MatVocab

Table Of Contents : A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z

Introduction

Several foundational elements required to achieve Sir Tim Berners-Lee’s vision for a semantic web are in place and available to the materials community. The semantic web, sometimes referred to as the web-of-data, focuses on ontologies as well as the linking data for machine-to-machine data interchange (implemented via RDF and OWL). Linkage between multiple datasets, files and their respective metadata can be established in an ad hoc fashion without having to adhere to specific database table structures. Linked data without context is of limited value. A semantic web for materials requires common vocabularies. An example of a common vocabulary is the Dublin Core (DC) ontology, a set of universally accepted metadata used to describe a resource (e.g. document).

The development and publishing of vocabulary using RDFS/OWL is one of the initial steps required to link relevant materials information across disparate (federated) sources. The development of common vocabularies could be jump-started via crowd sourcing and curated by materials subject matter experts (SME). Additionally, collaborative efforts with professional societies and other organizations (e.g. ASTM terminology standards, CEN, ASM, TMS, etc.) could be used to accelerate vocabulary/ontology development. Over time, multiple vocabularies would likely winnow down to key sets of generally accepted terms and mappings between terms having the same meaning. Taxonomies, a form of ontology, can express simple relationships in the materials domain.

More sophisticated relationships between materials processing, structure and properties can be expressed using complex ontologies. These ontologies need to be developed and implemented using World Wide Web Consortium (W3C) recommendations like RDF/OWL or widely accepted semantic technology standards such as time.owl and DC. As the above elements are being established on a larger scale, various forms of materials informatics could be developed to greatly expand the materials data and design space for the materials scientists and engineers.

Success requires innovative approaches during the development of agents to query linked materials data, applications to mash-up and integrate data, and reasoning/inferencing engines specifically tailored to the materials domain. Machine learning and other innovative “data hungry” approaches to extract knowledge could be developed and applied for materials design.