Dynamic Dictionary Encoding for Inflective Word Compression

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing encoding technologies face a decrease in compression ratio due to the increased capacity demands of static dictionaries when registering inflective forms of high-frequency words, leading to inefficient code assignment and compression.

Innovation Solution

The proposed method involves determining if encoding target data is in an inflective form and registering it in a dynamic dictionary, rather than a static dictionary, to assign dynamic codes, thereby improving the compression ratio by reducing the capacity demands on the dictionary.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If inflective forms of high-frequency words are registered in a static dictionary, then encoding accuracy is improved, but dictionary capacity requirements increase and compression ratio decreases

Engineering Contradiction:
Improveencoding accuracyVSAvoiddictionary capacity
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent segments the dictionary into two distinct parts: a static dictionary for basic word forms and a dynamic dictionary for inflective forms. This segmentation allows the static dictionary to remain compact while the dynamic dictionary handles the variable inflective forms, resolving the contradiction between encoding accuracy and dictionary capacity requirements.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a dynamic dictionary that can adaptively store inflective forms during the encoding process. This dynamic structure allows the system to handle inflective forms without permanently expanding the static dictionary, maintaining compression efficiency while improving encoding accuracy for high-frequency words with inflections.

Inventive Principle:
Principle #15Dynamics

2Productivity

If individual codes are assigned to words with high appearance frequency, then compression ratio is improved, but code assignment complexity increases when handling inflective forms

Engineering Contradiction:
Improvecompression ratioVSAvoidcode assignment complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent introduces a code conversion unit as an intermediary that translates inflective forms into their corresponding basic forms before code assignment. This mediator simplifies the code assignment process by reducing inflective forms to their base forms, which can then be efficiently encoded using the existing frequency-based coding scheme, thereby maintaining compression ratio while reducing complexity.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If inflective forms are handled using traditional encoding methods, then encoding completeness is improved, but processing time increases

Engineering Contradiction:
Improveencoding completenessVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent performs preliminary normalization of inflective forms to their basic forms during the encoding process. By pre-processing the inflective forms and storing them in the dynamic dictionary, the system avoids time-consuming processing during actual encoding operations, thus maintaining encoding completeness while reducing processing time.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS11373040B2Computer-readable recording medium, encoding device, index generating device, search device, encoding method, index generating method, and search method
Publication Date: 2022.06.28 FUJITSU LTD
  • US11373040B2 patent drawing
  • US11373040B2 patent drawing
  • US11373040B2 patent drawing

AI summary

An information processing device executes a process includes determining whether or not encoding target data is in an inflective form of a word when the encoding target data included in target sentence data is encoded; and registering the encoding target data and a code assigned to the encoding target data in a dynamic dictionary in association with each other, in a case where the encoding target data is in the inflective form of the word.