@inproceedings{he-etal-2023-z,
title = "{Z}-Code++: A Pre-trained Language Model Optimized for Abstractive Summarization",
author = "He, Pengcheng and
Peng, Baolin and
Wang, Song and
Liu, Yang and
Xu, Ruochen and
Hassan, Hany and
Shi, Yu and
Zhu, Chenguang and
Xiong, Wayne and
Zeng, Michael and
Gao, Jianfeng and
Huang, Xuedong",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.279",
doi = "10.18653/v1/2023.acl-long.279",
pages = "5095--5112",
abstract = "This paper presents Z-Code++, a new pre-trained language model optimized for abstractive text summarization. The model extends the state-of-the-art encoder-decoder model using three techniques. First, we use a two-phase pre-training to improve the model{'}s performance on low-resource summarization tasks. The model is first pre-trained using text corpora for language understanding, then is continually pre-trained on summarization corpora for grounded text generation. Second, we replace self-attention layers in the encoder with disentangled attention layers, where each word is represented using two vectors that encode its content and position, respectively. Third, we use fusion-in-encoder, a simple yet effective method of encoding long sequences in a hierarchical manner. Z-Code++ createsa new state-of-the-art on 9 of 13 text summarization tasks across 5 languages. Our model is parameter-efficient in that it outperforms the 600x larger PaLM540B on XSum, and the finetuned 200x larger GPT3175B on SAMSum. In zero-shot and few-shot settings, our model substantially outperforms the competing models.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="he-etal-2023-z">
<titleInfo>
<title>Z-Code++: A Pre-trained Language Model Optimized for Abstractive Summarization</title>
</titleInfo>
<name type="personal">
<namePart type="given">Pengcheng</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Baolin</namePart>
<namePart type="family">Peng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Song</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yang</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruochen</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hany</namePart>
<namePart type="family">Hassan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yu</namePart>
<namePart type="family">Shi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chenguang</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wayne</namePart>
<namePart type="family">Xiong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Michael</namePart>
<namePart type="family">Zeng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jianfeng</namePart>
<namePart type="family">Gao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xuedong</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anna</namePart>
<namePart type="family">Rogers</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jordan</namePart>
<namePart type="family">Boyd-Graber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Naoaki</namePart>
<namePart type="family">Okazaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Toronto, Canada</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>This paper presents Z-Code++, a new pre-trained language model optimized for abstractive text summarization. The model extends the state-of-the-art encoder-decoder model using three techniques. First, we use a two-phase pre-training to improve the model’s performance on low-resource summarization tasks. The model is first pre-trained using text corpora for language understanding, then is continually pre-trained on summarization corpora for grounded text generation. Second, we replace self-attention layers in the encoder with disentangled attention layers, where each word is represented using two vectors that encode its content and position, respectively. Third, we use fusion-in-encoder, a simple yet effective method of encoding long sequences in a hierarchical manner. Z-Code++ createsa new state-of-the-art on 9 of 13 text summarization tasks across 5 languages. Our model is parameter-efficient in that it outperforms the 600x larger PaLM540B on XSum, and the finetuned 200x larger GPT3175B on SAMSum. In zero-shot and few-shot settings, our model substantially outperforms the competing models.</abstract>
<identifier type="citekey">he-etal-2023-z</identifier>
<identifier type="doi">10.18653/v1/2023.acl-long.279</identifier>
<location>
<url>https://aclanthology.org/2023.acl-long.279</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>5095</start>
<end>5112</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Z-Code++: A Pre-trained Language Model Optimized for Abstractive Summarization
%A He, Pengcheng
%A Peng, Baolin
%A Wang, Song
%A Liu, Yang
%A Xu, Ruochen
%A Hassan, Hany
%A Shi, Yu
%A Zhu, Chenguang
%A Xiong, Wayne
%A Zeng, Michael
%A Gao, Jianfeng
%A Huang, Xuedong
%Y Rogers, Anna
%Y Boyd-Graber, Jordan
%Y Okazaki, Naoaki
%S Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2023
%8 July
%I Association for Computational Linguistics
%C Toronto, Canada
%F he-etal-2023-z
%X This paper presents Z-Code++, a new pre-trained language model optimized for abstractive text summarization. The model extends the state-of-the-art encoder-decoder model using three techniques. First, we use a two-phase pre-training to improve the model’s performance on low-resource summarization tasks. The model is first pre-trained using text corpora for language understanding, then is continually pre-trained on summarization corpora for grounded text generation. Second, we replace self-attention layers in the encoder with disentangled attention layers, where each word is represented using two vectors that encode its content and position, respectively. Third, we use fusion-in-encoder, a simple yet effective method of encoding long sequences in a hierarchical manner. Z-Code++ createsa new state-of-the-art on 9 of 13 text summarization tasks across 5 languages. Our model is parameter-efficient in that it outperforms the 600x larger PaLM540B on XSum, and the finetuned 200x larger GPT3175B on SAMSum. In zero-shot and few-shot settings, our model substantially outperforms the competing models.
%R 10.18653/v1/2023.acl-long.279
%U https://aclanthology.org/2023.acl-long.279
%U https://doi.org/10.18653/v1/2023.acl-long.279
%P 5095-5112
Markdown (Informal)
[Z-Code++: A Pre-trained Language Model Optimized for Abstractive Summarization](https://aclanthology.org/2023.acl-long.279) (He et al., ACL 2023)
ACL
- Pengcheng He, Baolin Peng, Song Wang, Yang Liu, Ruochen Xu, Hany Hassan, Yu Shi, Chenguang Zhu, Wayne Xiong, Michael Zeng, Jianfeng Gao, and Xuedong Huang. 2023. Z-Code++: A Pre-trained Language Model Optimized for Abstractive Summarization. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5095–5112, Toronto, Canada. Association for Computational Linguistics.