LEL2B: MapTask files for Tutorial Group 1, 12 students: 9am, 11.06 40 George Square, Ivo



Below are the set of dialogue files for your tutorial group. The file assignments below ensure that each student has at least 100-300 lines of dialogue across one or more files.

Deadline for annotation: 12pm NOON on Friday 15 November. Submit your annotations to your tutor who will compile the files into a single spreadsheet and send them to Hannah (Hannah.Rohde@ed.ac.uk).

What's in each file? The first row contains column headers describing the contents of that column (e.g., "dialogue" gives the name of the dialogue, "familiarity" indicates whether the speakers knew each other or not). The next 13 rows show the practice sample dialogue from class; this just provides a reminder of what you need to do for annotation. The following rows show the dialogue that needs annotation, with empty cells in the columns for the primary annotation ("expression", "referent", "mention", "indefiniteOrDefinite", "isPronoun") and the bonus annotation ("duration").

How to access your files? As will be explained in the week 8 course materials, each student will be working with one or more files. You'll download your file(s) onto your computer, open http://sheets.google.com, open a blank spreadsheet, and then go to the spreadsheet 'File' menu, select 'Import', find the 'Upload' tab, and select the file you downloaded. You will need to click the 'Import data' button.

What do you need to do? Once the file is open, inspect the top rows to understand the different columns. The first 13 annotated lines of dialogue should look familiar from class. You then need to continue the annotation process for the rest of the file. You have a choice of whether to complete only the primary annotation of referring expressions (finding and annotating the expressions for the referent/mention/indefiniteOrDefinite/isPronoun properties) or to also measure the acoustic duration (looking up identical repeated expressions to measure their durations in the .wav files here: https://groups.inf.ed.ac.uk/maptask/signals/dialogues/). For full details about the annotation process, see the annotation instructions from class. Real data is messy -- note down any questions that come up!

To get started: Click on the link(s) below to download the file(s) you'll be working on.

q1ec1.txt (80 lines)BLOOM
q1ec2.txt (150 lines)BLOOM
q1ec3.txt (292 lines)BRESLIN
q1ec4.txt (96 lines)GERMAN
q1ec5.txt (156 lines)GERMAN
q1ec6.txt (110 lines)KENDIX
q1ec8.txt (158 lines)KENDIX
q1ec7.txt (273 lines)LI
q1nc1.txt (689 lines)MARTINEZ DE AZCONA SALVAGO
q1nc2.txt (400 lines)MITCHELL
q1nc3.txt (407 lines)NORDAN
q1nc4.txt (229 lines)SELVAM
q1nc5.txt (192 lines)SLATTERY
q1nc6.txt (169 lines)SLATTERY
q1nc7.txt (480 lines)STOCKWELL
q1nc8.txt (265 lines)ZHANG