@inproceedings{ramirez-etal-2023-controllable,
title = "Controllable Generation of Dialogue Acts for Dialogue Systems via Few-Shot Response Generation and Ranking",
author = "Ramirez, Angela and
Agarwal, Kartik and
Juraska, Juraj and
Garg, Utkarsh and
Walker, Marilyn",
editor = "Stoyanchev, Svetlana and
Joty, Shafiq and
Schlangen, David and
Dusek, Ondrej and
Kennington, Casey and
Alikhani, Malihe",
booktitle = "Proceedings of the 24th Annual Meeting of the Special Interest Group on Discourse and Dialogue",
month = sep,
year = "2023",
address = "Prague, Czechia",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.sigdial-1.32",
doi = "10.18653/v1/2023.sigdial-1.32",
pages = "355--369",
abstract = "Dialogue systems need to produce responses that realize multiple types of dialogue acts (DAs) with high semantic fidelity. In the past, natural language generators (NLGs) for dialogue were trained on large parallel corpora that map from a domain-specific DA and its semantic attributes to an output utterance. Recent work shows that pretrained language models (LLMs) offer new possibilities for controllable NLG using prompt-based learning. Here we develop a novel few-shot overgenerate-and-rank approach that achieves the controlled generation of DAs. We compare eight few-shot prompt styles that include a novel method of generating from textual pseudo-references using a textual style transfer approach. We develop six automatic ranking functions that identify outputs with both the correct DA and high semantic accuracy at generation time. We test our approach on three domains and four LLMs. To our knowledge, this is the first work on NLG for dialogue that automatically ranks outputs using both DA and attribute accuracy. For completeness, we compare our results to fine-tuned few-shot models trained with 5 to 100 instances per DA. Our results show that several prompt settings achieve perfect DA accuracy, and near perfect semantic accuracy (99.81{\%}) and perform better than few-shot fine-tuning.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ramirez-etal-2023-controllable">
<titleInfo>
<title>Controllable Generation of Dialogue Acts for Dialogue Systems via Few-Shot Response Generation and Ranking</title>
</titleInfo>
<name type="personal">
<namePart type="given">Angela</namePart>
<namePart type="family">Ramirez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kartik</namePart>
<namePart type="family">Agarwal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Juraj</namePart>
<namePart type="family">Juraska</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Utkarsh</namePart>
<namePart type="family">Garg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marilyn</namePart>
<namePart type="family">Walker</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-09</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 24th Annual Meeting of the Special Interest Group on Discourse and Dialogue</title>
</titleInfo>
<name type="personal">
<namePart type="given">Svetlana</namePart>
<namePart type="family">Stoyanchev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shafiq</namePart>
<namePart type="family">Joty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Schlangen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ondrej</namePart>
<namePart type="family">Dusek</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Casey</namePart>
<namePart type="family">Kennington</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Malihe</namePart>
<namePart type="family">Alikhani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Prague, Czechia</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Dialogue systems need to produce responses that realize multiple types of dialogue acts (DAs) with high semantic fidelity. In the past, natural language generators (NLGs) for dialogue were trained on large parallel corpora that map from a domain-specific DA and its semantic attributes to an output utterance. Recent work shows that pretrained language models (LLMs) offer new possibilities for controllable NLG using prompt-based learning. Here we develop a novel few-shot overgenerate-and-rank approach that achieves the controlled generation of DAs. We compare eight few-shot prompt styles that include a novel method of generating from textual pseudo-references using a textual style transfer approach. We develop six automatic ranking functions that identify outputs with both the correct DA and high semantic accuracy at generation time. We test our approach on three domains and four LLMs. To our knowledge, this is the first work on NLG for dialogue that automatically ranks outputs using both DA and attribute accuracy. For completeness, we compare our results to fine-tuned few-shot models trained with 5 to 100 instances per DA. Our results show that several prompt settings achieve perfect DA accuracy, and near perfect semantic accuracy (99.81%) and perform better than few-shot fine-tuning.</abstract>
<identifier type="citekey">ramirez-etal-2023-controllable</identifier>
<identifier type="doi">10.18653/v1/2023.sigdial-1.32</identifier>
<location>
<url>https://aclanthology.org/2023.sigdial-1.32</url>
</location>
<part>
<date>2023-09</date>
<extent unit="page">
<start>355</start>
<end>369</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Controllable Generation of Dialogue Acts for Dialogue Systems via Few-Shot Response Generation and Ranking
%A Ramirez, Angela
%A Agarwal, Kartik
%A Juraska, Juraj
%A Garg, Utkarsh
%A Walker, Marilyn
%Y Stoyanchev, Svetlana
%Y Joty, Shafiq
%Y Schlangen, David
%Y Dusek, Ondrej
%Y Kennington, Casey
%Y Alikhani, Malihe
%S Proceedings of the 24th Annual Meeting of the Special Interest Group on Discourse and Dialogue
%D 2023
%8 September
%I Association for Computational Linguistics
%C Prague, Czechia
%F ramirez-etal-2023-controllable
%X Dialogue systems need to produce responses that realize multiple types of dialogue acts (DAs) with high semantic fidelity. In the past, natural language generators (NLGs) for dialogue were trained on large parallel corpora that map from a domain-specific DA and its semantic attributes to an output utterance. Recent work shows that pretrained language models (LLMs) offer new possibilities for controllable NLG using prompt-based learning. Here we develop a novel few-shot overgenerate-and-rank approach that achieves the controlled generation of DAs. We compare eight few-shot prompt styles that include a novel method of generating from textual pseudo-references using a textual style transfer approach. We develop six automatic ranking functions that identify outputs with both the correct DA and high semantic accuracy at generation time. We test our approach on three domains and four LLMs. To our knowledge, this is the first work on NLG for dialogue that automatically ranks outputs using both DA and attribute accuracy. For completeness, we compare our results to fine-tuned few-shot models trained with 5 to 100 instances per DA. Our results show that several prompt settings achieve perfect DA accuracy, and near perfect semantic accuracy (99.81%) and perform better than few-shot fine-tuning.
%R 10.18653/v1/2023.sigdial-1.32
%U https://aclanthology.org/2023.sigdial-1.32
%U https://doi.org/10.18653/v1/2023.sigdial-1.32
%P 355-369
Markdown (Informal)
[Controllable Generation of Dialogue Acts for Dialogue Systems via Few-Shot Response Generation and Ranking](https://aclanthology.org/2023.sigdial-1.32) (Ramirez et al., SIGDIAL 2023)
ACL