Lab Meeting, Spring 2016, Week 16

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

• Joe Fruehwald has a blog post with a really nice walk through of using list columns in dplyr (columns that can hold another data frame, a model output, etc.) with purrr::map for running the same model on subsets of data (like Wednesday showed us). -DK

• (If you're curious about how you might refactor an analysis along these lines, compare this analysis of vowel distributions at talker/gender/dialect level available here with the original version of the same analysis. I was surprised how much purrr and dplyr can work together to streamline and speed up code.) -DK

• New paper out looking at eye-tracking and speech planning. Reports that speakers' gaze during speech planning sensitive to what is in common ground and finds little evidence that this leads to delays in speech planning. Vanlangendonck, Willems, Menenti, & Hagoort, (2016). An early influence of common ground during speech planning. LCN, 3798, 1–10. doi:10.1080/23273798.2016.1148747

What we did over last week

Florian

Kodi

Andrew

1. Ran an MTurk experiment from Scott that he told me about on Wednesday and had all the data by Friday afternoon
2. Fixed some code in my list balancer that got broken by Amazon changing an API limit
3. Helped Dave get a couple PostgreSQL databases set up for his new experiment backend

4. Started changes to my experiment runner to give it an API and make it a general experiment runner instead of two slightly out of sync copies of the same codebase; one for SPR experiments and one for SocAlign experiments

5. Looked through the docs and source for the Amazon MTurk API 1.7.0 for Java to see if it's worth using it instead of or in addition to boto. Short term probably not, but if boto3 never adds MTurk support (currently they have no plans to) and they never merge my patch for boto2 (looking increasingly unlikely after a year), then quite possibly it's worth it long term.

Olga

1. Scheduling the post doc candidate talks and skype meetings
2. Finals

Esteban

1. Wrapped up initial pass testing of AutoVOT and got in contact with the creators for some thoughts on what to about it's reliability.
2. Wrapped up initial analyses for JML follow-up, just waiting on the remaining annotations before planning next steps.
3. Continued drafting up JML follow-up write-up for dissertation.

4. Read new paper by H.H. Clark in Psych Review on depiction

5. Attended the Pton language group meeting, read the back - and - forth in PNAS between Lidz and colleagues and Piantadosi & Kidd.

Dave

1. Did my final NSF GRFP activity report.

2. Wrote up hierarchical clustering based analyses of animal imaging project.

3. Improvements to experiment back end (check/record start, finish, abandon, submit; show informative error messages)

4. Deploy MTurk experiment to www (including setting up databases) and pilot

5. Drafted abstract for LabPhon workshop based on Kodi's analysis of Nationwide Speech Project data.

6. Refactored code for KL divergence and classification error for NSP vowels using purrr and friends.

7. Re-started work on implementing contrasts coding for Julia

Sarah

Dan

Amanda

1. Finished grading BCS259 Final Projects
2. Held office hours for, responded to emails about, and proctored BCS259 Final Exam

3. Worked on talk for EvoDev talk: looked at kid project, the coding the RAs have been doing

4. Worked on a Qualtrics tutorial
5. Thought about what I would want to be able to say to submit the follow up Mammals data to psychonomics
6. Met with a project coordinator at the RMSC about setting up a language learning booth to try to teach kids that language science is fun / to get more exposure for the lab with the community. Discussed possibly working with their summer camps, contemplating giving a future public talk about language research. I need to make some posters for adults to read about the tasks I'm planning on doing with kids (1 rule learning game akin to what Matt/Steve do with models and monkeys / what Gary Marcus did with infants, and 1 conceptual pact un/helpful hint task) and maybe also a poster for kids about the kinds of jobs language scientists do...

7. Re-skimmed Marcus et al (1999) on rule learning by 7 month olds for the above project.

Zach

1. Got C&G code from Linda, working on making it my own for a pilot of the pupillometry project

2. Broke homebrew, fixed homebrew, got lolcommits working
3. Stats finaaaaaallllll

Linda

1. More writing and analysis for pen in the mouth, exp 1.

2. Ran C&G new noise condition. Preprocessed the data.

3. Worked on stats final.
4. Wrote up latex tutorial for the final RA meeting of the semester.
5. Helped Thais debug some code.

Maryam

Wednesday

1. Writing
2. Did stats homework, glanced at stats final