#acl HlpLabGroup:read,write,delete,revert,admin All:read
#format wiki
#pragma section-numbers 4
#language en

= Lab Meeting, Autumn 2015, Week 6 =

This week in lab meeting I will present some new data from a set of mechanical turk studies about naming.  I'll present the analyses I've looked at so far and some preliminary interpretations, but I am mainly hoping to elicit feedback and suggestions both about the design and the analyses. No background reading is necessary. Just be natural. Be calm, thoughtful, receptive and compassionate, but also at the same time firm and relentlessly critical.

Some of the broad questions I am interested in are:
 
 1. What properties of a task cause people to encode visual information into linguistic representations or to adopt a strategy of naming? 
 1. How does the use of object naming as a strategy unfold across time as people come to understand the demands of the task they are embedded in? Are there individual differences in preference or predisposition for name encoding?
 1. How flexible are these types of memory/task-driven name representations? How do people accommodate to tasks or visual contexts in which activating the obvious label for an object will lead to confusability? 

The studies I'm presenting are a first attempt at looking at naming in an otherwise nonlinguistic task.

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

 * New book: [[https://www.palgrave.com/page/detail/researching-sociopragmatic-variability-kate-beeching/?sf1=barcode&st1=9781137373946 | ''Researching Sociopragmatic Variability'']] - KW


 * Pycha, A (2015), Co-articulatory Cues for Communication: An Investigation of Five Environments. ''Language and Speech'', Published online before print, doi:10.1177/0023830915603878. [[http://las.sagepub.com/content/early/2015/09/07/0023830915603878.abstract]] - "Although speakers did not exhibit greater temporal co-articulation in contexts that tend to diminish place cues, they did exhibit greater spatial co-articulation. This finding suggests that co-articulation serves specific communicative goals." - DFK

== What we did over last week ==

=== Florian ===
 1. '''Edited''' paper with Scott Fraundorf on adaptation to dialectal ''need''. '''Submitted''' to ''Journal of Memory and Language''.
 1. '''Edited''' Pajak et al. one final time. To be '''submitted''' this week to ''Language Learning''. (it's already accepted)
 1. '''Feedback''' on Maryam and Kodi's paper on adaptation and storage of adapted expectations across levels of linguistic analysis. The paper is now due for ''Language and Linguistics Compass'' but we'll probably need another 1-2 weeks. We also talked about the potential of another paper focusing on inference about social group membership and talker identity directed at sociolinguists. In that paper we would expand Dave's framework to more explicitly reason about such inferences.
 1. '''Feedback''' on Zach's presentation in spe about his ''What's in an accent?'' project.
 1. Met with Jane Gatewood (associate provost for global engagement) for post-mortem on CUHK visit. We've also heard back from [[http://www.se.cuhk.edu.hk/hmmeng_web/ | Helen Meng ]] (computer science, CUHK), confirming her interest in collaborations on language learning, online tutoring systems (they have some really cool ones), and the potential to apply them to speech retraining of patients with strokes that have affected their language abilities. Let me know if you're potentially interested in this project. We're also looking for graduate students interested in visiting Hong Kong to work on this project. 
 1. Met with [[https://sites.google.com/site/janegrimshaw/ | Jane Grimshaw ]] (syntax-semantics, Linguistics, Rutgers) to restart our work on the grammar-performance interface in complementizer mention (or from her perspective, the categorical choice between two different syntactic structures driven by nuanced semantics constraints of complement clause-embedding verbs). If anyone is interested in joining this project, let me know. We have two finished recall experiments. Our target audience are linguists.
 1. '''Wrote''' three more letters of recommendation and contacted relevant folks at UMich, UC Davis, UCSD. Got some enthusiastic encouragement back from some of these places about you MoFos.
 1. Finally submitted my promotion files. yay.

=== Kodi ===
 1. '''Developed''' detailed project plan (design notes, hypotheses, reasoning) for a series of experiments investigating adaptation to foreign accented speech and cross-talker generalization. Worked on '''preparing stimulus materials''' for this project.
 1. '''Developed''' detailed project plan for follow up experiments on socially-mediated syntactic alignment. The first step should be rollin' out soon!
 1. '''Met''' with Maryam, Chigusa and Florian about paper on adaptation and storage of adapted expectations across levels of linguistic analysis.
 1. Began '''revising''' said paper based on feedback from the group meeting. Still lots to do there.
 1. '''Met''' with Zach about stimulus materials for subtitle project. I think we have some awesome videos of heavily accented talkers teaching online courses.
 1. '''Met''' with Dave about several new projects: (i) using supervised and unsupervised distributional learning to assess listeners' expectations about a talker's vowel space; (ii) using a phonetic recalibration task with fricative variation to investigate the process by which listeners induce (or not) talker-independent expectations about variation that is often attributed to talker-specific causes (e.g., vocal tract differences); in other words, how do listeners induce talker-independent structure?; (iii) using dialect corpora to assess within-talker vs. between talker vowel variability (as a measure of listeners' priors about vowels), and then training a phonetic acquisition model to investigate how listeners develop expectations about the variance associate with vowel categories; and (iv) how to evaluate accent similarity using acoustic and ASR-based analyses.

