Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Taxonomies, Ontologies and Knowledge Graphs
Faceted Taxonomies: Start Search From Any Perspec>ve, Progressively Filter Results
Taxonomy • List or Hierarchy • BT/NT/RT rela>onships • Matrix and Facet forms • Visual naviga>on • Human friendly • Built for known needs
Thesaurus • BT/NT/RT/ USE/UF rela>onships
• Dic>onary naviga>on
Ontology • Maps unlimited types of rela>onship between
concepts • Difficult to build and maintain -‐ especially for
fluid, imprecise environments • Machine friendly not human friendly • Built for mul>ple poten>al needs
Taxonomy
• List or Hierarchy • BT/NT/RT rela>onships • Matrix and Facet forms • Visual naviga>on • Human friendly • Built for known needs
Breathe…. Now we’re going to talk about Knowledge Graphs, Taskonomies and
leading into Search Based Apps Ontology
Taxonomy Facets
Salient Rela7onships
Contextual Analysis
Content Sources
• Between facets • Between concepts • Between facets and
content • Between content
and content • Mapping is ac>ve
and passive
• People • Places • Ins>tu>ons
• Wikipedia • Weather server • Time server • Map server • Reviews • Web pages • Structured
databases
• Taskonomy – user needs • Disambigua>on cards • Frequent next ac>on
Relevance vs Salience
Relevance Retrieved results meet the need expressed in the query. Retrospec>ve, against defined needs. Supports finding stuff you are looking for.
Salience Having the quality of standing out as especially significant. Has con>nuing u>lity, meets as-‐yet unexpressed needs. Supports discovery of stuff you didn’t know existed.
Taxonomy vs Ontology vs Knowledge Graph
Taxonomy • BT/NT/RT rela>onships • Hierarchical • Visual naviga>on • Human friendly • Built for known need Thesaurus • BT/NT/RT/ USE/UF
rela>onships • Dic>onary naviga>on
Ontology • Maps unlimited types of
rela>onship between concepts
• Difficult to build and maintain -‐ especially for fluid, imprecise environments
• Machine friendly not human friendly
• Built for mul>ple poten>al needs
Knowledge Graph • Maps salient en>>es and
rela>onships • Can map to content as
well as concepts • Highly scalable, flexible,
evidence-‐based, built for known need
Taskonomy: an arrangement to enable common work tasks
“I’m prescribing medica>on for a pa>ent. I want to immediately iden>fy medica>on appropriate to a diagnosis, that we have in stock. I want to check contraindica>ons against my pa>ent’s medical history. I want to be aware of any incidents involving the selected medica>on, and I want to be able to advise the pa>ent on how to manage their daily life while on the medica>on.”
Search Based Applica>ons are fed by Taskonomies (+ Taxonomies + Knowledge Graphs)
Task
Pharmaceu>cals database
(subscrip>on)
ePharmacy (hospital)
Pa>entslikeus.com (social)
Pa>ent records system (hospital)
Incident repor>ng system (hospital)
“I’m prescribing medica>on for a pa>ent. I want to immediately iden>fy medica>on appropriate to a diagnosis, that we have in stock. I want to check contraindica>ons against my pa>ent’s medical history. I want to be aware of any incidents involving the selected medica>on, and I want to be able to advise the pa>ent on how to manage their daily life while on the medica>on.”
Pharmaceu>cals database
(subscrip>on)
ePharmacy (hospital)
Pa>entslikeus.com (social)
Pa>ent records system (hospital)
Incident repor>ng system (hospital)
“I’m prescribing medica>on for a pa>ent. I want to immediately iden>fy medica>on appropriate to a diagnosis, that we have in stock. I want to check contraindica>ons against my pa>ent’s medical history. I want to be aware of any incidents involving the selected medica>on, and I want to be able to advise the pa>ent on how to manage their daily life while on the medica>on.”
Taxonomies & Thesauri • Resolve vocabularies to common terms
• Support naviga>on
Knowledge graphs • Map rela>onships
between concepts & concepts and between concepts and content
• That are salient to the task at hand