Temporal Knowledge Base Without Reification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current knowledge base systems face inefficiencies in managing temporal knowledge due to the high overhead of reification methods, which are costly and incomplete in capturing temporal realities, especially in systems like Yago and DBpedia.
Innovation Solution
A temporal knowledge base system is introduced, comprising a Static Triple Store and a Temporal Triple Store, using a method that efficiently stores and retrieves temporal knowledge without reification, employing an Entity Table, a Predicate Table, and a Triple Table to manage temporal subjects, predicates, and objects with time references, allowing for efficient storage and querying of temporal triples.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If reification method is used to represent temporal knowledge in RDF, then temporal facts can be captured, but storage overhead increases significantly (5-fold or 9-fold triple overhead)
Solution Approach 1:
The patent extracts temporal information from the triple structure and stores it separately in dedicated temporal columns (valid_time_start, valid_time_end, transaction_time_start, transaction_time_end) within the same triple store. This separation eliminates the need for reification while preserving temporal accuracy, as temporal attributes are stored directly alongside the triple data rather than requiring additional triples to represent time information.
Solution Approach 2:
The patent adds temporal dimensions to the traditional triple structure by introducing time columns that extend the data model. Instead of using reification which adds temporal information through additional triples (horizontal expansion), the patent vertically integrates temporal information into the existing triple structure by adding time attributes to the triple representation, thereby reducing the number of triples needed while maintaining temporal expressiveness.
2Reliability
If reification is used to capture complete temporal information, then temporal realities are accurately represented, but system complexity and overhead increase
Solution Approach 1:
The patent extracts temporal information from the triple structure and stores it separately in dedicated temporal columns (valid_time_start, valid_name_end, transaction_name_id, transaction_name_id). This separation eliminates the need for reification while preserving temporal accuracy, as temporal attributes are stored directly alongside the triple data rather than requiring additional triples to represent time information.
Solution Approach 2:
The patent creates a universal temporal triple store structure that handles both valid time and transaction time within the same storage mechanism. The temporal columns can represent different time semantics (valid time for when facts are true, transaction time for when facts are recorded) using the same data structure, eliminating the need for separate reification mechanisms for different temporal scenarios.
3Quantity of substance
If traditional RDF triples are used without temporal extensions, then storage is efficient, but temporal knowledge cannot be accurately represented
Solution Approach 1:
The patent adds temporal dimensions to the traditional triple structure by introducing time columns that extend the data model. Instead of using reification which adds temporal information through additional triples (horizontal expansion), the patent vertically integrates temporal information into the existing triple structure by adding time attributes to the triple representation, thereby reducing the number of triples needed while maintaining temporal expressiveness.
Solution Approach 2:
The patent transforms the static triple structure into a dynamic one by incorporating temporal validity and transaction time columns. This allows the triple store to represent facts that change over time, with each triple having associated time ranges that define when the fact is valid and when it was recorded. The dynamic nature is achieved through time-range queries rather than through dynamic triple creation as in reification.
Data Source
AI summary
A knowledge base system is provided for storing and retrieving temporal knowledge in the form of RDF triples without reification. Any resource in an RDF triple (subject predicate object) is converted into a temporal resource that includes a name and a time period components. Hence, a temporal RDF triple (temporal subject temporal predicate temporal object) is formed. The knowledge base system includes three tables for storing temporal triples: A triple Table, An Entity Table, and a Predicate Table. A temporal triple is stored in Subject, Temporal Predicate, and Object columns of the Triple Table and in the Temporal Entity column of the Entity Table. In various embodiments, an object relational, a relational, or NoSQL DBMS or a native RDF triple store system can implement the knowledge base.