=== Andrew ===
 1. Temporarily gave up fighting the stupid JS bug I've been fighting for over a week on CrowdExp and moved on to implementing more views. Now it will generate a template XLSX file for stimuli based on stimulus design, assuming the page where you input it weren't broken.
 2. Went through the stack of ~10 hard drives laying around my office and tried to figure out what they were from, labeling those I could identify.

=== Olga ===
 1. Updated the hyperlinks and graphic on the tutorial
 2. Mainly worked in Chigusa's lab this week

=== Esteban ===

 1. Helped review session for BCS 152.
 1. Helped edit the BCS 152 exam and answer key.
 1. Continued piloting eye tracking experiment with Bethany.
 1. Got feedback on said eye-tracking experiment from Mike.
 1. Met with Jasmeen about her work and mine.
 1. Continued revisions and post-hoc analyses for paper on VOT change and feedback

=== Dave ===

 1. '''Applied''' for first round of jobs (after lots of '''editing''' and '''writing''' application materials).
 1. '''Met''' with Kodi to discuss a bunch of project ideas. Most concretely, we talked about how to do a distributional learning experiment with vowels instead of VOT. We talked about a bunch of more pie-in-the-sky stuff, too.
 1. '''Outlined''' discussion and intro for fMRI animals paper.
 1. Corrected '''proofs''' for selective adaptation paper.

=== Sarah ===
 1. Met with Mike and worked on rewriting SET '''manuscript'''.
 1. BCS 111 stuff
 1. '''Recorded''' new sample stimuli for Lookslike followup.
 1. Began setting up and testing the next Gotta condition on the updated computer in 417a.

=== Dan ===

=== Amanda ===
 1. '''Revised''' and '''responded''' to reviewer comments on my mammals paper with Frontiers; decided to '''read''' a few more papers to address a few comments.
 1. '''Read''' Emiel Krahmer's squib in Computational Linguistics, basically he is just suggesting that computational linguists try to read and learn from psycholinguistic work in the same fields, and vice versa, as we could learn a lot from one another, and points out what is currently holding the field back from doing so (mostly that psycholing is looking a big picture stuff, and comp ling isn't necessary ready to answer the big questions).
 1. '''Recorded''' Chelsea for a new version of my kid task, talked to Chigusa about it.
 1. Started '''prepping''' my presentation for KurTan lab meeting about my overall interests.
 1. '''Worked''' a bit on the Kinder Lab website.
 1. Participated in / aided in / attended my childhood best friend's wedding.

=== Zach ===

 1. '''Sent''' out a giant, glorious email to my RAs instructing them all on their new task in exquisite detail.
 1. '''Met''' with Florian and Linda about eye-tracking. Learned how to do spaghetti-plot tracking by myself, instead of using giant functions.
 1. Started working the '''eye-tracking data''' into plots.  They don't bode well.
 1. Working hardcore on '''getting all my graphs and plots up to snuff''' for my presentation.
 1. Lots and lots of '''emails and scheduling'''.
 1. [Forgot to add:] '''Graded''' BCS 152 exams.

=== Linda ===
 1. Finished coding transcription task (Hear sentence; transcribe it)
 1. Finished coding freestyle classification task (Hear x speakers; drag and drop to group speakers based on how similar they sound).
 1. Finished coding subtitle training task (See transcription before or after audio plays).
 1. Met with Logan (grad student in CS) to talk about avatar project and brainstorm experimental ideas. Did some brief literature searching for relevant papers. 
 1. Worked on my presentation on Hagoort and Van Berkum 2007 for Language class. Gave said presentation.
 1. Finally parsed all the eye-tracking data into ms-ms format. 
 1. Met with Florian and Zach about said data. I have a much better idea of how to start analyzing eye-tracking data now.

=== Maryam ===

=== Wednesday ===

=== Idoia ===
 1. Analysed the data from a SPR experiment on attraction effects on Basque
 1. Went over the linguistic background questionnaire of the 80 participants to make sure everything ok there
 1. Getting to know RMarkdown
 1. Calculating average dependency lengths in Basque and Spanish data
 1. Worked on BCS 501 and 511