@inproceedings{kabra-elenberg-2023-domain,
title = "Domain Private Transformers for Multi-Domain Dialog Systems",
author = "Kabra, Anmol and
Elenberg, Ethan",
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.402",
doi = "10.18653/v1/2023.findings-emnlp.402",
pages = "6049--6061",
abstract = "Large, general purpose language models have demonstrated impressive performance across many different conversational domains. While multi-domain language models achieve low overall perplexity, their outputs are not guaranteed to stay within the domain of a given input prompt. This paper proposes \textit{domain privacy} as a novel way to quantify how likely a conditional language model will leak across domains. We also develop policy functions based on token-level domain classification, and propose an efficient fine-tuning method to improve the trained model{'}s domain privacy. Experiments on membership inference attacks show that our proposed method has comparable resiliency to methods adapted from recent literature on differentially private language models.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kabra-elenberg-2023-domain">
<titleInfo>
<title>Domain Private Transformers for Multi-Domain Dialog Systems</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anmol</namePart>
<namePart type="family">Kabra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ethan</namePart>
<namePart type="family">Elenberg</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>Large, general purpose language models have demonstrated impressive performance across many different conversational domains. While multi-domain language models achieve low overall perplexity, their outputs are not guaranteed to stay within the domain of a given input prompt. This paper proposes domain privacy as a novel way to quantify how likely a conditional language model will leak across domains. We also develop policy functions based on token-level domain classification, and propose an efficient fine-tuning method to improve the trained model’s domain privacy. Experiments on membership inference attacks show that our proposed method has comparable resiliency to methods adapted from recent literature on differentially private language models.</abstract>
<identifier type="citekey">kabra-elenberg-2023-domain</identifier>
<identifier type="doi">10.18653/v1/2023.findings-emnlp.402</identifier>
<location>
<url>https://aclanthology.org/2023.findings-emnlp.402</url>
</location>
<part>
<date>2023-12</date>
<extent unit="page">
<start>6049</start>
<end>6061</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Domain Private Transformers for Multi-Domain Dialog Systems
%A Kabra, Anmol
%A Elenberg, Ethan
%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 kabra-elenberg-2023-domain
%X Large, general purpose language models have demonstrated impressive performance across many different conversational domains. While multi-domain language models achieve low overall perplexity, their outputs are not guaranteed to stay within the domain of a given input prompt. This paper proposes domain privacy as a novel way to quantify how likely a conditional language model will leak across domains. We also develop policy functions based on token-level domain classification, and propose an efficient fine-tuning method to improve the trained model’s domain privacy. Experiments on membership inference attacks show that our proposed method has comparable resiliency to methods adapted from recent literature on differentially private language models.
%R 10.18653/v1/2023.findings-emnlp.402
%U https://aclanthology.org/2023.findings-emnlp.402
%U https://doi.org/10.18653/v1/2023.findings-emnlp.402
%P 6049-6061
Markdown (Informal)
[Domain Private Transformers for Multi-Domain Dialog Systems](https://aclanthology.org/2023.findings-emnlp.402) (Kabra & Elenberg, Findings 2023)
ACL