Merchantry Knowledge

Stable public release

edition:knowledge-systems

Browse the canonical public views and the evidence-eligible claims bound to release release:edition:knowledge-systems:7aaba55a11d29659.

4 claims

Knowledge representation

A knowledge graph can represent interlinked entities, events, concepts, their properties and types, and semantic relations.
Knowledge graphs may use ontologies as a schema layer to support logical inference and retrieve implicit knowledge.
A virtual knowledge graph can answer graph queries over existing relational databases or data lakes when source-to-ontology mappings are correct.
Entity alignment addresses records in different graphs that refer to the same real-world entity.
4 claims

Knowledge governance

Content and document-management systems can model roles, tasks, and validation criteria, while workflow systems can represent knowledge creation, use, and maintenance processes.
Knowledge audits can assess explicit and tacit knowledge, expertise, skills, and intellectual capital to identify strengths and gaps.
Metascience seeks to improve research quality and has five main research areas: methodology, reporting, reproducibility, evaluation, and incentives.
Integrity controls can establish integrity and publisher identity without resolving whether a publisher is credible.
4 claims

Knowledge protection

Formal protection includes intellectual-property instruments, access constraints, and protection of communication channels, systems, and storage.
Informal protection includes secrecy, norms, complexity, lead time, and human-resource practices.
Formal protection is described as more suitable for established, codified knowledge, while informal protection is described as more suitable for complex or difficult-to-codify knowledge.
Knowledge protection must balance sharing, learning, and collaboration against misuse, appropriation, or loss.
5 claims

Implementation constraints

Combining large ontologies or sources can create inconsistency, and ordinary deductive logic can make arbitrary conclusions derivable from contradiction.
Vast inputs, duplicated semantic terms, vague concepts and queries, and uncertain values challenge automated reasoning.
Formal inference can be brittle when required chain elements are missing.
Formal knowledge representation imposes cognitive overhead; tacit and changing knowledge add to the engineering problem, and domain- or organization-specific ways of expressing knowledge require community agreement.
Semantic metadata and geolocation can reduce anonymity, and machine-interpretable information can enable automated censorship.