Contextual Concept Inference for NLP Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current natural language processing systems are unable to infer the meaning of unknown concepts from surrounding context, limiting their ability to understand concepts at the sentence or paragraph level without requiring initial training or retraining.
Innovation Solution
A computer-implemented method that identifies sentences associated with unknown concepts, compares them to similar chains of known concepts, and infers meaning by measuring contextual similarity, allowing for autonomous and incremental learning of unknown concepts without the need for initial training.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If current natural language processing systems use contextual-based methods to predict missing parts from word sequence patterns, then they can process known concepts, but they cannot effectively understand concepts from unknown sentences or paragraphs
Solution Approach 1:
The patent moves the analysis from the word/sentence level to the concept level. Instead of predicting missing words from surrounding text, the system extracts concepts from sentences and compares concept chains across different textual dimensions (multiple sentences, multiple documents) to infer unknown concepts. This dimensional shift enables the system to handle unknown concepts by operating at a higher level of abstraction where patterns can be recognized across broader contexts.
Solution Approach 2:
The patent introduces concepts as an intermediary layer between raw text and meaning interpretation. By extracting concepts from sentences and using these concepts as mediators for comparison and inference, the system can bridge the gap between unknown text and known meanings. The concept chain serves as an intermediary structure that allows indirect reasoning about unknown concepts through their relationships with known concepts.
2Adaptability or versatility
If the system compares concept chains from multiple sentences to infer unknown concepts, then it achieves better understanding, but the computational complexity increases
Solution Approach 1:
The patent segments the text processing task into distinct stages: sentence-level concept extraction, chain-level comparison, and inference-generation. By dividing the complex task of understanding unknown concepts into these manageable segments, the system can process information incrementally rather than attempting to analyze entire documents at once. This segmentation reduces computational complexity by breaking down the problem into smaller, more tractable sub-problems.
Solution Approach 2:
The patent performs preliminary concept extraction from all sentences before attempting to infer unknown concepts. By pre-processing the text to identify and extract concepts from known sentences, the system creates a ready-to-use concept database that can be quickly queried during inference. This preliminary action avoids the need to re-analyze entire sentences during the inference process, significantly reducing computational complexity.
Data Source
AI summary
A computer-implemented method for learning unknown concepts during natural language processing is disclosed, including identifying a sentence associated with an unknown concept, selecting a first sequential set of sentences from a first document, including the sentence associated with the unknown concept, one sentence prior, and subsequent to the sentence associated with the unknown concept, selecting a second sequential set of sentences from a second document, including a sentence associated with a known concept, and one sentence prior and subsequent to the sentence associated with the known concept, comparing concepts associated with the first sequential set of sentences and second sequential set of sentences, determining whether an inference can be made between the unknown concept associated with the sentence from the first document and the sentence associated with the known concept associated with the sentence from the second document, and tagging the unknown concept associated with the known concept.