Authors & Publish chair page, blog	Paper formatting page, blog	Initial paper submission chair, page, blog	Final paper submission chair, page, blog	Problematic papers - corrections page, blog
	Author [PDF,CrossCheck] tests page, blog	IEEE PDF eXpress - paper format chair, page, blog	IEEE electronic Copyright (eCf) chair, page, blog	Attendee downloads of papers page, blog
	Conference registration page, blog	Travel visas to Hungary page, blog	Conference presentations page, blog	HELP contacts WCCI2020, system
Non-Author actions	Paper reviews - authors' perspective page, blog	IEEE CrossCheck text similarity chair, page, blog	IEEE Xplore web-publish chair, page, blog
	IEEE Conference Application chair	IEEE Letter of Acquisition chair	IEEE Publication Form chair	Software systems page

Conference Guides for [IEEE-CIS, INNS] conferences : Although this guide was initially set up for IJCNN2019, and was used for IEEE WCCI 2020, hopefully some of the basic information may still help [organising committee members, authors] up to 2023-2025 (assuming a half-life of links and process information of 3-5 years?). The Authors' and Publications menus have been combined, allowing authors to see the Publications Chair perspective as well. I am no longer keeping this up to date. Bill Howell, 11Dec2020

IEEE CrossCheck analysis

Authors should be [warned, armed, prepared]!
"Text [self, external] % similarity" versus "Plagiarism"
Pre-posting of papers to arxiv.org and other sites
CrossCheck analysis
CrossCheck rejection criteria
[Questions, recommendations, experience, links] that you have could be interesting
IEEE comments and links

WARNING, WARNING!! Don't get caught by the CrossCheck similarity analysis!!
For IJCNN2017&2019, far too many great papers were rejected in an entirely avoidable way because of the %self-similarity criteria, which strongly outnumber %external-similarity rejections. Authors DEFINITELY SHOULD PRE-SCREEN THEIR PAPERS PRIOR TO SUBMISSION, preferably with iThenticate-based tools like CrossCheck, or other similar tools (NTU's TurnitIn is an example that gives similar, albeit lower overall %similarity, to CrossCheck). For authors who do not have access to CrossCheck or other equivalents, what we really need is a [free, simple] pre-check tool that they can test against a collection of pdfs that they used to prepare their paper. However, though it may be easy (or not) to program something simple :

Would it be of sufficient quality to reliably extract the criteria?
Maybe there is something already out there for that purpose?
Maybe iThenticate has something to apply to a paper that has been written and a few related publications by the authors?

"Text-similarity" [acceptance, rejection] of papers is still a required screening for the IEEE and the conference [General, Program, Technical] Co-Chairs. Although the initial reason for CrossCheck (iThenticate) may have been for plagiarism, excessive [self, external] text-similarity is itself of concern. Whether journal copyrights or publishing industry agreements also have a role in this I don't know.

The Technical Co-Chairs run ALL conference papers through the IEEE CrossCheck Portal, which analyses each paper for [text, data, analysis] that is [repeated, similar to] to content in a HUGE database of published papers. An overall rating of each paper is produced by CrossCheck, allowing the Technical Co-Chairs to focus on a reduced set of papers to examine in detail. Papers failing this initial screen are sent to the IEEE Intellectual Property group (IEEE-IP), which also has a say.

"Text [self, external] % similarity" versus "Plagiarism"

A distinction must be made between "text [self, external] % similarity" and "plagiarism". (Regrettably, I did not exclusively use the phrase "text-similarity to start with!). I assume that the common understanding of the word "plagiarism" is the taking of [concepts, derivations, formulae, data, results, conclusions] from other authors' work without proper recognition (citation and quotation), with some degree of tolerance for the usage of text even if not quoted (this does appear to a fairly common practice, if not the ideal).

IEEE CrossCheck (iThenticate-based) is really a text-similarity tool, even if IEEE seemingly developed it for plagiarism checks and refers to it as such. It is a relatively [fast, powerful, consistent] tool for basic text-similarity analysis. However, to me it is a BIG mistake to make a cut-off simply based on the overall %similarity, as many perfectly acceptable papers would be rejected, and this would treat [self, external] similarities the same. Furthermore, it took a huge amount of my time to do ~450 IJCNN2019 CrossCheck analysis, including the subjective measure of a text "chunk", "reaching into" [group, source] similarities to determine whether they were [self, external] similarities, and to ascertain whether similarities were from authors' arxiv.org pre-postings by the same authors (or at least one common author). This really isn't a sustainable approach, and I don't recommend it for other conferences. Extensions of the CrossCheck results to provide those "criteria" automatically would make it practical.

