AI 용어집
인공지능 완전 사전
Fact Checking
Systematic process of evaluating the veracity of a claim by comparing it to reliable sources and factual data.
Question-Answering System
Computer application designed to provide accurate answers to questions posed in natural language, based on a knowledge base.
Evidentiary Proof
Set of sources, citations, or concrete data used by a system to support or refute a given claim.
Disinformation
Deliberate spread of false or misleading information with the intent to deceive or harm, which QA systems seek to identify.
Misinformation
Inaccurate information shared without intent to harm, often due to ignorance or misunderstanding, detectable by verification systems.
Verified Knowledge Base
Structured and validated data corpus serving as a truth reference for a fact-checking question-answering system.
Contradiction Detection
Mechanism for identifying logical inconsistencies between a claim and information contained in reference sources.
Subject-Predicate-Object Triplet
Fundamental data structure in information extraction, representing a factual relationship (e.g., 'The Eiffel Tower - is located in - Paris').
Verifiability
Measure of the degree to which a claim can be confirmed or disproved by external and reliable evidence.
Reliable Source
Document, website, or database recognized for its accuracy and impartiality, used as a reference for fact verification.
Evidence Extraction
Technique of isolating relevant sentences or passages in a corpus of documents that support or contradict a claim.
Veracity Classification
Task of categorizing a claim into predefined classes such as 'True', 'False', 'Mixed', or 'Not verifiable'.
Stance Detection
Automatic analysis of the position of a text (favorable, unfavorable, or neutral) with respect to a target claim.
Knowledge Graph
Network of entities and semantic relationships used to model and query real-world facts in a structured manner.
Multi-source Verification
Approach consisting of cross-referencing information from multiple independent sources to increase the robustness of verification.