@inproceedings{varshney-etal-2022-prompt,
title = "Prompt Augmented Generative Replay via Supervised Contrastive Learning for Lifelong Intent Detection",
author = "Varshney, Vaibhav and
Patidar, Mayur and
Kumar, Rajat and
Vig, Lovekesh and
Shroff, Gautam",
editor = "Carpuat, Marine and
de Marneffe, Marie-Catherine and
Meza Ruiz, Ivan Vladimir",
booktitle = "Findings of the Association for Computational Linguistics: NAACL 2022",
month = jul,
year = "2022",
address = "Seattle, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.findings-naacl.84/",
doi = "10.18653/v1/2022.findings-naacl.84",
pages = "1113--1127",
abstract = "Identifying all possible user intents for a dialog system at design time is challenging even for skilled domain experts. For practical applications, novel intents may have to be inferred incrementally on the fly. This typically entails repeated retraining of the intent detector on both the existing and novel intents which can be expensive and would require storage of all past data corresponding to prior intents. In this paper, the objective is to continually train an intent detector on new intents while maintaining performance on prior intents without mandating access to prior intent data. Several data replay-based approaches have been introduced to avoid catastrophic forgetting during continual learning, including exemplar and generative replay. Current generative replay approaches struggle to generate representative samples because the generation is conditioned solely on the class/task label. Motivated by the recent work around prompt-based generation via pre-trained language models (PLMs), we employ generative replay using PLMs for incremental intent detection. Unlike exemplar replay, we only store the relevant contexts per intent in memory and use these stored contexts (with the class label) as prompts for generating intent-specific utterances. We use a common model for both generation and classification to promote optimal sharing of knowledge across both tasks. To further improve generation, we employ supervised contrastive fine-tuning of the PLM. Our proposed approach achieves state-of-the-art (SOTA) for lifelong intent detection on four public datasets and even outperforms exemplar replay-based approaches. The technique also achieves SOTA on a lifelong relation extraction task, suggesting that the approach is extendable to other continual learning tasks beyond intent detection."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="varshney-etal-2022-prompt">
<titleInfo>
<title>Prompt Augmented Generative Replay via Supervised Contrastive Learning for Lifelong Intent Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Vaibhav</namePart>
<namePart type="family">Varshney</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mayur</namePart>
<namePart type="family">Patidar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rajat</namePart>
<namePart type="family">Kumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lovekesh</namePart>
<namePart type="family">Vig</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gautam</namePart>
<namePart type="family">Shroff</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2022-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: NAACL 2022</title>
</titleInfo>
<name type="personal">
<namePart type="given">Marine</namePart>
<namePart type="family">Carpuat</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marie-Catherine</namePart>
<namePart type="family">de Marneffe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ivan</namePart>
<namePart type="given">Vladimir</namePart>
<namePart type="family">Meza Ruiz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Seattle, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Identifying all possible user intents for a dialog system at design time is challenging even for skilled domain experts. For practical applications, novel intents may have to be inferred incrementally on the fly. This typically entails repeated retraining of the intent detector on both the existing and novel intents which can be expensive and would require storage of all past data corresponding to prior intents. In this paper, the objective is to continually train an intent detector on new intents while maintaining performance on prior intents without mandating access to prior intent data. Several data replay-based approaches have been introduced to avoid catastrophic forgetting during continual learning, including exemplar and generative replay. Current generative replay approaches struggle to generate representative samples because the generation is conditioned solely on the class/task label. Motivated by the recent work around prompt-based generation via pre-trained language models (PLMs), we employ generative replay using PLMs for incremental intent detection. Unlike exemplar replay, we only store the relevant contexts per intent in memory and use these stored contexts (with the class label) as prompts for generating intent-specific utterances. We use a common model for both generation and classification to promote optimal sharing of knowledge across both tasks. To further improve generation, we employ supervised contrastive fine-tuning of the PLM. Our proposed approach achieves state-of-the-art (SOTA) for lifelong intent detection on four public datasets and even outperforms exemplar replay-based approaches. The technique also achieves SOTA on a lifelong relation extraction task, suggesting that the approach is extendable to other continual learning tasks beyond intent detection.</abstract>
<identifier type="citekey">varshney-etal-2022-prompt</identifier>
<identifier type="doi">10.18653/v1/2022.findings-naacl.84</identifier>
<location>
<url>https://aclanthology.org/2022.findings-naacl.84/</url>
</location>
<part>
<date>2022-07</date>
<extent unit="page">
<start>1113</start>
<end>1127</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Prompt Augmented Generative Replay via Supervised Contrastive Learning for Lifelong Intent Detection
%A Varshney, Vaibhav
%A Patidar, Mayur
%A Kumar, Rajat
%A Vig, Lovekesh
%A Shroff, Gautam
%Y Carpuat, Marine
%Y de Marneffe, Marie-Catherine
%Y Meza Ruiz, Ivan Vladimir
%S Findings of the Association for Computational Linguistics: NAACL 2022
%D 2022
%8 July
%I Association for Computational Linguistics
%C Seattle, United States
%F varshney-etal-2022-prompt
%X Identifying all possible user intents for a dialog system at design time is challenging even for skilled domain experts. For practical applications, novel intents may have to be inferred incrementally on the fly. This typically entails repeated retraining of the intent detector on both the existing and novel intents which can be expensive and would require storage of all past data corresponding to prior intents. In this paper, the objective is to continually train an intent detector on new intents while maintaining performance on prior intents without mandating access to prior intent data. Several data replay-based approaches have been introduced to avoid catastrophic forgetting during continual learning, including exemplar and generative replay. Current generative replay approaches struggle to generate representative samples because the generation is conditioned solely on the class/task label. Motivated by the recent work around prompt-based generation via pre-trained language models (PLMs), we employ generative replay using PLMs for incremental intent detection. Unlike exemplar replay, we only store the relevant contexts per intent in memory and use these stored contexts (with the class label) as prompts for generating intent-specific utterances. We use a common model for both generation and classification to promote optimal sharing of knowledge across both tasks. To further improve generation, we employ supervised contrastive fine-tuning of the PLM. Our proposed approach achieves state-of-the-art (SOTA) for lifelong intent detection on four public datasets and even outperforms exemplar replay-based approaches. The technique also achieves SOTA on a lifelong relation extraction task, suggesting that the approach is extendable to other continual learning tasks beyond intent detection.
%R 10.18653/v1/2022.findings-naacl.84
%U https://aclanthology.org/2022.findings-naacl.84/
%U https://doi.org/10.18653/v1/2022.findings-naacl.84
%P 1113-1127
Markdown (Informal)
[Prompt Augmented Generative Replay via Supervised Contrastive Learning for Lifelong Intent Detection](https://aclanthology.org/2022.findings-naacl.84/) (Varshney et al., Findings 2022)
ACL