HCIL Symposium: Day 2

Workshop: Humans and the Semantic Web

Session 1: Ontology Visualization Tools

  1. Olivier Bodenreider, National Library of Medicine: Visualization Tools for the Unified Medical Language System (SemNav), the Gene Ontology (GenNav)and RxNorm

    • UMLS – used my NLM (using MESH data), to visualize semantic relationships
  2. Catherine Plaisant, HCIL, University of Maryland: Graph visualization
    • SpaceTree – how to see branches of a hierarchical classification tree?
    • Click node to open new branch is an easy-to-read format
    • (Can this be used for site navigation or sitemap?)
    • Unlike MS Explorer, this system will open up and show more branches (as long as it can easily be seen within the existing window)
    • TreePlus – takes clustering (example of animal food chain) and makes it hierarchical
    • Moves branches to show information without becoming cluttered
  3. David Wang, MIND Lab, University of Maryland: CropCircles: Topology Sensitive Visualization of OWL Class Hirarchies
    • Ontology visualization is used to view relationships and hierarchies in a usable fashion
    • Class hierarchies – gives users a sense of context when browsing and connects related concepts
    • The problem with hierarchical structures is that it can be very hard t read
    • What happens when there are multiple inheritances?
    • Compared tool to Treemap and Spacetree
      • Spacetree does very well in navigational tasks
      • Spacetree and CropCircles are fairly good at returning to visited node

Session 2: Web and Annotation Tools

  1. Tim Finin, eBiquity Lab, University of Maryland Baltimore County: Swoogle: A Semantic Web search engine

    • Swoogle is the effort to make sense of the Semantic Web
    • swoogle.umbc.edu
    • Support Semantic Web developers
    • Searching specialized collections
    • Support Semantic Web tools
    • SPARQL is a language like SQL for the Web
    • Used for the SPIRE project
  2. Daniel Krech, MIND Lab, University of Maryland: Redfoot: Content Management for the Semantic Web
    • Started as an RDF store/viewer/editor framework
    • CMS built on Semantic Web technologies
    • Sharing content between CMSs is often laborious
    • Solves problem of sharing and reusing content
    • CMS face helps shield humans from the SW bits intended for machines
    • Redfoot functionality
      • generic editor
      • context management
      • authentication support
      • comments, bookmarks
      • edit content
      • blog
      • recipes, ingredients, DOAP
    • Redfoot built to explore how the SW can be used to analyze terrorist activity
  3. Michael Grove, MIND Lab, University of Maryland: Image Annotation on the Semantic Web Discussion
    • PhotoStuff
    • Multimedia ontologies
    • It’s a pain to mark up photos for the web – takes time
    • PhotoStuff takes automated info captured by digital camera
    • Rich opportunity to add a wealth of metadata
    • Use of Natural Language Processing to show additional info in image popups
    • Adds descriptions automatically

Session 3: Semantic Web in Practice

  1. Gary Berg-Cross, Building Ontology Patterns for the Semantic Web: Practical Issues

    • Ontologies start with taxonomic pattern
    • Taxonomies can be tangled (cross-applicability)
    • Subsumption
      • Rigidity
      • Identity
      • Unity
      • Dependency
    • Struggle to make an “untangled” taxonomy
    • Ontologies may use meta-properties to distinguish between objects and process
    • Focus on space-region, object, event, time-interval
  2. David Zaccagnini, Knowledge Discovery using Semantic Web technologies and text mining in life sciences
    • Text mining and the semantic web
    • Automate the assignment of semantics to unstructured content
    • PubMed example – humans cannot make sense of a resulting hitlist of thousands of hits
    • Their system combines proprietary and public ontologies
    • Tool identifies concepts and adds these concepts into the system
  3. Brian Caruso and Brian Lowe, Mann Library, Cornell University: VIVO: An ontology-driven life sciences research portal
    • Attempts to consolidate information about life sciences at Cornell University
    • Results are organized as entities
    • vivo.cornell.edu
  4. Duane Degler, IPGems: Current research and trends in interaction design for the Semantic Web
    • “New acquaintances or old friends?” – metaphor to HCI
    • “The doctor’s appointment” – Let user enter into his/her handheld web browser a natural language request that the computer can interpret based on information already known about the user (calendar, contacts, location, time, distance, etc.)
    • The concept of the agent that conducts business on your behalf
    • Concept of trust with the agent – will the agent get it right?

Session 4: Panel Discussion

  • Alfredo Morales, Cerebra
  • Duane Degler, IPGems
  • Walt Truszkowski, Goddard Space Flight Center
  • Harry Chen, Image Matters

Walt Truszkowski discussed how his office is using personal web space and applying SM technologies. For example, a user can go to the site (private, employee only) and choose an agent. As this agent a question and the agent will return possible answers.

Harry Chen discussed geospatial information on the SM. GIS has become mainstream with the advent of Google Maps, etc. Any person can access this type of information.

Q: What’s so special about the Semantic Web?

The group discussed how the Semantic Web is growing, but there are growing pains. No conclusions drawn

Semantic Web User Interaction Workshop, November 6, 2006

Tags: [, , , , , , , , , , ]