Jul, 2017

V 0.35

Engine

  • Detection of facts candidates in the form of triple

Preprint

  • Validate facts extracted from text. Delete triple
  • Validate facts extracted from text. Create triple
  • Show all facts of the semanticized paper

Jun, 2017

V 0.33

Engine

  • Word sense disambiguation (WSD) of NER candidates

Preprint

  • Further extension of JATS biomedical metadata

Application for journals

  • Generating HTML with biomedical microdata

May, 2017

V 0.321

Engine

  • Authors recognition and labeling

Preprint

  • Authors authentication via ORCID
  • Improved backend architecture
  • Improved preprint load time with preliminary semantic objects preprocessing

Apr, 2017

V 0.32

Generic

  • Fixes and improvements
  • Updated Statistics in the Ontologies Research

Mar, 2017

V 0.31

Engine

  • Authors identification with ORCID IDs
  • Uniprot’s TrEMBL NER
  • Extended terms variants generation to increase papers recall in the Information Retrieval task

Preprint

  • Marking text chunk as possible biomedical concept
      Do you see protein, chemical element, drug or laboratory equipment in the text but there is no relevant object in the external ontologies? Just highlight it and select the checkbox “Mark as term”. sci.AI engine will periodically look for the match in the knowledge graph. This way you’ll influence growth of the ontologies
  • Adding custom ontology and ID
      Do you see element that is not found in currently supported ChEBI but you know it is present PubChem? Do you know any custom but publically available Enterprise ontology with relevant element? Add ontology name and ID of the object to the custom fields, we’ll validate and connect it to the knowledge graph. Later it might influence adding the ontology to the officially supported list
  • Applying the same labeling to all similar chunks
      Validated one “TNFa” in the text and there are 100 more? Apply changes in bulk
  • Exporting semanticized paper in JATS format
      Journals can embed it directly into the editorial process. There is no standard schema for the biomedical objects in JATS now. Standardization is the future works

Feb, 2017

V 0.3

Engine

  • Uniprot’s Swiss-Prot, ChEBI, ICD-10, Gene Ontology, MeSH, Drugbank named entities recognition (NER)
  • Labeling of the recognized proteins, chemical elements, drugs and MeSH entities in the text

Preprint

  • sci.AI Preprint to read and validate semanticized papers
  • Google ID single sign on to the sci.AI preprint
      Login with your Google ID to the http://app.sci.ai/
  • Exporting semanticized paper in custom XML format

Google Add-on

  • Semanticize terms via Google Docs add-on
  • Google Docs and sci.AI processor integration