AMI Corpus Non-Scenario Meetings

List of all non-scenario meeting sets included in the AMI hub and a brief description of each

Idiap Room

IB4001-5 Four participants plan a fictitious office move that is simplified from the actual constraints a larger group faces. IB4001-2 and IB4003-4 have different groups of people. IB4005 retains one person each from the previous groups and adds two new participants.

IB4010, IB4011 Four participants discuss the selection of films to show for a fictitious movie club.

IN1001 Three colleagues discuss web-based video browsing for the CINETIS project. (Audio recordings feature strong French accents.)

IN1002 Four speech research colleagues discuss posterior probability methods.

IN1005 Four colleagues discuss the use of Probabilistic Latent Semantic Analysis methods for indexing web pages.

IN1007 Four colleagues talk about keyword spotting, Frequency-Domain Linear Prediction, and the Spectral Transform Linear Prediction technique.

IN1008, IN1009 Four colleagues discuss audio-based interactive tables and their use with microphone arrays for the CRAFT & AMI projects. (Audio recordings feature strong French accents.)

IN1012 Four speech research colleagues share their experiences from the Interspeech 2005 conference and talk about full covariance, and the differences between variable scale and conventional fixed scale analysis for feature computation.

IN1013 Four speech research colleagues talk about different approaches for coding spectral information.

IN1014 Four colleagues discuss conference room recording equipment. (Recording stops before the end of the meeting.)

IN1016 Four speech research colleagues talk about the use of slide content to enhance speech recognition.

Edinburgh Room

EN2001a,b,d,e, EN2002a-d MSc students discuss a group project design. The EN2001 series features meetings with four or five participants. Meetings from EN2002 have three or four participants.

EN2003a Three linguistics students plan a postgraduate workshop.

EN2004a Four colleagues involved with the AMI project's transcription effort discuss the procedure for checking transcripts. (All participants are female.)

EN2005a Four dialogue theory research colleagues discuss using game and rhetorical structure methods to obtain localized information about discourse structure from meetings.

EN2006a-b Four colleagues discuss annotation conventions for an astronomy project. (Participants are not properly seated in EN2006a.)

EN2009b-d Researchers discuss how to get data out of a joint eye tracking system for analysis. The participants include a post-doctoral psychologist, a computer programmer, a senior psycholinguist (EN2009d only), and a data processing specialist.