tagging model uses machine learning algorithms to automatically analyze query terms, determine their relevance to POIs, and filter out irrelevant information, substituting mechanical user actions with intelligent automated processing

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If the query includes irrelevant terms, then the POI recall fails, but manual query modification reduces recall efficiency

Engineering Contradiction:

ImprovePOI recall efficiencyVSAvoidquery input simplicity

Core Design Contradiction:

ProductivityVSEase of operation

Solution Approach 1:

The tagging model acts as an intermediary between the user's raw query and the POI database. It receives the original query, automatically identifies and removes irrelevant terms through intelligent analysis, and passes the cleaned query to the POI recall system, thereby maintaining both operational simplicity and recall efficiency

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary processing of the query before POI recall is attempted. By pre-identifying and removing irrelevant terms through the tagging model, the query is optimized in advance, ensuring that the subsequent POI recall operation can proceed efficiently without failure

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS11093531B2Systems and methods for recalling points of interest using a tagging model

Publication Date: 2021.08.17 BEIJING DIDI INFINITY TECH & DEV CO LTD

AI summary

The present disclosure relates to systems and methods for recalling one or more POIs. The systems and methods may receive a query inputted by a user via a terminal and determine one or more terms by processing the received query. The systems and methods may further label the one or more terms using a trained tagging model. The systems and methods may determine one or more POs based on the labeled one or more terms, and display the one or more terms on the terminal.