21
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 644753 (KConnect). Webinar: KConnect Cloud Market: Medical Text Analytics and Semantic Search

KConnect Cloud Market: Medical Text Analytics and Semantic Search as a Services

Embed Size (px)

Citation preview

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 644753 (KConnect).

Webinar: KConnect Cloud Market: Medical Text Analytics and Semantic Search

Outline

KConnect Project Overview KConnect Cloud Market Cloud Market Services Usage Scenarios Resources Questions

Kconnect Project Overview Website: http://www.kconnect.eu/ Focus on medical and clinical domain Multilingual text analytics and semantic search Services cover:

Semantic annotations Semantic search and log analysis Document classification Machine translation

Services are delivered: Locally on premises On the Cloud

Scope of WP2

Cloud Architecture overview

applications

applications

Web UI

routing nodes

Mimir nodes storagenotifications

Docker repository

account / quota management

metadata store

text analytics

document queue

Knowledge Base

Machine translation

Knowledge base exploration Text analytics services

Semantic BioMedical Tagger Healthcare Tagger Bio-Yodie Hungarian KConnect pipeline Swedish KConnect pipeline

Fully-managed Mimir-as-a-Service Machine translation

Cloud Market Services

KConnect Cloud Market

Using the Cloud Platform

KConnect Cloud – Register

Free and quick account registration

Get API key Annotate 250MB of

documents per month Execute 5000 SPARQL

queries

KConnect Cloud – Select service

Select medical service to use

Test with a sample text to see the results online

KConnect Cloud – API Keys

Issue an API key to access the services

API keys provide flexible authentication management

Keys can be enabled, disabled and removed

Unified Medical Language System (UMLS) Vocabulary More than 130 vocabularies and ontologies Multiple language support

Other language data sets (Swedish, Hungarian, Chech, etc)

Data is modelled in RDF Data can by queried using SPARQL Access via REST API at:

https://kconnect-kb.s4.ontotext.com/v1/sparql

KnowledgeBase Service

Example SPARQL queries Get a list of all data sets Get a list of all semantic types of UMLS Retrieve a list of concepts from a specific semantic

type Retrieve a list of concepts from a single UMLS data

sub set Get a hierarchy of terms

Demo of both KB UI and using POSTMAN REST API plug-in

KnowledgeBase Service

Text analytics Populate gazetteers Multi-language support Concept meta-data provider for the annotations

Semantic search Auto-suggest Hierarchical search Relations

KnowledgeBase Service – Usage

Available Services Semantic Biomedical Tagger Healthcare Tagger Bio-Yodie

Body_Location or Region; Body Part Organ or Organ Component; Body Space or Junction; Body System; Tissue; Anatomical Abnormality; Acquired Abnormality; Congenital Abnormality; Finding; Injury or Poisoning; Pathologic Function; Disease or Syndrome; Mental or Behavioral Dysfunction; Cell or Molecular Dysfunction; Experimental Model of Disease; Sign or Symptom; Neoplastic Process; Diagnostic Procedure; Laboratory Procedure; Laboratory or Test Result; Research Activity; Molecular Biology Research Technique; Clinical Drug; Pharmacologic Substance; Antibiotic

NegEx Algorithm – Negation, Temporality, Experiencer

Swedish KConnect Pipeline Hungarian KConnect Pipeline Machine Translation

Demo with KBUI and PostMan REST API plugin

Text Analysis Services

KConnect Cloud – Access the APIs

API_KEY = <api key>KEY_SECRET = <secret>SERVICE_ENDPOINT = https://$API_KEY:[email protected]/v1/hu-application CONTENT_TYPE = text/plainJSON_REQUEST = "{\"document\" : \"A ribonukleinsav (RNS) a DNS -hez hasonló polimer óriásmolekula, amely sok ismétlődő egységből épül fel. Egységei a ribonukleotidok .\", \"documentType\" : \“$CONTENT_TYPE\"}" curl -X POST -H "Content-Type: application/json" -d “$JSON_REQUEST” “$SERVICE_ENDPOINT" -o response.json

{ "text":"A ribonukleinsav (RNS) a DNS -hez …. all.” , “entities” :{ "Disease":[{"indices":[327,330],"type": "Disease", "source":"umls", "cui":"C0023449", "tui":"T191", "tui-name":"Neoplastic Process"}]}}

Command line input:

Output:

KConnect Cloud – Access the APIsRequest:

Response:

Instantly deployment of new Mimir servers when needed

Fully-managed setup Zero administration: automated operations,

maintenance & upgrades High availability & resilience Cost efficient Isolation of the multi-tenant Mimir services Fair use of shared resources

Mimir-as-a-Service

Mimir-as-a-Service Management

Demo Application

document processing

semantic Search

Mimir storage

text analytics

document queue

Knowledge Base

annotated documents

metadataexploration

RDF-ized annotations

Populate gazetteers

Resources

Java, Python & C# SDKs http://docs.s4.ontotext.com/display/S4docs/Add-ons

Sample code Java, C#, NodeJS, JavaScript, Python, PHP, Groovy cUrl examples http://docs.s4.ontotext.com/display/S4docs/Sample+Code

GATE plug-in Online documentation

http://docs.s4.ontotext.com

KConnect Cloud Market

Thank You For Your Attention!

Questions?