Cognitive System Material Insight Ranking via Alignment Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cognitive computing systems often uncover numerous relationships in a knowledge base but struggle to automatically identify material insights, relying on human analysts to determine significance, and frequently highlight common knowledge rather than new or hidden relationships.
Innovation Solution
A method that calculates a 'degree of alignment' for discovered relationships based on evidence, allowing the system to prioritize and present material insights, using metrics such as recentness and alignment with general knowledge, and allowing user-defined criteria to adjust the importance of relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If cognitive systems highlight the strongest relationships based on common knowledge, then the system provides reliable and well-known information, but it fails to identify material insights and new or hidden relationships
Solution Approach 1:
The patent segments the evaluation of relationships into multiple independent dimensions: (1) strength of relationship based on corpus evidence, (2) degree of alignment with commonly known relationships, and (3) recency of the relationship. This segmentation allows the system to separately evaluate and combine these factors to identify material insights that would be missed by focusing on a single dimension.
Solution Approach 2:
The patent introduces a 'degree of alignment' parameter that measures how well a discovered relationship matches commonly known relationships. By changing the evaluation parameter from simply relationship strength to include alignment degree, the system can distinguish between well-known relationships and material insights, resolving the contradiction between reliability and information loss.
2Productivity
If the system ranks relationships by strength alone, then the processing is simple and fast, but the system cannot prioritize material insights over common knowledge
Solution Approach 1:
The patent performs preliminary action by pre-computing and storing the degree of alignment for each relationship with commonly known relationships in a dictionary. This pre-computation allows the system to quickly retrieve alignment scores during relationship ranking without performing complex analyses in real-time, maintaining processing speed while improving measurement precision.
Solution Approach 2:
The patent introduces an intermediary mechanism - the 'degree of alignment' metric and dictionary - that mediates between the simple relationship strength measurement and the complex task of identifying material insights. This intermediary provides a computationally efficient way to evaluate whether a relationship is merely common knowledge or represents a material insight.
3Measurement precision
If human analysts manually determine significance of relationships, then the system can identify material insights accurately, but the system requires extensive human intervention and reduced automation
Solution Approach 1:
The patent implements self-service by enabling the cognitive system to automatically evaluate and rank relationships using the degree of alignment metric and recency information. The system serves itself by comparing discovered relationships against the pre-stored dictionary of commonly known relationships, eliminating the need for human analysts to manually determine significance while maintaining high accuracy in identifying material insights.
4Quantity of substance
If the system evaluates all discovered relationships equally, then the system maintains comprehensive coverage, but the system cannot prioritize new or timely information
Solution Approach 1:
The patent applies dynamics by making the relationship ranking adaptive and time-sensitive. The system dynamically adjusts the importance of relationships based on their recency and degree of alignment, allowing the ranking criteria to change over time. This dynamic approach enables the system to prioritize new or timely information while maintaining comprehensive coverage of all discovered relationships.
Data Source
AI summary
Certain relationships representing material insights are identified from among a set of discovered relationships. Cognitive discovery of relationships in a knowledge base, or corpus, are ranked according to one or more metrics indicative of material insights, including recentness and degree of alignment.