Log In Sign Up

CS-NET at SemEval-2020 Task 4: Siamese BERT for ComVE

In this paper, we describe our system for Task 4 of SemEval 2020, which involves differentiating between natural language statements that confirm to common sense and those that do not. The organizers propose three subtasks - first, selecting between two sentences, the one which is against common sense. Second, identifying the most crucial reason why a statement does not make sense. Third, generating novel reasons for explaining the against common sense statement. Out of the three subtasks, this paper reports the system description of subtask A and subtask B. This paper proposes a model based on transformer neural network architecture for addressing the subtasks. The novelty in work lies in the architecture design, which handles the logical implication of contradicting statements and simultaneous information extraction from both sentences. We use a parallel instance of transformers, which is responsible for a boost in the performance. We achieved an accuracy of 94.8 89


page 1

page 2

page 3

page 4


SemEval-2020 Task 4: Commonsense Validation and Explanation

In this paper, we present SemEval-2020 Task 4, Commonsense Validation an...

Does It Make Sense? And Why? A Pilot Study for Sense Making and Explanation

Introducing common sense to natural language understanding systems has r...

Temporal Common Sense Acquisition with Minimal Supervision

Temporal common sense (e.g., duration and frequency of events) is crucia...

ISCAS at SemEval-2020 Task 5: Pre-trained Transformers for Counterfactual Statement Modeling

ISCAS participated in two subtasks of SemEval 2020 Task 5: detecting cou...

Tapping BERT for Preposition Sense Disambiguation

Prepositions are frequently occurring polysemous words. Disambiguation o...

LMVE at SemEval-2020 Task 4: Commonsense Validation and Explanation using Pretraining Language Model

This paper describes our submission to subtask a and b of SemEval-2020 T...

CUHK at SemEval-2020 Task 4: CommonSense Explanation, Reasoning and Prediction with Multi-task Learning

This paper describes our system submitted to task 4 of SemEval 2020: Com...

1 Credits

This document has been adapted from the instructions for COLING-2020 proceedings compiled by Fei Liu and Liang Huang, which are, in turn, based on COLING-2018 proceedings compiled by Xiaodan Zhu and Zhiyuan Liu, which are, in turn, based on the instructions for COLING-2016 proceedings compiled by Hitoshi Isahara and Masao Utiyama, which are, in turn, based on the instructions for COLING-2014 proceedings compiled by Joachim Wagner, Liadh Kelly and Lorraine Goeuriot, which are, in turn, based on the instructions for earlier ACL proceedings, including those for ACL-2014 by Alexander Koller and Yusuke Miyao, those for ACL-2012 by Maggie Li and Michael White, those for ACL-2010 by Jing-Shing Chang and Philipp Koehn, those for ACL-2008 by Johanna D. Moore, Simone Teufel, James Allan, and Sadaoki Furui, those for ACL-2005 by Hwee Tou Ng and Kemal Oflazer, those for ACL-2002 by Eugene Charniak and Dekang Lin, and earlier ACL and EACL formats. Those versions were written by several people, including John Chen, Henry S. Thompson and Donald Walker. Additional elements were taken from the formatting instructions of the

International Joint Conference on Artificial Intelligence


2 Introduction

Place licence statement here for the camera-ready version. See Section 3.9 of the instructions for preparing a manuscript.

The following instructions are directed to authors of papers submitted to SemEval-2020 or accepted for publication in its proceedings. All authors are required to adhere to these specifications. Authors are required to provide a Portable Document Format (PDF) version of their papers. The proceedings are designed for printing on A4 paper.

SemEval papers should have the exact same format as COLING 2020 papers with two exceptions:

  1. The review process is single-blind. Submissions are not anonymous and should use the ACL camera-ready formatting.

  2. Paper titles should follow the required template. System description papers submitted by task participants should have a title of the form “[SystemName] at SemEval-2020 Task [TaskNumber]: [Insert Paper Title Here]”. Task description papers submitted by task organizers should have a title of the form “SemEval-2020 Task [TaskNumber]: [Task Name]”.

Authors from countries in which access to word-processing systems is limited should contact the publication co-chairs Derek F. Wong (, Yang Zhao ( and Liang Huang ( as soon as possible.

We may make additional instructions available at and Please check these websites regularly.

3 General Instructions

Manuscripts must be in single-column format. Type single-spaced. Start all pages directly under the top margin. See the guidelines later regarding formatting the first page. The lengths of manuscripts should not exceed the maximum page limit described in Section 5. Do not number the pages.

