Triangulated Semantic Vector Forecasting for Evidence Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current databases with summarized articles lack trend and forecasted states, leading to imprecise retrieval of supporting or refuting evidence, with large-scale evidence retrieval being recall-oriented and prone to false positives, especially when dealing with trending words over time.
Innovation Solution
A method is developed to generate triangulated current and forecasted keywords, concepts, and entities by projecting semantic meanings into word vectors, regressing vector values over time to produce forecasted vectors, and averaging static vectors, which are then used to calculate the volume of a triangle formed by forecasted, static, and bootstrapped vectors to determine the best evidence for user inquiries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If large-scale evidence retrieval is performed using traditional databases with summarized articles, then the quantity of retrieved evidence increases, but the precision and relevance of the evidence decreases due to recall-oriented approach and false positives
Solution Approach 1:
The patent transforms static article summaries into dynamic forecasted semantic representations by changing the temporal parameter. It uses vector regression to project current semantic states into future states, allowing the system to retrieve evidence based on forecasted trends rather than static historical data, thereby improving precision while maintaining quantity
Solution Approach 2:
The patent introduces a temporal dimension to traditional evidence retrieval by creating forecasted vectors that represent future semantic states. This transforms the retrieval from a static spatial search to a dynamic spatiotemporal search, enabling the system to identify trending concepts and reduce false positives while maintaining comprehensive coverage
2Device complexity
If traditional static vector representations are used for article retrieval, then the system complexity remains low, but the system cannot capture trending words and forecasted states over time
Solution Approach 1:
The patent transforms static vector representations into dynamic forecasted vectors by applying temporal regression. The system continuously updates semantic representations to reflect current trends and forecasted states, allowing it to adapt to changing contexts while maintaining a relatively simple architectural framework
Solution Approach 2:
The patent performs preliminary vector regression to forecast future semantic states before actual retrieval occurs. By pre-calculating forecasted vectors and identifying trending concepts in advance, the system prepares adaptive retrieval capabilities without requiring complex real-time processing during query execution
Data Source
AI summary
Computer-implemented method includes developing, via a processor, a words model from a plurality of natural language text based articles relating to a subject and generating, via the processor, a static vector based upon the words model. The computer-implemented method further includes developing, via the processor, an actual articles model from actual articles, generating, via the processor, a bootstrapped vector using the actual articles model, generating, via the processor, a n-dimensional depth item using the static vector and the bootstrapped vector, and determining, via the processor, evidence based on the n-dimensional depth item. The computer-implemented method still further includes presenting, via the processor and a display, the evidence base upon an input query from a user.