Bootstrapped Knowledge Acquisition for Domain Entity-Action Compatibility

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current computing systems face challenges in acquiring and utilizing domain knowledge, particularly in understanding real-world common sense knowledge, such as entity-action compatibility, due to limited natural language processing capabilities and the need for extensive manual data population, leading to sparse knowledge coverage and errors.

Innovation Solution

A method for bootstrapping a small amount of manually acquired domain knowledge to automatically build a larger set of domain knowledge through automated learning processes, using pre-condition annotated action terms to infer attributes of entities from natural language content, expanding the knowledge base with feature correlations and compatibility information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual data population is used to build domain knowledge base, then knowledge accuracy can be ensured, but human effort and time consumption increase significantly

Engineering Contradiction:

Improveknowledge accuracyVSAvoidtime consumption

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary action by automatically inferring entity attributes and action compatibility relationships from existing natural language content before manual verification. This pre-population reduces the scope of manual data entry while maintaining accuracy through subsequent validation steps.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements self-service by enabling automated bootstrap learning that uses pre-condition annotated action terms to independently infer domain knowledge without continuous manual intervention. The system serves itself by learning from existing content and expanding the knowledge base autonomously.

Inventive Principle:

Principle #25Self-service

2Loss of information

If extensive manual data population is performed to achieve comprehensive knowledge coverage, then knowledge completeness improves, but complexity of the system increases

Engineering Contradiction:

Improveknowledge coverageVSAvoidsystem complexity

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The system introduces an intermediary mechanism - pre-condition annotated action terms - that mediates between natural language content and the domain knowledge base. This intermediary enables automated inference of entity attributes and relationships, reducing the need for direct manual population of all knowledge elements.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system achieves multi-functionality by using the same automated inference mechanism to populate multiple types of knowledge elements simultaneously - entity attributes, action compatibility relationships, and procedural knowledge - thereby improving coverage without proportionally increasing system complexity.

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Productivity

If automated learning processes are used to expand knowledge base, then productivity increases, but measurement precision of knowledge attributes may decrease

Engineering Contradiction:

Improveknowledge expansion speedVSAvoidattribute accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system implements feedback by validating automatically inferred knowledge against existing domain constraints and rules. The bootstrap learning process uses pre-condition annotations as feedback signals to verify inferred attributes, ensuring accuracy while maintaining high productivity through automated operations.

Inventive Principle:

Principle #23Feedback

4Reliability

If common sense knowledge is incorporated to constrain entity actions, then reliability of generated content improves, but device complexity increases due to additional knowledge requirements

Engineering Contradiction:

Improvecontent reliabilityVSAvoidknowledge base complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system applies local quality by implementing constraints specifically at the entity-action interface rather than throughout the entire knowledge base. Pre-condition annotations are applied locally to action terms, enabling reliability improvement for specific entity-action pairs without requiring global complexity increases.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS10832591B2Evaluating user responses based on bootstrapped knowledge acquisition from a limited knowledge domain

Publication Date: 2020.11.10 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

Mechanisms for training a human user to perform an operation and provided. The mechanisms generate a domain specific knowledge base comprising a set of entities and corresponding domain specific attributes and expand the domain specific knowledge base to include values for the domain specific attributes through an automated bootstrap learning process that performs natural language processing and analysis of natural language content using a set of pre-condition annotated action terms, thereby generating an expanded domain specific knowledge base. The mechanisms evaluate an input from another device identifying an action associated with an entity in the set of entities, based on a retrieved domain specific attribute value and the retrieved pre-condition annotation from the expanded domain specific knowledge base. The mechanisms output a notification to a user computing device indicating whether the input is correct or incorrect to thereby train a user associated with the user computing device.