IW Meeting 2012-07-12

From Inference Web

Jump to: navigation, search

Contents

Meeting Information

Agenda

Attendees

  • Tim
  • Jim
  • Patrice
  • Deborah
  • James M
  • Cynthia
  • Patrick (regrets)
  • Paulo

Meeting Preparation

Around the room

 * Add a section for yourself 2 hours before meeting.  * Mark any discussion point that you would like to raise during meeting (with DURING MEETING).   * Otherwise, assume that others will read the rest before meeting.   * Also, please be considerate and read others' discussion points before the meeting starts.

Please put in at least 2 things that you would like out of a next generation explanation and provenance environment as fodder for our discussion on next generation Inference Web. Please do this no later than wednesday July 11 by noon eastern.  Then also please review everyone else's wish list

( BTW, we last discussed this topic at http://inference-web.org/wiki/IW_Meeting_2012-05-24 )

James

  • Finished FUSE storyboarding draft 
  • Revisions to FUSE online demo.
  • Developing first draft of summarization module for FUSE
    • Exploring generalizations to SPCDIS

Next generation explanation/provenance:

  • Summarization interface for tabular data collection analysis
    • Potential benefits in both SPCDIS + FUSE projects
  • Support module for enabling user tweaking of summarizations, based on given information retrieval tasks.

Tim

Requirement:

  • connectivity and discoverability in an environment where agents do not know of others' activities.

Approaches:

  • "provenance ping-back" to inform upstream agents about downstream activities. Revolves around POSTing PROV to the host of the data/services that was used.
  • Applying the FRIR stack. Uncoordinated, content-based identifiers that lead to consolidation across disparate provenance assertions.

Jim

  • Ontology as API for data cube browser, including provenance explanations of the vis that help define what is shown
  • Lit review for Linked Science

Will do:

  • Datacube provenance Lit review for Linked Science

Requirements:

  • Actionable provenance, especially in data cube contexts (is this provenance-as-plan?) Yes
    • reproducability
    • Tools that can act on provenance information to inform their use (this is broadly applicable, as in the following three existing uses)
      • RDFa specification for the vis to create
      • DataFAQs uses this approach for its configurations
      • VisKo is leaning towards this, too.

Cynthia

  • Auto-generated FUSE explanation.
  • IWBrowser extension of FUSE domain.

Patrice

  • 1 - Vocabulary Services available, e.g., BioPortal's ability to pull out pieces of the ontology through a rest service.   perhaps bioportal's ability to do this for provenance.  ..... this could move to a provenance portal.
    • catalog of provenance vocabs
    • avoid the proliferation of PROV extensions
    • it puts us into the center of the ongoing extensions
    • "Provenance Portal"
    • ontology level vs. instance level (Tim: but the instance level should just be a crawl the the provenance ping-backs)
    • We need to do this asap, since PROV just happened a couple of hours ago.
    • http://www.w3.org/2011/prov/wiki/ProvImplementations
    • Additional facet when showing any ontology class in Provenance Portal: show real-life examples, know what systems are exercising the class, etc. (fed by provenance-ping back)
    • TODO: Jim to install or estimate effort to install an instance of BioPortal locally.  Consider extension of prov with a couple of PML features
    • TODO: Tim to try it out on XCEDE-DM and PML {2 \union 3}
  • 2 - Documentation on comparison/contrast of PML with W3C Prov

Paulo

support http://globalchange.gov/  (Curt Tilmes 'n'crew)

  • Deborah: ++++++1
  • Tim: Curt wants "rhetoric ontology" for components of doc, linking claims to evidence (and attribution of each step) (Tim sent some links, no followup, though)
  • DOE's part of globalchange.gov is top dude.
  • 820 authors, 4k datasets, 100s organizations.
  • starting with high level provenance
  • only care about high level provenance,  people, affiliations, portions of documents, ...
  • focus on helping curt getting it right, and collect "what is missing?"
    • PML 3 might be driven by use cases that come out of the annual report provenance.
  • no word yet on the scalable provenance proposal from paulo, deborah, and yolanda
    • perhaps curt will join our meetings.  can contact gerhardt - BER - biology and environmental resources within office of science within doe  seed funding .
    • paulo will meet with him next week at esip.  deborah will be at esip on wed-thurs.
    • this annual report is to be released in dec 2012 and every year after that
    • TODO: Deborah and Paulo do stuff here.
    • question is the report reproducibport reproducible?    not today

Tim: I don't doubt that we can 1) do it and 2) make an impact and 3) get some great credit and visibility, but I've already spent my budget for this kind of effort.

High level

  • 0 - easy to use (not necessasily to understand)
  • 1 - to be inviting to other communities (PNNL people other than Paulo, CyberShARE, NASA, USC/ISI, etc.)
  • 2 - is scalable with support for abstractions (demonstrated example of use in support of so-called "big data")

More technical

  • 3 - supports incomplete/inconsistent views/points of view
  • 4 - provides support for data quality/uncertantity quantification
  • 5 - supports human activities/non-automated processes
  • 6 - still supports formal proofs
  • 7 - supports verification/validation

Very technical

  • 8 - integrated into PNNL's CAT
  • 9 - compatible with PROV-O

Patrick

Deborah

Next Generation Requests:

  • 1 - Easy to use by a broad range of people..... possibly meaning an onion design
  • 2 - Extensible
  • 3 - Leverages our strengths and research interests
  • 4 - Is fundable
  • 5 - Graphical representation of source usage.

Outstanding Items

Today's discussions

Publication Plans

Tim:  SeRSy 2012 - Semantic Technologies meet Recommender Systems & Big Data  (ISWC workshop) Due Jul 31

Jim: Global Health Explorer/Twitter+datacube (wt) Poster to ISWC (due 8/2), Datacube Provenance Lit review to LISC (Linked Science, @ ISWC) (due 7/31)

Deborah:   eScience with Ping et al on SemantAqua,  AGU future of information mgmt,  AGU more tbd

James: AGU (SPCDIS), ISWC Demo (possibly for FUSE)

Offline info

  • Patrice offline mostly july 16 - august 9
  • Deborah July 27 - August 5
  • Tim: ASAP, ALAP
  • Jim offline (no meetings, no work) 7/16 - 7/20
  • Cynthia 8/20 - 8/24

Administrative

Next week - mapping progress!  at least from tim and james

MEETING POINTS

Outstanding Items

Facts about IW Meeting 2012-07-12RDF feed
Date12 July 2012  +
Personal tools
Navigation