3.1 Electronically-available Resources

Prepare your PDF files using LaTeX with the official SemEval 2020 style files, which are adopted from COLING 2020: coling2020.sty (document style) and coling.bst (bibliography style). These files are available in at You will also find the document you are currently reading (semeval2020.pdf) and its LaTeX source code (semeval2020.tex) in

3.2 Format of Electronic Manuscript

For the production of the electronic manuscript you must use Adobe’s Portable Document Format (PDF). PDF files are usually produced from LaTeX using the pdflatex command. If your version of LaTeX produces Postscript files, you can convert these into PDF using ps2pdf or dvipdf. On Windows, you can also use Adobe Distiller to generate PDF.

Please make sure that your PDF file includes all the necessary fonts (especially tree diagrams, symbols, and fonts for non-Latin characters). When you print or create the PDF file, there is usually an option in your printer setup to include none, all or just non-standard fonts. Please make sure that you select the option of including ALL the fonts. Before sending it, test your PDF by printing it from a computer different from the one where it was created. Moreover, some word processors may generate very large PDF files, where each page is rendered as an image. Such images may reproduce poorly. In this case, try alternative ways to obtain the PDF. One way on some systems is to install a driver for a postscript printer, send your document to the printer specifying “Output to a file”, then convert the file to PDF.

It is of utmost importance to specify the A4 format (21 cm x 29.7 cm) when formatting the paper. When working with dvips, for instance, one should specify -t a4.

If you cannot meet the above requirements for the production of your electronic submission, please contact the publication co-chairs as soon as possible.

3.3 Layout

Format manuscripts with a single column to a page, in the manner these instructions are formatted. The exact dimensions for a page on A4 paper are:

  • Left and right margins: 2.5 cm

  • Top margin: 2.5 cm

  • Bottom margin: 2.5 cm

  • Width: 16.0 cm

  • Height: 24.7 cm

Papers should not be submitted on any other paper size. If you cannot meet the above requirements for the production of your electronic submission, please contact the publication co-chairs above as soon as possible.

3.4 Fonts

For reasons of uniformity, Adobe’s Times Roman font should be used. In LaTeX2e this is accomplished by putting


in the preamble.

Type of Text Font Size Style
paper title 15 pt bold
author names 12 pt bold
author affiliation 12 pt
the word “Abstract” 12 pt bold
section titles 12 pt bold
document text 11 pt
captions 11 pt
sub-captions 9 pt
abstract text 11 pt
bibliography 10 pt
footnotes 9 pt
Table 1: Font guide.

3.5 The First Page

Centre the title, author’s name(s) and affiliation(s) across the page. Do not use footnotes for affiliations. Do not include the paper ID number assigned during the submission process. Do not include the authors’ names or affiliations in the version submitted for review.

Title: Place the title centred at the top of the first page, in a 15 pt bold font. (For a complete guide to font sizes and styles, see Table 1.) Long titles should be typed on two lines without a blank line intervening. Approximately, put the title at 2.5 cm from the top of the page, followed by a blank line, then the author’s names(s), and the affiliation on the following line. Do not use only initials for given names (middle initials are allowed). Do not format surnames in all capitals (e.g., use “Schlangen” not “SCHLANGEN”). Do not format title and section headings in all capitals as well except for proper names (such as “BLEU”) that are conventionally in all capitals. The affiliation should contain the author’s complete address, and if possible, an electronic mail address. Start the body of the first page 7.5 cm from the top of the page.

The title, author names and addresses should be completely identical to those entered to the electronical paper submission website in order to maintain the consistency of author information among all publications of the conference. If they are different, the publication co-chairs may resolve the difference without consulting with you; so it is in your own interest to double-check that the information is consistent.

Abstract: Type the abstract between addresses and main body. The width of the abstract text should be smaller than main body by about 0.6 cm on each side. Centre the word Abstract in a 12 pt bold font above the body of the abstract. The abstract should be a concise summary of the general thesis and conclusions of the paper. It should be no longer than 200 words. The abstract text should be in 11 pt font.

Text: Begin typing the main body of the text immediately after the abstract, observing the single-column format as shown in the present document. Do not include page numbers.

Indent when starting a new paragraph. Use 11 pt for text and subsection headings, 12 pt for section headings and 15 pt for the title.

Licence: Include a licence statement as an unmarked (unnumbered) footnote on the first page of the final, camera-ready paper. See Section 3.9 below for details and motivation.

3.6 Sections

Headings: Type and label section and subsection headings in the style shown on the present document. Use numbered sections (Arabic numerals) in order to facilitate cross references. Number subsections with the section number and the subsection number separated by a dot, in Arabic numerals. Do not number subsubsections.

Citations: Citations within the text appear in parentheses as [Gusfield:97] or, if the author’s name appears in the text itself, as Gusfield Gusfield:97. Append lowercase letters to the year in cases of ambiguity. Treat double authors as in [Aho:72], but write as in [Chandra:81] when more than two authors are involved. Collapse multiple citations as in [Gusfield:97, Aho:72]. Also refrain from using full citations as sentence constituents. We suggest that instead of

[Gusfield:97] showed that …”

you use

“Gusfield Gusfield:97 showed that …”

If you are using the provided LaTeX and BibTeX style files, you can use the command \newcite to get “author (year)” citations.