For "non-trivial" (beyond text) plagiarism from external sources (other authors), CrossCheck is a great tool that goes into far more depth and breadth than the summary criteria and analysis used for [WCCI,IEEE-CEC, IJCNN, IEEE-FUZZ], and it is one approach to [start, support] plagiarism investigations. But given the impractical nature of doing even proper text-similarity analysis with CrossCheck, I can't see this happening for all papers, even just those with overall similarity >30%.

Our main check on "non-trivial" text-similarity is still probably the peer reviewers, who are not provided with CrossCheck results at this time (~1,500 papers for IJCNN2019). Being realistic, text-similarity checks are probably not uniformly done to the same quality by all peer reviewers, and the results are probably highly variable.

Could CrossCheck text-similarity screening results be used to alert the peer reviewers of selected papers to go into more detail?
But the summary CrossCheck printouts are INADEQUATE for text-similarity investigations - they would need full log-in to CrossCheck to do this, and settign them up is non-trivial and may not be practical for more than a handful of people? (unless you've got money?)
Would peer reviewers put in sufficient time to do this, especially if two or more papers they were peer reviewing were to be checked?

Pre-posting of papers to arxiv.org and other sites

The IEEE has specifically allowed authors to post their conference paper, prior to the paper submission, to arxiv.og for several years, but NOT other sites. By 2019, they now more generally allow conference-approved sites to be used for pre-posting. Do NOT assume that you can simply post anywhere - it is safest to post to arxiv.org, and for IJCNN2019 only the following sites have been specifically allowed :

hal.archives-ouvertes.fr, biorxiv.org, tel.archives-ouvertes.fr, www.aclweb.org

ASK your conference committee BEFORE you post to other sites! Each [WCCI,IEEE-CEC, IJCNN, IEEE-FUZZ] or other conference may be different!

CrossCheck analysis

Authors are still adapting to ever-stronger constraints on text-similarity criteria, but it's still a great surprise to many! I made the mistake of note emphasizing this enough for IJCNN2019, as it is very frustrating for the authors and the Organizing Committee. As stated above, we lose far too many papers because of this!

A great view of the results of CrossCheck analysis for IJCNN2019 is in the IJCNN2019 CrossCheck spreadsheet.

I hope that this table in particular will [inform, educate] authors of the current state of the problem, so don't get caught!

Of ~930 IJCNN 2017 Anchorage Alaska paper submissions (~620 accepted? - I may have the wrong number there), going by these very approximate indications (refer to the IEEE FAQ below) :

Level One (>50% similarity) - 16 examples (almost all rejections - the one case I remember that passed only passed after reqording, given it's novel and significant nature), almost all self-plagiarism cases, 2 with large content (48%,27%) from other authors!, one cited unpublished class notes of same authors. At these levels, self-plagiarism is so high you sometimes have to look twice to see that there actually is [novel, substantive] new content!
Level Two (20-50%) - ~260 examples! Again, from a very quick random look, overwhelmingly self-plagiarism, but there were some papers with significant similarity to "external" papers by other authors. This possible "external" material is generally at a much low percentage (<<10%) and concerns the text rather than results.
Level Three (10-20%) - 170 I didn't analyse the nature of the similiarity ratings for this level.
Level Four&Five - I didn't analyse these!

CrossCheck rejection criteria

As a rough guide of how this was applied for IJCNN2019 (see the spreadsheet link above for the exact formulae):

