@inproceedings{basile-etal-2023-zero,
title = "Zero-Shot Data Maps. Efficient Dataset Cartography Without Model Training",
author = "Basile, Angelo and
Franco-Salvador, Marc and
Rosso, Paolo",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.554",
doi = "10.18653/v1/2023.findings-emnlp.554",
pages = "8264--8277",
abstract = "Data Maps (Swayamdipta, et al. 2020) have emerged as a powerful tool for diagnosing large annotated datasets. Given a model fitted on a dataset, these maps show each data instance from the dataset in a 2-dimensional space defined by a) the model{'}s confidence in the true class and b) the variability of this confidence. In previous work, confidence and variability are usually computed using training dynamics, which requires the fitting of a strong model to the dataset. In this paper, we introduce a novel approach: Zero-Shot Data Maps based on fast bi-encoder networks. For each data point, confidence on the true label and variability are computed over the members of an ensemble of zero-shot models constructed with different {---} but semantically equivalent {---} label descriptions, i.e., textual representations of each class in a given label space. We conduct a comparative analysis of maps compiled using traditional training dynamics and our proposed zero-shot models across various datasets. Our findings reveal that Zero-Shot Data Maps generally match those produced by the traditional method while delivering up to a 14x speedup. The code is available [here](https://github.com/symanto-research/zeroshot-cartography).",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="basile-etal-2023-zero">
<titleInfo>
<title>Zero-Shot Data Maps. Efficient Dataset Cartography Without Model Training</title>
</titleInfo>
<name type="personal">
<namePart type="given">Angelo</namePart>
<namePart type="family">Basile</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marc</namePart>
<namePart type="family">Franco-Salvador</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paolo</namePart>
<namePart type="family">Rosso</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2023</title>
</titleInfo>
<name type="personal">
<namePart type="given">Houda</namePart>
<namePart type="family">Bouamor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Juan</namePart>
<namePart type="family">Pino</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kalika</namePart>
<namePart type="family">Bali</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Singapore</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Data Maps (Swayamdipta, et al. 2020) have emerged as a powerful tool for diagnosing large annotated datasets. Given a model fitted on a dataset, these maps show each data instance from the dataset in a 2-dimensional space defined by a) the model’s confidence in the true class and b) the variability of this confidence. In previous work, confidence and variability are usually computed using training dynamics, which requires the fitting of a strong model to the dataset. In this paper, we introduce a novel approach: Zero-Shot Data Maps based on fast bi-encoder networks. For each data point, confidence on the true label and variability are computed over the members of an ensemble of zero-shot models constructed with different — but semantically equivalent — label descriptions, i.e., textual representations of each class in a given label space. We conduct a comparative analysis of maps compiled using traditional training dynamics and our proposed zero-shot models across various datasets. Our findings reveal that Zero-Shot Data Maps generally match those produced by the traditional method while delivering up to a 14x speedup. The code is available [here](https://github.com/symanto-research/zeroshot-cartography).</abstract>
<identifier type="citekey">basile-etal-2023-zero</identifier>
<identifier type="doi">10.18653/v1/2023.findings-emnlp.554</identifier>
<location>
<url>https://aclanthology.org/2023.findings-emnlp.554</url>
</location>
<part>
<date>2023-12</date>
<extent unit="page">
<start>8264</start>
<end>8277</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Zero-Shot Data Maps. Efficient Dataset Cartography Without Model Training
%A Basile, Angelo
%A Franco-Salvador, Marc
%A Rosso, Paolo
%Y Bouamor, Houda
%Y Pino, Juan
%Y Bali, Kalika
%S Findings of the Association for Computational Linguistics: EMNLP 2023
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F basile-etal-2023-zero
%X Data Maps (Swayamdipta, et al. 2020) have emerged as a powerful tool for diagnosing large annotated datasets. Given a model fitted on a dataset, these maps show each data instance from the dataset in a 2-dimensional space defined by a) the model’s confidence in the true class and b) the variability of this confidence. In previous work, confidence and variability are usually computed using training dynamics, which requires the fitting of a strong model to the dataset. In this paper, we introduce a novel approach: Zero-Shot Data Maps based on fast bi-encoder networks. For each data point, confidence on the true label and variability are computed over the members of an ensemble of zero-shot models constructed with different — but semantically equivalent — label descriptions, i.e., textual representations of each class in a given label space. We conduct a comparative analysis of maps compiled using traditional training dynamics and our proposed zero-shot models across various datasets. Our findings reveal that Zero-Shot Data Maps generally match those produced by the traditional method while delivering up to a 14x speedup. The code is available [here](https://github.com/symanto-research/zeroshot-cartography).
%R 10.18653/v1/2023.findings-emnlp.554
%U https://aclanthology.org/2023.findings-emnlp.554
%U https://doi.org/10.18653/v1/2023.findings-emnlp.554
%P 8264-8277
Markdown (Informal)
[Zero-Shot Data Maps. Efficient Dataset Cartography Without Model Training](https://aclanthology.org/2023.findings-emnlp.554) (Basile et al., Findings 2023)
ACL