As reviewing will be double-blind, the submitted version of the papers should not include the authors’ names and affiliations. Furthermore, self-references that reveal the author’s identity, e.g.,

“We previously showed [Gusfield:97] …”

should be avoided. Instead, use citations such as

“Gusfield Gusfield:97 previously showed … ”

Please do not use anonymous citations and do not include any of the following when submitting your paper for review: acknowledgements, project names, grant numbers, and names or URLs of resources or tools that have only been made publicly available in the last 3 weeks or are about to be made public and would compromise the anonymity of the submission. Papers that do not conform to these requirements may be rejected without review. These details can, however, be included in the camera-ready, final paper.

In LaTeX, for an anonymized submission, ensure that \colingfinalcopy at the top of this document is commented out. For a camera-ready submission, ensure that \colingfinalcopy at the top of this document is not commented out.

References: Gather the full set of references together under the heading References; place the section before any Appendices, unless they contain references. Arrange the references alphabetically by first author, rather than by order of occurrence in the text. Provide as complete a citation as possible, using a consistent format, such as the one for Computational Linguistics or the one in the Publication Manual of the American Psychological Association [APA:83]. Use of full names for authors rather than initials is preferred. A list of abbreviations for common computer science journals can be found in the ACM Computing Reviews [ACM:83].

The LaTeX and BibTeX style files provided roughly fit the American Psychological Association format, allowing regular citations, short citations and multiple citations as described above.

  • Example citing an arxiv paper: [rasooli-tetrault-2015].

  • Example article in journal citation: [Aho:72].

  • Example article in proceedings: [borsch2011].

Appendices: Appendices, if any, directly follow the text and the references (but see above). Letter them in sequence and provide an informative title: Appendix A. Title of Appendix.

3.7 Footnotes

Footnotes: Put footnotes at the bottom of the page and use 9 pt text. They may be numbered or referred to by asterisks or other symbols.111This is how a footnote should appear. Footnotes should be separated from the text by a line.222Note the line separating the footnotes from the text.

3.8 Graphics

Illustrations: Place figures, tables, and photographs in the paper near where they are first discussed, rather than at the end, if possible. Colour illustrations are discouraged, unless you have verified that they will be understandable when printed in black ink.

Captions: Provide a caption for every illustration; number each one sequentially in the form: “Figure 1. Caption of the Figure.” “Table 1. Caption of the Table.” Type the captions of the figures and tables below the body, using 11 pt text.

Narrow graphics together with the single-column format may lead to large empty spaces, see for example the wide margins on both sides of Table 1. If you have multiple graphics with related content, it may be preferable to combine them in one graphic. You can identify the sub-graphics with sub-captions below the sub-graphics numbered (a), (b), (c) etc. and using 9 pt text. The LaTeX packages wrapfig, subfig, subtable and/or subcaption may be useful.

3.9 Licence Statement

As in COLING-2014, COLING-2016, and COLING-2018, we require that authors license their camera-ready papers under a Creative Commons Attribution 4.0 International Licence (CC-BY). This means that authors (copyright holders) retain copyright but grant everybody the right to adapt and re-distribute their paper as long as the authors are credited and modifications listed. In other words, this license lets researchers use research papers for their research without legal issues. Please refer to for the licence terms.

Depending on whether you use British or American English in your paper, please include one of the following as an unmarked (unnumbered) footnote on page 1 of your paper. The LaTeX style file (coling2020.sty) adds a command blfootnote for this purpose, and usage of the command is prepared in the LaTeX source code (coling2020.tex) at the start of Section “Introduction”.

We strongly prefer that you licence your paper as the CC license above. However, if it is impossible you to use that license, please contact the COLING-2020 publication co-chairs Derek F. Wong (, Yang Zhao ( and Liang Huang (, before you submit your final version of accepted papers. (Please note that this license statement is only related to the final versions of accepted papers. It is not required for papers submitted for review.)

4 Translation of non-English Terms

It is also advised to supplement non-English characters and terms with appropriate transliterations and/or translations since not all readers understand all such characters and terms. Inline transliteration or translation can be represented in the order of: original-form transliteration “translation”.

5 Length of Submission

The maximum submission length is 9 pages (A4) of content for long papers and 4 pages (A4) of content for short papers, plus an unlimited number of pages for references (for both long and short papers). Authors of accepted papers will be given additional space in the camera-ready version to reflect space needed for changes stemming from reviewers comments.

Papers that do not conform to the specified length and formatting requirements may be rejected without review.






The acknowledgements should go immediately before the references. Do not number the acknowledgements section. Do not include this section when submitting your paper for review.