Jolanta Bachan

Let's watch! Successes, Challenges and Opportunities for Speech Technology in Conversational Agents by Shehzad Mevawalla, Amazon Alexa, a keynote speech at Interspeech 2020

Download the dialog and annotate dialogue acts.

Guidelines for using ISO standard 24617-2 by Harry Bunt, Tilburg University, 2019

Recording a dialogue on ZOOM on two seperate channels:
ZOOM -> Profil użytkownika (w prawym górnym rogu) -> Ustawienia
Nagranie -> Zaznacz okienko: Nagrywaj osobny plik audio każdego uczestnika

Apps for drawing diagrams: OpenOffice Draw, DrawIO: app.diagrams.net, Konipio Club: kinopio.club

Analyse your dialogue on dialogue act level, calculate statistics => export TextGrid file format to CSV format, textgrid2csv converter

1. Finite State Machine Designer
2. Build a Finate-State Transducer of your dialogue => NDFST interpreter
3. Write a report of your work

INPUT: s s u s u s u s u

# METADATA: Demo FST file for NDFST interpreter.
# The interpreter features generous use of white space, error messages for FST
# configuration file input, optional prettyprinted trace for fst run.

# Initial state:
initial = q0
# Set of terminal states:
terminal = q5
# Transition quadruples :
fst =
   q0, s, self_intro, q1;
   q1, s, info_seek, q2;
   q2, u, info_provide, q1;
   q2, u, info_provide, q3;
   q3, s, ferewall, q4;
   q4, u, ferewall, q5

--------------------
s - system
u - user
INPUT: s s u s u u s u s u

# METADATA: Demo FST file for NDFST interpreter.
# The interpreter features generous use of white space, error messages for FST
# configuration file input, optional prettyprinted trace for fst run.

# Initial state:
initial = q0
# Set of terminal states:
terminal = q3, q5
# Transition quadruples :
fst =
   q0, s, self_intro, q1;
   q1, s, info_seek, q2;
   q2, u, info_provide, q1;
   q2, u, info_provide, q2;
   q2, u, info_provide, q3;
   q3, s, ferewall, q4;
   q4, u, ferewall, q5