Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Rosenbaum, Andy, Kharazmi, Pegah, Banijamali, Ershad, Zeng, Lu, DiPersio, Christopher, Wei, Pan, Oz, Gokmen, Chung, Clement, Owczarzak, Karolina, Triefenbach, Fabian, Hamza, Wael
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2412.05388
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909419157061632
author Rosenbaum, Andy
Kharazmi, Pegah
Banijamali, Ershad
Zeng, Lu
DiPersio, Christopher
Wei, Pan
Oz, Gokmen
Chung, Clement
Owczarzak, Karolina
Triefenbach, Fabian
Hamza, Wael
author_facet Rosenbaum, Andy
Kharazmi, Pegah
Banijamali, Ershad
Zeng, Lu
DiPersio, Christopher
Wei, Pan
Oz, Gokmen
Chung, Clement
Owczarzak, Karolina
Triefenbach, Fabian
Hamza, Wael
contents We present CALICO, a method to fine-tune Large Language Models (LLMs) to localize conversational agent training data from one language to another. For slots (named entities), CALICO supports three operations: verbatim copy, literal translation, and localization, i.e. generating slot values more appropriate in the target language, such as city and airport names located in countries where the language is spoken. Furthermore, we design an iterative filtering mechanism to discard noisy generated samples, which we show boosts the performance of the downstream conversational agent. To prove the effectiveness of CALICO, we build and release a new human-localized (HL) version of the MultiATIS++ travel information test set in 8 languages. Compared to the original human-translated (HT) version of the test set, we show that our new HL version is more challenging. We also show that CALICO out-performs state-of-the-art LINGUIST (which relies on literal slot translation out of context) both on the HT case, where CALICO generates more accurate slot translations, and on the HL case, where CALICO generates localized slots which are closer to the HL test set.
format Preprint
id arxiv_https___arxiv_org_abs_2412_05388
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle CALICO: Conversational Agent Localization via Synthetic Data Generation
Rosenbaum, Andy
Kharazmi, Pegah
Banijamali, Ershad
Zeng, Lu
DiPersio, Christopher
Wei, Pan
Oz, Gokmen
Chung, Clement
Owczarzak, Karolina
Triefenbach, Fabian
Hamza, Wael
Computation and Language
Artificial Intelligence
Machine Learning
We present CALICO, a method to fine-tune Large Language Models (LLMs) to localize conversational agent training data from one language to another. For slots (named entities), CALICO supports three operations: verbatim copy, literal translation, and localization, i.e. generating slot values more appropriate in the target language, such as city and airport names located in countries where the language is spoken. Furthermore, we design an iterative filtering mechanism to discard noisy generated samples, which we show boosts the performance of the downstream conversational agent. To prove the effectiveness of CALICO, we build and release a new human-localized (HL) version of the MultiATIS++ travel information test set in 8 languages. Compared to the original human-translated (HT) version of the test set, we show that our new HL version is more challenging. We also show that CALICO out-performs state-of-the-art LINGUIST (which relies on literal slot translation out of context) both on the HT case, where CALICO generates more accurate slot translations, and on the HL case, where CALICO generates localized slots which are closer to the HL test set.
title CALICO: Conversational Agent Localization via Synthetic Data Generation
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2412.05388