Unification-based Reconstruction of Multi-hop Explanations for Science QuestionsCitation formats

Standard

Unification-based Reconstruction of Multi-hop Explanations for Science Questions. / Valentino, Marco; Thayaparan, Mokanarangan; Freitas, Andre.

Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics. Vol. Main Volume Association for Computational Linguistics, 2021. p. 200-211.

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Harvard

Valentino, M, Thayaparan, M & Freitas, A 2021, Unification-based Reconstruction of Multi-hop Explanations for Science Questions. in Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics. vol. Main Volume, Association for Computational Linguistics, pp. 200-211.

APA

Valentino, M., Thayaparan, M., & Freitas, A. (2021). Unification-based Reconstruction of Multi-hop Explanations for Science Questions. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics (Vol. Main Volume, pp. 200-211). Association for Computational Linguistics.

Vancouver

Valentino M, Thayaparan M, Freitas A. Unification-based Reconstruction of Multi-hop Explanations for Science Questions. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics. Vol. Main Volume. Association for Computational Linguistics. 2021. p. 200-211

Author

Valentino, Marco ; Thayaparan, Mokanarangan ; Freitas, Andre. / Unification-based Reconstruction of Multi-hop Explanations for Science Questions. Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics. Vol. Main Volume Association for Computational Linguistics, 2021. pp. 200-211

Bibtex

@inproceedings{33f17bf5da844e09813a79a415c2d4cb,
title = "Unification-based Reconstruction of Multi-hop Explanations for Science Questions",
abstract = "This paper presents a novel framework for reconstructing multi-hop explanations in science Question Answering (QA). While existing approaches for multi-hop reasoning build explanations considering each question in isolation, we propose a method to leverage explanatory patterns emerging in a corpus of scientific explanations. Specifically, the framework ranks a set of atomic facts by integrating lexical relevance with the notion of unification power, estimated analysing explanations for similar questions in the corpus.An extensive evaluation is performed on the Worldtree corpus, integrating k-NN clustering and Information Retrieval (IR) techniques. We present the following conclusions: (1) The proposed method achieves results competitive with Transformers, yet being orders of magnitude faster, a feature that makes it scalable to large explanatory corpora (2) The unification-based mechanism has a key role in reducing semantic drift, contributing to the reconstruction of many hops explanations (6 or more facts) and the ranking of complex inference facts (+12.0 Mean Average Precision) (3) Crucially, the constructed explanations can support downstream QA models, improving the accuracy of BERT by up to 10% overall.",
author = "Marco Valentino and Mokanarangan Thayaparan and Andre Freitas",
year = "2021",
month = apr,
day = "21",
language = "English",
volume = " Main Volume",
pages = "200--211",
booktitle = "Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics",
publisher = "Association for Computational Linguistics",
address = "United States",

}

RIS

TY - GEN

T1 - Unification-based Reconstruction of Multi-hop Explanations for Science Questions

AU - Valentino, Marco

AU - Thayaparan, Mokanarangan

AU - Freitas, Andre

PY - 2021/4/21

Y1 - 2021/4/21

N2 - This paper presents a novel framework for reconstructing multi-hop explanations in science Question Answering (QA). While existing approaches for multi-hop reasoning build explanations considering each question in isolation, we propose a method to leverage explanatory patterns emerging in a corpus of scientific explanations. Specifically, the framework ranks a set of atomic facts by integrating lexical relevance with the notion of unification power, estimated analysing explanations for similar questions in the corpus.An extensive evaluation is performed on the Worldtree corpus, integrating k-NN clustering and Information Retrieval (IR) techniques. We present the following conclusions: (1) The proposed method achieves results competitive with Transformers, yet being orders of magnitude faster, a feature that makes it scalable to large explanatory corpora (2) The unification-based mechanism has a key role in reducing semantic drift, contributing to the reconstruction of many hops explanations (6 or more facts) and the ranking of complex inference facts (+12.0 Mean Average Precision) (3) Crucially, the constructed explanations can support downstream QA models, improving the accuracy of BERT by up to 10% overall.

AB - This paper presents a novel framework for reconstructing multi-hop explanations in science Question Answering (QA). While existing approaches for multi-hop reasoning build explanations considering each question in isolation, we propose a method to leverage explanatory patterns emerging in a corpus of scientific explanations. Specifically, the framework ranks a set of atomic facts by integrating lexical relevance with the notion of unification power, estimated analysing explanations for similar questions in the corpus.An extensive evaluation is performed on the Worldtree corpus, integrating k-NN clustering and Information Retrieval (IR) techniques. We present the following conclusions: (1) The proposed method achieves results competitive with Transformers, yet being orders of magnitude faster, a feature that makes it scalable to large explanatory corpora (2) The unification-based mechanism has a key role in reducing semantic drift, contributing to the reconstruction of many hops explanations (6 or more facts) and the ranking of complex inference facts (+12.0 Mean Average Precision) (3) Crucially, the constructed explanations can support downstream QA models, improving the accuracy of BERT by up to 10% overall.

M3 - Conference contribution

VL - Main Volume

SP - 200

EP - 211

BT - Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics

PB - Association for Computational Linguistics

ER -