Partial Query Prediction System for Online Dictionaries

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing online dictionary services lack efficient methods for predicting and suggesting complete queries based on partial user input, particularly across languages, and do not effectively handle misspellings or provide personalized results.

Innovation Solution

A distributed system that processes partial search queries by mapping them to a chunk table and token table, using a prediction database and dictionary database to generate and rank predicted complete queries, which includes handling misspellings and personalization based on user query logs and language pair selection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If online dictionary services require users to enter complete words or terms for translation, then translation accuracy is improved, but user operation time and convenience deteriorate

Engineering Contradiction:
Improvetranslation accuracyVSAvoiduser operation time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system pre-computes and stores prediction results for partial queries in advance. When a user types a partial query, the system retrieves pre-computed predictions from the prediction database, eliminating the need for real-time computation and enabling fast suggestions before the user completes typing.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system introduces a prediction interface as an intermediary between the user's partial input and the complete query submission. This prediction interface provides suggested completions that guide users toward accurate translations without requiring them to manually type the complete query, thus maintaining accuracy while reducing time.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Ease of operation

If the system provides prediction suggestions for partial queries, then user convenience is improved, but system complexity increases

Engineering Contradiction:
Improveuser convenienceVSAvoidsystem complexity
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

The system segments the query processing functionality into distinct modules: a prediction module that handles partial query suggestions, a translation module that processes complete queries, and separate database components (prediction database and dictionary database). This modular architecture manages complexity by dividing functions into independent, manageable units.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The prediction interface acts as an intermediary layer between the user and the complex translation system. It simplifies user interaction by providing guided completions while managing the complexity of query processing, database searches, and translation operations in the background without exposing this complexity to the user.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If the system uses comprehensive databases for query predictions, then prediction accuracy is improved, but data processing time increases

Engineering Contradiction:
Improveprediction accuracyVSAvoiddata processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system pre-computes prediction results for partial queries and stores them in advance in the prediction database. This preliminary computation eliminates the need for real-time analysis of the complete dictionary database when providing suggestions, enabling fast retrieval based on pre-processed data while maintaining high prediction accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system segments the large dictionary database into a smaller, optimized prediction database that contains only the necessary prediction data for partial queries. This segmentation reduces the amount of data that needs to be processed in real-time while maintaining prediction accuracy by storing pre-computed results.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP2313838B1Dictionary suggestions for partial user entries
Publication Date: 2019.05.22 GOOGLE LLC
  • EP2313838B1 patent drawingFigure 1
  • EP2313838B1 patent drawingFigure 2
  • EP2313838B1 patent drawingFigure 3

AI summary

In accordance with some embodiments, a server receives a partial search query from a client device that is located remotely from the server, and predicts from the partial search query a set of predicted complete queries relevant to the partial search query, where the predicted complete queries comprise previously submitted complete queries submitted by a community of users. The partial search query and the set of predicted complete queries are in the same language. In addition, the server obtains translations of at least a subset of the set of predicted complete queries, where the translations are in a different language from the set of predicted complete queries, and conveys both the set of predicted complete queries and the corresponding translations to the client device.