Blockchain Fact Verification Using IPFS and Named Entity Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing blockchain technology is ineffective in verifying the factual accuracy of document fragments or media content, such as manipulated images or videos, as it can only trace the origin of a document but not verify the truthfulness of its claims.
Innovation Solution
A method utilizing a private Inter Planetary File System (IPFS) for document storage, combined with blockchain and machine learning algorithms, to identify named entities, retrieve relevant documents, and perform verification by analyzing text style, context, and grammar, providing a verdict with supporting evidence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If blockchain technology is used to trace document origins by storing hash codes, then document origin tracing capability is improved, but the ability to verify factual accuracy of document claims remains ineffective
Solution Approach 1:
The patent merges blockchain technology with machine learning algorithms and IPFS storage to create a hybrid system that combines the strengths of each component: blockchain provides immutable origin tracing, IPFS provides decentralized document storage, and machine learning algorithms provide factual accuracy verification. This combination resolves the contradiction by adding factual verification capability while preserving origin tracing reliability.
Solution Approach 2:
The patent introduces machine learning algorithms as an intermediary layer between blockchain and factual verification. The algorithms process document claims, compare them against known facts, and generate verification results, thereby enabling factual accuracy verification without compromising the blockchain's origin tracing function.
2Adaptability or versatility
If machine learning algorithms are used to verify document claims, then factual accuracy verification is improved, but system complexity increases
Solution Approach 1:
The patent segments the system into distinct functional modules: IPFS for document storage, blockchain for origin tracking and metadata storage, and machine learning algorithms for factual verification. This segmentation allows each component to perform its specific function independently, making the overall complex system more manageable and maintainable.
Solution Approach 2:
The patent designs the system so that the machine learning algorithms can handle multiple types of verification tasks (claim verification, entity validation, relationship checking) using a unified architecture, reducing the need for separate specialized systems for each verification function.
3Measurement precision
If multiple named entities and metadata are extracted and processed, then verification accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary extraction and indexing of named entities and their relationships from authenticated documents before verification is needed. This pre-processing creates a ready-to-query knowledge base that can be quickly accessed during actual verification, reducing processing time while maintaining high accuracy.
Solution Approach 2:
The patent focuses the machine learning algorithms on verifying only the specific claims and entities relevant to each document, rather than processing all possible information. By targeting verification efforts at local areas of highest importance (key claims and named entities), the system achieves high accuracy with reduced processing time.
Data Source
AI summary
A method for performing fact verification includes receiving a document for verification; identifying and extracting, on the computer network, at least two named entities from the document for verification and associated information as metadata; identifying, off-chain, identifiers of relevant documents corresponding to each of the at least two named entities, the relevant documents being authenticated documents including one or more of the at least two named entities; identifying a predetermined number of relevant documents among the identified relevant documents based on a number of the at least two named entities present and a number of occurrences for each of the named entities; and determining whether the document to be verified is supported by the predetermined number of relevant documents.