CS295J/Week 2.11

tuesday class notes

• introductions, now that shopping is over (make my own seating chart)
• admin
  • bring laptops -- will help us be responsive in class
  • interactive wiki review (projector? tv?)

• pre-proposal/proposal points
  • what is cool to you about the proposal
  • what would you want to do?
  • what might you want to do, but you don't understand
  • what research result would you like to accomplish? (different from

what you want to do -- research result is more constrained)

• • did you notice discontinuities? (different authors)
  • knowledge transfer? broader impact? intellectual merit?
  • cut and paste from solicitation

• how did you spend your time for class? (no expectation of total, yet -- still in first week!?

• what makes a paper good/relevant?

• what papers have you added and why?

• theory of vis description for defining research in choosing papers

thursday class notes

• give my practiced 2-minute Ernestine presentation from notes:
  • just like the proposal, this paper used a model of human behavior to estimate task time for phone operators
  • task was completing an operator-assisted call
  • tried 2 different user interfaces -- old and new
  • first tried GOMS (Goals, Operators, Methods, Selection) to describe human tasks, but it didn't account for our ability to do multiple things at once
    • eg talk to you while I move my hand to the board
  • then CPM-GOMS (adds Cognitive, Perceptual, Muscle) to get parallelism
    • each type of operation, C, P, M, can proceed in parallel
  • they got results that compared the two user interfaces accurately and matched real user studies that were also done for both interfaces
  • relates to proposal because it does same thing and illustrates many of the difficulties and complexities

• critique that presentation
  • what was good?
    • goals for presentation
      • related to our research
      • fit in 2 minutes
      • indicate further directions
  • what could have been improved

• should we discuss this further?
  • 5: yes, 4: maybe, 3: if we have nothing else to do, 2: no, 1: I'll leave if we do
  • go around table starting at my left

• practice
  • how many practiced their presentation? Hold up as many fingers as the number of times you practiced:
  • pair up, spread out,
    • 2:00 present to other person
    • 2:00 other person provide feedback afterward
    • 2:00 revise your notes
    • 2:00 repeat
    • 2:00 other person provide feedback afterward
      • 10:00 swap

• go through actual papers
  • 25:00 with discussion scoring after each one

• • Caroline: Cognitive strategies and eye movements for searching hierarchical computer displays
    • lots of "what"
    • validate intuitions about searching (proove hypotheses?)
    • "EPIC" architecture (like CPM_GOMS)
    • eye-tracking works
    • what model? what problems? how could we use it? What projects could we do with it? what would we need to do them?
  • van Wedeen paper
    • mapping whole-brain structure
    • project is about visualizing this information (picture would be great)
    • use as a semi-automated starting point for high-level regions?
    • lots of "what" and comparison to other similar methods
    • missing is visual analysis (makes a case for us to use it in our application)
    • network

• • Steve: theory of fluid visualization
    • is interesting, built up author as expert, position paper
    • two pieces: first describes how to use perceptual theories to predict human perforance
    • second: define tasks, construct tasks, iterate over mappings
    • limitation: no specific experiment (but does have experiments that come in the future)
    • dhl: nice acks (me, caroline's advisor :-)
    • dhl: a theory for part of the low-level perceptual human behavior

• • Michael: forearm electromyography
    • demonstrates that forearm modeling can be done as part of a UI
    • lots of "what"
    • freeing up hands could help users be more efficient
    • shortcoming: doesn't address the issue of cognition (a bit of a stretch in addressing)
    • dhl: slightly tangential, perhaps, but could use as human data capture (like eye tracking and fNIRS and event tracking...)

• • (no owner): intent to speak

• • diem: fNIRS, your brain on interfaces (brain heat imaging)
    • another human measurement, maybe of brain load

• • Hua: effectiveness of graph visualizations
    • concept of mental effort and evaluation of it
    • response time relationship to mental effort -- combines them
    • "mental effort" concept -- could we use it? goes to a higher level than muscle motions (other kinds of measures?)

• • Clara: visual attention review
    • experiment about how quickly users can adjust to confusing layouts
    • results are relevant as guide for designing HCI
    • eye-tracking treatment in the paper relevant to our use potential use of eye tracking
    • dhl: nice staging metaphor -- user experience: guide, but subtly
    • dhl: covert attention -- cool
    • dhl: all read this in more detail?

• • Jenna: color scale
    • tasks will be similar to what their experiments tested
    • speed-accuracy tradeoff in use of color
    • specific guidelines for using color

• • nathan: computational visual attention systems
    • overt and covert visual attention
    • what models there are for visual attention
    • vision and robotics applications
    • should be applicable to visualization, too -- knowing what folks are looking at should be good at predicting where we are showing too much or too little

• • Chen: wizard of oz for gestural interface for 3D selection of neural pathway estimates
    • good design process to emulate

• • Wenjun: maeda book
    • minimalist design
    • managing complexity is simplicity
    • dhl: could we figure out how this connects to cognition
    • illusion of waiting -- can be frustrated
    • put more emotion into design

• discussion of winner(s)
  • this will carry forward into next class, too

• moving forward
  • format ok? longer presentations? shorter?
  • another week like this one but more informed?
  • a week reading everything from one conference?
  • something else?