17747463. AUTOMATED FACT CHECKING USING ITERATIVE KNOWLEDGE BASE QUERYING simplified abstract (INTERNATIONAL BUSINESS MACHINES CORPORATION)
AUTOMATED FACT CHECKING USING ITERATIVE KNOWLEDGE BASE QUERYING
Organization Name
INTERNATIONAL BUSINESS MACHINES CORPORATION
Inventor(s)
G P Shrivatsa Bhargav of Bengaluru (IN)
Saswati Dana of Bangalore (IN)
Dinesh Khandelwal of Indore (IN)
Dinesh Garg of Ajmer Road (IN)
AUTOMATED FACT CHECKING USING ITERATIVE KNOWLEDGE BASE QUERYING - A simplified explanation of the abstract
This abstract first appeared for US patent application 17747463 titled 'AUTOMATED FACT CHECKING USING ITERATIVE KNOWLEDGE BASE QUERYING
Simplified Explanation
The patent application describes a method for decomposing a natural language assertion into a question and answer pair. This involves translating the question into a structured knowledge graph query and performing an iterative process of querying the knowledge graph and evaluating the responses.
- The method decomposes a natural language assertion into a question and answer pair.
- It translates the question into a structured knowledge graph query.
- The method performs an iterative process of querying the knowledge graph and evaluating the responses.
- In each iteration, the method retrieves a predicted answer and checks its similarity with the initial answer.
- If the similarity does not meet a threshold criterion, the query is altered and used for the next iteration.
- The method also generates an assertion correctness score based on the confidence scores obtained during the iterative process.
Potential Applications
- Natural language processing systems
- Question-answering systems
- Knowledge graph-based search engines
Problems Solved
- Difficulty in decomposing natural language assertions into question and answer pairs
- Inefficient querying of knowledge graphs for accurate answers
- Lack of confidence scoring for factual assertions
Benefits
- Improved accuracy in decomposing natural language assertions
- Efficient querying of knowledge graphs for accurate answers
- Confidence scoring to assess the correctness of assertions
Original Abstract Submitted
An embodiment includes decomposing a natural language assertion into a natural language question and answer pair that includes an initial question and an initial answer. The embodiment translates the initial question into a structured knowledge graph query and then performs an iterative process comprising iterative querying of a knowledge graph and evaluating of corresponding query responses resulting in respective confidence scores. A first iteration of the iterative process comprises querying of the knowledge graph to retrieve a first predicted answer, then determining whether a degree of similarity between the initial answer and the first predicted answer meets a threshold criterion. If not, the first predicted query is altered and used for querying the knowledge graph in a subsequent iteration of the iterative process. The embodiment also generates an assertion correctness score indicative of a degree of confidence that the assertion is factual using the respective confidence scores.
