IW Meeting 2011-12-08

From Inference Web

Jump to: navigation, search

Contents

Meeting Information

Agenda

  • IEEE IS paper
  • FUSE Updates
  • W. Chen's CHSI project - is weijing on?  Jim M can give the context for this.  demo on an 3 and jan 5

Attendees

  • Tim is here will be spotty
  • Deborah
  • Jim
  • James (green)

Deborah is only online a bit. We're going to just type here, since there are more people here.

Discussions

IEEE IS paper

Tim can look at this Friday and Saturday. Jim will be working on this Friday and evenings.

I'm really excited that this paper was as well received as it was. The comments were extremely constructive and I think will help us make this paper really great. I've made a google doc for the reviewer comments:

https://docs.google.com/document/d/1ax-lV43LePYM8WH9ypW6MoLgnR17wvvq3ejGDHlTRNM/edit?hl=en_US

I need each of the authors to do the following:

Would be good to have these instructions at the top of https://docs.google.com/document/d/1ax-lV43LePYM8WH9ypW6MoLgnR17wvvq3ejGDHlTRNM/edit?hl=en_US Jim: +1

  • Look over the comments and find fixes that you're willing to do. There are a bunch of text changes that we can just go and do pretty quickly.
  • Mark those fixes with a comment: "mine" if you're taking it on, "complete" if you just went and fixed it.
  • Use those comments as an ongoing status of the change task.
  • Add "complete" as a reply when the task is done.
  • Add a response to reviewer after that. I'll aggregate these into the response to reviewers that has to go with the paper.

Since these are mostly small changes, I think we should stay with the LaTeX version of the paper for this revision, unless anyone has objections. The URL for the paper itself is here:

https://docs.google.com/document/d/1VnK8I1EJIjLBYoGsDF9sH_HDf8RVhXrCddN2lxit3yw/edit

The due date is next friday, so we have to move on this.

Tim saw a comment about inconsistent URIs, which can review the concrete examples and diagrams for consistency.


W. Chen CHSI project

The goal of Community Health Status Indicator(CHSI) is to provide an overview of key health indicator for local communities. To present indicators like deaths due to heart disease and cancer.

W. Chen needs to let us know what she(?) needs.

Tim recreated a CHSI dataset and dug up Dominic's CHSI dataset.

Jim McCusker is working with Weijng and should be able to give an overview of the work and needs

Tim created http://logd.tw.rpi.edu/weijing_chen_chsi_coordination_page b/c there isn't any other materials describing the project. The Drupal formatting is terrible.


Jim - will you put up information describing the project on the page tim generated?

Sure. -- Jim.

I can work to enhance it, but need to know what to focus on.

Jim is the lead 

Jim is to give me  report today.  we will be getting weekly reports from weijing and jim on thursday mornings.  (i havenot seen the one from this morning. If Tim is to help out, he needs to be cc'd. An ongoing coordiantion page would be best for him to see what needs to be done.

jim should take the lead on these - both the reports that are behind and need to be ramped up and the cordinaion page. there are demos the first week of jan 

Jim's most recent update

I'm working on extending my RDF visualizer so that it can easily be used to display explanations. I'm going to ask Weijing to put together some provenance for her dataset and vis so I can test that the visuals are appropriate and I can start rolling up the details.

You can see one example bit of provenance here:

http://sparql.tw.rpi.edu/services/rdfviewer/?url=http%3A%2F%2Fdl.dropbox.com%2Fu%2F9752413%2Fresearchdemo%2Fpatents%2Fderivation.ttl

That's from the slides I had sent out before, it's much more powerful now, and it is (in theory) embeddable using a couple of lines of javascript in any HTML page.

One idea is that we can use DBpedia entries for various statistical algorithms and visualizations that Weijing will be using. I'm going to be adding a follow-your-nose capability to allow the user to expand the graph based on what things are already being linked to.


Did you meet with Weijing yesterday? I thught the report above was from what was done previously and that we would start getting thursday morning reports and also thought you had a meeting planned for yesterday. - Deborah

I was traveling Sunday-last night, so no. - Jim

Next steps

when is the xt meetingplanned? - Deborah

No date set yet. I'll set one up with her for tomorrow, through a call or email. - Jim


This project needs to pick up the pace.


Maybe I can let her focus on getting the data converted, while I work on some sort of analysis and visualization.

ok - just make sure she is makingprogress.

