Week 12 April 26, 2011: Project/Paper Presentations (Paper and Presentation (Slides) or Project and Presentation (Slides))

Printer-friendly version

Titanpad Notes

This week we offer the opportunity for initial project presentations. In the past, it has worked very well when there is a 'second chance' at presenting final projects. Having an opportunity to go back and modify the presentation or project based on feedback is an integral part of scientific work, however, in many situations, there isn't a second chance. One of the goals of this approach is for you to learn and experience the benefits of feedback, thus encouraging you through your experience, to create for yourselves, the practice sessions ahead of the final presentations. For example, if preparing for a major presentation event, make believe that the deadline is one month before the actual deadline. This will get you in the habit of planning ahead and affording you to incorporate feedback as well as set the paper or presentation aside, coming back to it later with "fresh eyes." Another advantage is that it would also allow you to get feedback from colleagues who have the same final deadline but not at the deadline time, when they would not be available.

Reading Assignment for this class:

  • Review/re-read The Craft of Research, Chapter 8, Claims
  • Read classmates claims (see below).

Written Assignment for this class:

  • By April 20th, articulate in writing, three claims about how in your project, you are using and the benefits you are obtaining, from semantic technologies. Include an example for each claim that you will present in class. Post these claims to the website for your classmates to review.
    1. Jin - Water Quality Portal
      • Claim: Water Quality Portal provides automatic reasoning and inference over water sources and facilities to determine weather or not a water source is polluted, or a facility is violating regulation. This automatic reasoning and inference is enabled by the Semantic Web Technology: OWL ontology.
        Example: In EPA regulation, it specifies any water source cannot contain arsenic 0.01 mg/l or more. Therefore, we use the ontology we designed to encode these information, and use encoded information combine with the ontology to provide automatic reasoning using Jena and Pellet reasoner.
      • Claim: Water Quality Portal mashes data from different data sources, and perform reasoning over these data as well as visualize the data.
        Example: The data we are collecting from include sources like USGS website, EPA website, and different state regulations websites such as California State Regulation.
      • Claim: Water Quality Portal provides Provenance information about the water sources, and utilize these provenance information to help user query the data base on the data source.
        Example: The user may be interested at applying California Regulation constraints to USGS water data. In this case, the user can specify the request through Water Quality Portal interface, the back-end will perform queries as well as reasoning based on the request.
    2. Group - Peter, Ping, Scott - Health
      • Claim: In this system, we utilize semantic technology to better model the domain of health care. We develop the ontology for the project in OWL using Protege.
        Example: Our ontology defines classes for domain concepts like medical practitioners, patients, problems, and also the relationship between the domain concepts like which patients go to which medical practitioners, and which drugs can be applicable to which problems.
      • Claim: This system supports specialty-based EHR Visualization via semantic technology.
        Example: The user who is a medical practitioner by default, inputs his/her id to log in the system. The system can infer the drugs the user is interested in based on the user's specialty by issuing a sparql query to the triple store.
      • Claim: This system can be easily extended, because the semantic technologies we adopted have good extensibility.
        Example: To extend the system to include more medical practitioner specialties, problems, drugs, we only need to add classes into the ontology and instances into the triple store.
      • Eric - S2S Methods Paper
        • Claim: S2S leverages semantics to provide a metadata vocabulary, and, in future versions, will use linked data practices for dereferenceable framework metadata.
          Example: The S2S ontology provides terms for describing a user interface widget, including properties referencing required scripts and function calls. In the next version of the framework, unregistered URIs will be dereferenced to retrieve metadata.
        • Claim: S2S provides explicit semantics for the relationships between web services, search parameters, and user interface widgets.
          Example: The S2S ontology has terms to associate services with the query interfaces they implement, query interfaces with the search parameters they allow and descriptions of the output formats they produce, and widgets with the output formats and search parameters they support.
        • Claim: S2S leverages RDFS reasoning to infer sub-class transitivity for type checking, and, in future versions, will use OWL-DL to perform instance integrity checking using cardinality and value qualifiers.
          Example: There are two types of query interfaces being used in the current version of S2S, result queries and parameter value queries. Before a query interface instance can be used for a specific purpose, S2S checks the types that an instance belongs to. For the integrity checking claim, in the next version of the framework, we will leverage a DL reasoner to assist framework component creators in the development of RDF metadata for components.
  • Review other team's claims. Post at least one question on each other's project claims by April 24.
  • Review questions from classmates on your claims, and on April 26 present your project overview your sem web claims

Presentation Assignment for this class:

  • Present your plans for your project or paper by April 26
  • Presentations need to include the following slides:
    • Overview of slide of project or paper
    • Project architecture slide of project or outline of paper slide
    • Overview of claims of benefits of semantic technologies (3 or more if desired)
    • One example of the benefit for each claim
    • Current project status including what is remaining to be done
    • For the group project, a slide identifying the contributions of each person
Name Questions for Groups Slides
Eric Water Quality Portal: You are working with GBs of water quality data (and presumably O[10^9] triples). How are you performing reasoning at this scale? Is it done dynamically using SPARQL (i.e., applying custom rule sets over the data at query time), or is it done at data ingest time (i.e., querying over precomputed inferences)?
Health Group: Does the ontology you created for the health domain align well with other medical domain ontologies? Do you foresee any limitations in the extensibility/mappability of the ontology to other models for medical specialities, drugs, etc.?
Eric's Slides [Download]
Jin S2S: What vocabulary you provide? what do you mean by dereferenceable framework metadata? In the third claim, how does S2S support RDFS reasoning, why should we care these inference and reasoning?
Health Group:What are the main contributions of your ontology? What problems does the ontology solve that other ontology won't be able to solve? Does the ontology support any OWL type reasoning, or the ontology is more like to define medical vocabulary?
Peter Water Quality Portal:Very interesting work. No questions at this time.
S2S: Based on the claims and example, I have no idea what you are doing. Can you use non-semantic web specific terminology to explain what exactly you are doing? (I have a good idea of it because you've explained it to me before, but these claims and examples provide no substantial meaning for your project.)
Scott Water Quality Portal:The purpose and use is clear. It works quite nicely, and is successful. I have trouble understanding how a series of upper and lower limits in an ontology provides meaningful semantics; or semantics above those of a text file.

S2S:Why are these extra levels of semantics necessary? It seems to me that it's getting out of hand.
Health Group Slides
Ping S2S:What are the difference between S2S and Swoogle? Did you developed the metadata vocabulary of S2S?