Interaction data

We provide eaf label files (read by Elan), mpg videos, wav audio signals. We exploit the following ressources recorded during interactive scenarios:
  1. Sentence repeating game. Computer-mediated face-to-face interactions (one target female speaker vs. ten different interlocutors: Sol, Nico, NG, NatVal, Nat, Marion, Claire, CL, Chris & Barbara): eaf, wav, mpg. The game consists in asking an instructor to read aloud one semantically unpredictable sentence to a respondent who has to repeat it back with no errors. One interaction consists in 10 games followed by another series of 10 games where the roles are exchanged. Speech and gaze of each dyad are monitored.
  2. Put-that-there. Face-to-face interactions (one male instructor vs. three different manipulators: pierre, frederic & mael): eaf, wav, mpg_pierre, mpg_frederic, mpg_mael. The game consists in asking an instructor to reproduce a cube arrangement (consisting of 10 cubes) in a chessboard by a multimodal "put-that-there". He must instruct a manipulator - who has no access to the arrangement - to perform the move. One interaction consists in 10 games. Speech, gaze, hand & head movements of the instructor are monitored.
  3. RL/RI. Neuropsychological interviews (one female interviewer vs. five different subjects). The RL/RI  (see below) consists in testing mnesic abilities by a series of immediate piloted recall and delayed free recall and piloted recalls of 16 items.

Neuropsychological tests and evaluation protocols

Our robots will attempt to conduct the following neuropsychological tests:
  1. test RL/RI16: examination sheet (French)
  2. Mini-Mental State Examination: examination sheet (French)
  3. Montreal Cognitive Assessment (MoCA): examination sheet (French, see sample results) & examination sheet (English)