I'll take a look at what she has so far, and try to take it from there, then.

Jim, will you be able to point me to what I should be doing? I'm not looking to join another project, but am happy to help if it is easier for me to do.

Yes. I'll take a look at Weijing's datasets and see what's been done and what's left to do.


tim if you can helmake sure weijing is getting any data in - that would be great.  (and of course any follow up identified by jim and/or weijing. I can walk her through conversion stuff. I can use it as an excuse to exercise the documentation and tools.

That would be great. I may kick in some work on that front to make sure the modeling supports the actual user-facing work.

I don't plan to do much acontextual enhancement - there is so much in there that it would take forever, while virtually none of it would be used.

Well, I mean that I'll fix up what Weijing has done to make sure that it works well with the applications we have in mind.


great - can we set a time next week to either demo anything or at least go over current plans?

Would the next IW call be appropriate?

Next Actions:

  • JPM: Acquire and assess the current data conversion for use in a demo. (Early 12/9)
  • TL: Offer help to WC on enhancement to help her with current issues or blockers. (TL sent her an email this afternoon) (done)
  • JPM: Build initial visualization to explore available data. (over weekend/before wednesday)
  • DM: encourage WC to be proactive
  • WC: provide an update to JPM on status (Wednesday)
  • JPM: provide an update via email to DM on status. (Thursday Morning)
  • JPM: provide a demo of existing work, review of existing plans (Thursday afternoon)
  • JPM: set up meeting with WC and DM for update on Wednesday at 4:45. (email sent)




FUSE Site Visit

Sure.  If it would be  helpful, James have access to both slides and notes taken during the visit.   Would need to upload these, though.  Below are some highlights, which I  think are relevant.  Deborah, let me know if you want something else  added.

  • 2-day meeting at BAE Systems offices
  • Talks by RPI, as well as collaborating groups.  Each talk covered following:
    • A review of current research directions
    • Waypoints for FUSE work, for December and January
  • Feedback was provided by government visitors (including FUSE PM Dewey Murdick) on each talk
  • In RPI talk, we reviewed storyboarding work, current research focuses, and gave following waypoints:
  • December
    • Developed initial designs for representing provenance information
    • Completed storyboarding for selected challenge questions
    •   Generated some initial PML-based proofs for representing FUSE explanations
  • June
    • Development of a prototype for presenting provenance info from FUSE
    • Once we get data from CADRE, we will then be able to start working on the real stuff.
  • Additionally, some discussion on how related RPI research may be applied to FUSE
    • There was agreement that provenance/explanation abstraction would be useful
      • Both James Michaelis / Jim McCusker's work may tie into this


      •   yes - jim m - we do notneed you this year on this but next year (jan or early feb) we might have a conversation about how we can leverage jim /tim and james' abstracton work (abstraction work is a bit longer term in Tim's plan, but I'm happy to discuss this)

great

  • On this week's FUSE telecon, Olga mentioned that she will need revised waypoints from each team by Monday (12/12) - so she can review them, and get get them to Dewey.
    • I am currently looking into this, but will probably need to do a review pass with Deborah, Joanne and Jim Hendler before sending to Olga

james m - if you can draft  something i can iterate on sunday.   (also i think yu have the last way point slide that i presented.  that should be a starting point.  let me know if you need that.  (i can not send now - internet too bad)

  • Alright, I'll plan on doing that.  Since the waypoints do involve infrastructure work as well, I'll try to get Jim Hendler to do a review pass.  perfect

Deborah: Additionally, I'm starting to talk w/ Dan hunter about CADRE - to get a better idea of how hypothesis triggering/management works.

From this, I'm hoping to get a better idea of how CADRE manages positive/negative evidence for hypotheses. great- cynthia should be on cc lists on this line of work as well)  Alright, I'll plan on that

i amgoing to have to shut down in about 3 minutes and walk to another location. send email if you need me for something after i appear to drop off.  i can answer later.

DM: is a demo of an initial viz/explorer at next weeks IW call good timing? yes

i think you and i and weijing might also do alonger call / meeting but deftely this demon n thurs iw meeting is great? Sure, we can do that Wednesday during her office hoursi am goingto need to leave winlow by 5:45 or 6.  if we can start say 4:45 or latest 5, that is ok? Fine with me, I'll check with WCperfect thx.

Facts about IW Meeting 2011-12-08RDF feed
Date8 December 2011  +
Personal tools
Navigation