Paper Title:
IndoToD: A Multi-Domain Indonesian Benchmark For End-to-End Task-Oriented Dialogue Systems
Published on:
2 November 2023
Primary Category:
Computation and Language
Paper Authors:
Muhammad Dehan Al Kautsar,
Rahmah Khoirussyifa' Nurdini,
Samuel Cahyawijaya,
Genta Indra Winata,
Ayu Purwarianti
Introduces IndoToD, a new benchmark for Indonesian task-oriented dialogue
Created by extending and translating two English datasets (CamRest676 and SMD)
Covers nearly 1,000 dialogues across 4 domains: restaurant search, navigation, scheduling, weather
Allows monolingual, cross-lingual and bilingual training and evaluation of dialogue systems
Building Indonesian Task-Oriented Dialogue Systems
This paper introduces a new benchmark dataset called IndoToD to support building and evaluating task-oriented dialogue systems for Indonesian. The authors created the dataset by extending two existing English datasets through a process of delexicalization, translation, and lexicalization. The benchmark contains nearly 1,000 dialogues spanning four domains. Experiments demonstrate the value of IndoToD for monolingual, cross-lingual, and bilingual training and evaluation of task-oriented dialogue systems.
No comments yet, be the first to start the conversation...
Sign up to comment on this paper