Lab Meeting, Spring 2015, Week 11

Awe- and Aw-inspiring readings (or just stuff you think deserves a mention)

What we did over last week

Florian

1. Finished and presented my CUNY plenary.
2. Met with a bunch of people at CUNY to talk about their projects and collaborations. This included a chat with Vera Demberg about Zach's idea to use ICA to measure syntactic adaptation in spoken language comprehension.
3. Wrote large chunks of a paper on dependency length and information density in natural languages compared to chance (joint work with Dan Gildea) and outline further studies we need to run before we can submit.
4. Met with Dean Williams, Chigusa, and Dick about upcoming UofR fund raising meeting in Florida.
5. Met with Brad and Dick about planning of new NICHD training grant proposal.
6. Met with Emily Morgan to provide feedback on her NRSA post-doc proposal (for a joint project with Davy Temperley and me on music and language production) and discussed revisions and time line.
7. Edited press release for Weatherholtz et al. paper.
8. Finished another (final) edit of the paper with Gina Kuperberg (on the role of prediction) and the cover letter to the editor. We're submitting today or tomorrow.

Andrew

1. Moved subject running system to MariaDB from SQLite in hopes of a performance gain.
2. Put Tal's new experiment up.
3. Started looking into hardware for a replacement webserver, because it looks like that's the real performance problem.

Olga

Sarah

1. Edited NRSA research strategy and sent to Marta Kutas.
2. Edited draft of NRSA specific aims.
3. Edited draft of NRSA activities plan.
4. Guest lectured in Ling 227 on Monday (the 16th) and Wednesday (the 18th).

5. Attempted to set up the latest edition to the Gotta eye-tracking experiments. Ran into issues with PsychToolBox; further debugging is needed.

Dave

1. Finalized-finalized CUNY poster pdf + presented. Interest was light but people seemed to be as surprised as we were that labels didn't matter. Got some interesting suggestions about looking at RTs (did a quick look and nothing really jumped out RE labeled vs. unlabeled or supervised vs. unsupervised, but didn't directly compare with category boundary analysis).

2. Made an R package with all the data an analysis scripts for the supervised/unsupervised learning study, which is on github: https://github.com/kleinschmidt/phonetic-sup-unsup (you can install in R with devtools::install_github('kleinschmidt/phonetic-sup-unsup')).

3. Made the methods report for the CUNY poster analyses all purty, and also posted to github: https://github.com/kleinschmidt/cuny_2015. To recreate the report, run kintr::rocco('poster.Rmd'). This will install the analysis package and run the script, including fitting a big lmer model, FYI

4. Finished first pass at clearner fMRI classifier interface, but got stuck on how to do the group-level analyses because SPM is a dumbass
5. Incorporated papa's feedback on diss outline and revised it.

6. Met w/ Florian about how to infer prior beliefs based on adaptation data (and how/why modeling hasn't worked out like we'd hoped). Comes down to difficulty with jointly inferring category parameters based on exposure statistics and subjects' responses.

7. Worked on increasing the robustness and test coverage of the formula (for, e.g., regression models) interface in Julia.

Esteban

1. Finished up CUNY poster
2. Did some follow-up stats on the data from that poster in reference to time course effects
3. Presented said poster
4. Met with Scott Seyfarth about finalizing stimuli for a new study.
5. Presented my poster at CUNY
6. Went through my pre-doc NRSA resubmission summary statement and contacted my program officer
7. Met with Susanne Gahl, Jennifer Arnold, Craig Chambers, and others at CUNY

Amanda

1. Ran two of the remaining studies for the mammals project. Met with Mike and Chigusa to talk about the findings of these studies.
2. Met with Zach and tried to put together all of the stuff we need to meet with our advisors about the Zamanda project.
3. Recorded some sentences for the Zamanda project.
4. Read through Cory and Dan's grant proposals for the mock review we had this week. Made plans with Greg to discuss my grant this upcoming week.
5. Made lecture slides for my guest lecture in Scott Grimm's Intro to Pragmatics class for this Tuesday.

6. Read Fedzechkina, Jaeger, & Newport (2012) in prep to make my guest lecture for Lang Dev this Thursday. Started working on the slides.

7. Responded to a ton of student emails, set up meetings with students, etc to help them with their Lang Dev CHILDES homework.
8. Set up a survey and collected responses for the group projects for Lang Dev.
9. Started setting up the (hopefully) final (production) study for the mammals project. Thought seriously about reading it (in Chigusa's words, I opened the paper, but haven't started looking at it yet...).
10. Looked up a paper by Emiel Krahmer based on people's recommendations from CUNY (he's also the editor on my Frontiers paper).
11. Found out how much Frontiers papers cost. In case you're wondering, it's a lot ($1900 or ~$1500 for a special topics paper). Also, in case you were wondering the max. length for a paper is supposed to be 12,000 words (the real reason I was looking up things on their website).
12. Started to think about Quals.
13. Discovered that there is an NSF DIG in linguistics again. Proposed to Mike and Chigusa that I should aim for the January 15th, 2016 deadline (post-quals, with enough time to resubmit, if necessary for the July 15th deadline before the start of my 4th year).

Maryam

Zach

1. Recorded stimuli for the Zamanda experiment, this time with Amanda's voice
2. Kludged through the hellscape of Praat scripting yet again (If you want to splice vast quantities of b/p sentences, I can share the code to brew big batches. Or pig patches.)
3. Put stuff together for Zamanda
4. Got the Game of Thrones analysis to the point where I should be able to analyze it with stats
5. Studied for upcoming midterm
6. More pygaze stuff!

Linda

1. Held RA meeting where I met Lauren (new RA)

2. Worked with Zach to get PyGaze output working

3. Wrote tool for the RAs to mark coordinates of images within visual scenes... I'm calling it PixelPy: https://github.com/llinda/pixelpy

4. Addressed mturk worker emails. Basically, a memory issue -- I tested on a machine with lots of RAM so didn't notice how greedy the experiment was. Fixed.
5. Created a more fine-grained (31 step) s-sh continuum. I'm using steps 14-19 (where larger = more s like) for the next iteration. These should correspond to the most ambiguous regions from exp2.

6. Ran 64 subjects on mturk. Pen and Mouth #3. Literally just got the results, but here's a preliminary graph -- looks a helluva lot better than previous iterations. Also my favorite comment from a worker: "Lose the pen next time, makes her look superficial." exp3.png

7. Prepared (in prog) material for review session for midterm 2 for Steve's class. Finished writing exam questions.