A common thread of retrieval-augmented methods in the existing literature
focuses on retrieving encyclopedic knowledge, such as Wikipedia, which
facilitates well-defined entity and relation spaces that can be modeled.
However, applying such methods to commonsense reasoning tasks faces two unique
challenges, i.e., the lack of a general large-scale corpus for retrieval and a
corresponding effective commonsense retriever. In this paper, we systematically
investigate how to leverage commonsense knowledge retrieval to improve
commonsense reasoning tasks. We proposed a unified framework of
retrieval-augmented commonsense reasoning (called RACo), including a newly
constructed commonsense corpus with over 20 million documents and novel
strategies for training a commonsense retriever. We conducted experiments on
four different commonsense reasoning tasks. Extensive evaluation results showed
that our proposed RACo can significantly outperform other knowledge-enhanced
method counterparts, achieving new SoTA performance on the CommonGen and CREAK
leaderboards.
Description
Retrieval Augmentation for Commonsense Reasoning: A Unified Approach
%0 Generic
%1 yu2022retrieval
%A Yu, Wenhao
%A Zhu, Chenguang
%A Zhang, Zhihan
%A Wang, Shuohang
%A Zhang, Zhuosheng
%A Fang, Yuwei
%A Jiang, Meng
%D 2022
%K llm retrieval
%T Retrieval Augmentation for Commonsense Reasoning: A Unified Approach
%U http://arxiv.org/abs/2210.12887
%X A common thread of retrieval-augmented methods in the existing literature
focuses on retrieving encyclopedic knowledge, such as Wikipedia, which
facilitates well-defined entity and relation spaces that can be modeled.
However, applying such methods to commonsense reasoning tasks faces two unique
challenges, i.e., the lack of a general large-scale corpus for retrieval and a
corresponding effective commonsense retriever. In this paper, we systematically
investigate how to leverage commonsense knowledge retrieval to improve
commonsense reasoning tasks. We proposed a unified framework of
retrieval-augmented commonsense reasoning (called RACo), including a newly
constructed commonsense corpus with over 20 million documents and novel
strategies for training a commonsense retriever. We conducted experiments on
four different commonsense reasoning tasks. Extensive evaluation results showed
that our proposed RACo can significantly outperform other knowledge-enhanced
method counterparts, achieving new SoTA performance on the CommonGen and CREAK
leaderboards.
@misc{yu2022retrieval,
abstract = {A common thread of retrieval-augmented methods in the existing literature
focuses on retrieving encyclopedic knowledge, such as Wikipedia, which
facilitates well-defined entity and relation spaces that can be modeled.
However, applying such methods to commonsense reasoning tasks faces two unique
challenges, i.e., the lack of a general large-scale corpus for retrieval and a
corresponding effective commonsense retriever. In this paper, we systematically
investigate how to leverage commonsense knowledge retrieval to improve
commonsense reasoning tasks. We proposed a unified framework of
retrieval-augmented commonsense reasoning (called RACo), including a newly
constructed commonsense corpus with over 20 million documents and novel
strategies for training a commonsense retriever. We conducted experiments on
four different commonsense reasoning tasks. Extensive evaluation results showed
that our proposed RACo can significantly outperform other knowledge-enhanced
method counterparts, achieving new SoTA performance on the CommonGen and CREAK
leaderboards.},
added-at = {2023-08-17T15:01:44.000+0200},
author = {Yu, Wenhao and Zhu, Chenguang and Zhang, Zhihan and Wang, Shuohang and Zhang, Zhuosheng and Fang, Yuwei and Jiang, Meng},
biburl = {https://www.bibsonomy.org/bibtex/251083bf5538e34f78740fefbd2530a56/lisa-ee},
description = {Retrieval Augmentation for Commonsense Reasoning: A Unified Approach},
interhash = {f60c8f62a844894d8d272d9b467b3455},
intrahash = {51083bf5538e34f78740fefbd2530a56},
keywords = {llm retrieval},
note = {cite arxiv:2210.12887Comment: EMNLP 2022 (main)},
timestamp = {2023-08-17T15:01:44.000+0200},
title = {Retrieval Augmentation for Commonsense Reasoning: A Unified Approach},
url = {http://arxiv.org/abs/2210.12887},
year = 2022
}