@inproceedings{nogueira-etal-2020-document,
title = "Document Ranking with a Pretrained Sequence-to-Sequence Model",
author = "Nogueira, Rodrigo and
Jiang, Zhiying and
Pradeep, Ronak and
Lin, Jimmy",
editor = "Cohn, Trevor and
He, Yulan and
Liu, Yang",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2020",
month = nov,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2020.findings-emnlp.63/",
doi = "10.18653/v1/2020.findings-emnlp.63",
pages = "708--718",
abstract = "This work proposes the use of a pretrained sequence-to-sequence model for document ranking. Our approach is fundamentally different from a commonly adopted classification-based formulation based on encoder-only pretrained transformer architectures such as BERT. We show how a sequence-to-sequence model can be trained to generate relevance labels as {\textquotedblleft}target tokens{\textquotedblright}, and how the underlying logits of these target tokens can be interpreted as relevance probabilities for ranking. Experimental results on the MS MARCO passage ranking task show that our ranking approach is superior to strong encoder-only models. On three other document retrieval test collections, we demonstrate a zero-shot transfer-based approach that outperforms previous state-of-the-art models requiring in-domain cross-validation. Furthermore, we find that our approach significantly outperforms an encoder-only architecture in a data-poor setting. We investigate this observation in more detail by varying target tokens to probe the model`s use of latent knowledge. Surprisingly, we find that the choice of target tokens impacts effectiveness, even for words that are closely related semantically. This finding sheds some light on why our sequence-to-sequence formulation for document ranking is effective. Code and models are available at pygaggle.ai."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="nogueira-etal-2020-document">
<titleInfo>
<title>Document Ranking with a Pretrained Sequence-to-Sequence Model</title>
</titleInfo>
<name type="personal">
<namePart type="given">Rodrigo</namePart>
<namePart type="family">Nogueira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhiying</namePart>
<namePart type="family">Jiang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ronak</namePart>
<namePart type="family">Pradeep</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jimmy</namePart>
<namePart type="family">Lin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2020-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2020</title>
</titleInfo>
<name type="personal">
<namePart type="given">Trevor</namePart>
<namePart type="family">Cohn</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yulan</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yang</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Online</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>This work proposes the use of a pretrained sequence-to-sequence model for document ranking. Our approach is fundamentally different from a commonly adopted classification-based formulation based on encoder-only pretrained transformer architectures such as BERT. We show how a sequence-to-sequence model can be trained to generate relevance labels as “target tokens”, and how the underlying logits of these target tokens can be interpreted as relevance probabilities for ranking. Experimental results on the MS MARCO passage ranking task show that our ranking approach is superior to strong encoder-only models. On three other document retrieval test collections, we demonstrate a zero-shot transfer-based approach that outperforms previous state-of-the-art models requiring in-domain cross-validation. Furthermore, we find that our approach significantly outperforms an encoder-only architecture in a data-poor setting. We investigate this observation in more detail by varying target tokens to probe the model‘s use of latent knowledge. Surprisingly, we find that the choice of target tokens impacts effectiveness, even for words that are closely related semantically. This finding sheds some light on why our sequence-to-sequence formulation for document ranking is effective. Code and models are available at pygaggle.ai.</abstract>
<identifier type="citekey">nogueira-etal-2020-document</identifier>
<identifier type="doi">10.18653/v1/2020.findings-emnlp.63</identifier>
<location>
<url>https://aclanthology.org/2020.findings-emnlp.63/</url>
</location>
<part>
<date>2020-11</date>
<extent unit="page">
<start>708</start>
<end>718</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Document Ranking with a Pretrained Sequence-to-Sequence Model
%A Nogueira, Rodrigo
%A Jiang, Zhiying
%A Pradeep, Ronak
%A Lin, Jimmy
%Y Cohn, Trevor
%Y He, Yulan
%Y Liu, Yang
%S Findings of the Association for Computational Linguistics: EMNLP 2020
%D 2020
%8 November
%I Association for Computational Linguistics
%C Online
%F nogueira-etal-2020-document
%X This work proposes the use of a pretrained sequence-to-sequence model for document ranking. Our approach is fundamentally different from a commonly adopted classification-based formulation based on encoder-only pretrained transformer architectures such as BERT. We show how a sequence-to-sequence model can be trained to generate relevance labels as “target tokens”, and how the underlying logits of these target tokens can be interpreted as relevance probabilities for ranking. Experimental results on the MS MARCO passage ranking task show that our ranking approach is superior to strong encoder-only models. On three other document retrieval test collections, we demonstrate a zero-shot transfer-based approach that outperforms previous state-of-the-art models requiring in-domain cross-validation. Furthermore, we find that our approach significantly outperforms an encoder-only architecture in a data-poor setting. We investigate this observation in more detail by varying target tokens to probe the model‘s use of latent knowledge. Surprisingly, we find that the choice of target tokens impacts effectiveness, even for words that are closely related semantically. This finding sheds some light on why our sequence-to-sequence formulation for document ranking is effective. Code and models are available at pygaggle.ai.
%R 10.18653/v1/2020.findings-emnlp.63
%U https://aclanthology.org/2020.findings-emnlp.63/
%U https://doi.org/10.18653/v1/2020.findings-emnlp.63
%P 708-718
Markdown (Informal)
[Document Ranking with a Pretrained Sequence-to-Sequence Model](https://aclanthology.org/2020.findings-emnlp.63/) (Nogueira et al., Findings 2020)
ACL