>=50% similarity - All such papers are rejected EXCEPT when special conditions apply :
- similarities come from an arXiv posting by at least one of the same authors. This includes cases where the entire paper was pre-posted to arXiv.org before submission. At present, other arXiv-like sites are not accepted, although the IEEE may add other sites to the list in the future.
- a paper draws extensively from a thesis by one of the authors
- a paper is close to 50% similarity, and mostly involves a diverse set of source papers (mostly the authors' previous papers), and fine-grained (eg sentence to paragraph level) similarities rather than >1/2 page "chunks" of continguous text from the same source.
30% < similarity < 50% - Rejection is primarily based on :
- "chunks" of contiguous text >= 1/2 a page in length, rather than [phrase, sentence, paragraph] level, with an emphasis on [theory development, data, results, discussion, conclusions], and less of a focus on [Introduction, Literature Review, References]
- AND one of :
  - When the maximum "chunk" size is >=1/2 page (1 full column of text) :
    - >10% maximum similarity with a source by different authors
    - >20% maximum similarity with a source by the same authors (at least one author is common to both)
  - When the maximum "chunk" size is <1/2 page (1 full column of text) :
    - >15% maximum similarity with a source by different authors
    - >25% maximum similarity with a source by the same authors

Note that proper quotation of a chunk of similar text would help address the above criteria at least partially, but is rarely seen in papers.

If you use snippets of text from another author "of significant length" (certainly >=1.2 page (=1 column of text in 2-column paper), then properly quote and reference it. If your paper has even modest levels of material that might be from other authors and isn't cited, then don't be surprised if it is rejected.

[Questions, recommendations, experience, links] that you have could be interesting

Please post them to the "IEEE CrossCheck blog" (see also the link in the Menu at the top of this webpage).

Passing the CrossCheck test is one of :
Several pre-requisites are required BEFORE your final paper can be accepted! Check the menu items above for details :

Acceptance by the [Program, Technical] Chairs of your final paper, that includes required changes arising from the review (if any).
IEEE PDF eXpress Plus - your paper must pass the paper format tests when finally submitted.
IEEE electronic Copyright form (eCf) - SIGNED (by your oganisation) and accepted (by the IEEE)
CrossCheck text-similarity - Authors DEFINITELY SHOULD pre-test their papers before initial submission! This is a BIG cause for rejection of great papers, and it is almost entirely avoidable!.
[WCCI,IEEE-CEC, IJCNN, IEEE-FUZZ] conference registration - At least one of the co-authors must be registered to attend the conference and present the paper (oral or poster).

IEEE comments and links

IEEE CrossCheck Portal - understanding the text-similarity checks
From an IEEE webpage :

"... CrossCheck compares submitted manuscripts against a very large database of published technical papers (as well as over 6 billion web pages), and provides editors with a summary report that highlights the similarity to previously published work. The publisher can then follow up to isolate and review the high-scoring papers as necessary.

Plagiarism detection systems are only as effective as the amount and quality of the source content within them. CrossCheck's significant advantage over all other similar services is that it includes the indexed, full-text content of participating CrossCheck member publishers.

IEEE policy requires that all accepted papers must be checked for plagiarism.

The IEEE CrossCheck Portal is available to all conference organizers and periodical editors to help screen manuscripts for plagiarized material. The IPR Office partnered with the IEEE Publications Technology department to develop the CrossCheck Portal as a stand-alone web application that can be used by any publications volunteer at any time. ..."

IEEE's "A Plagiarism FAQ"
provides a great description of the plagiarism problem, and how [editors, author-victims, readers] can deal with the problem. Their guidelines include five levels of misconduct :

Level One pertains to the uncredited verbatim copying of a full paper, or the verbatim copying of a major portion (> 50%), or verbatim copying within more than one paper by the same author(s).
Level Two pertains to the uncredited verbatim copying of large portion (between 20-50%) or verbatim copying within more than one paper by the same author(s).
Level Three pertains to the uncredited verbatim copying of individual elements (paragraph(s), sentence(s), illustration(s), etc.) resulting in a significant portion (up to 20%) within a paper
Level Four pertains to uncredited improper paraphrasing of pages or paragraphs
Level Five pertains to the credited verbatim copying of a major portion of a paper without clear delineation (e.g., quotes or indents)

You cannot jump over a cliff with this, as :

You can see that the massively dominant portion are "self plagiarism" cases.
There will typically be similarities in all papers, no matter how unique and original.
Many similarities are due to using common [phrases, expressions, styles], and it is natural for these to occur, especially with your own work, but also by convention.
Many, if not most, hits occur for introductory descriptions having little to do with the original content of the work's [data, analysis, conclusions].
Serious [external, self] plagiarism of results is possible without tripping the thresholds above (for example - slightly changed [data, analysis, results, wording], so it is still very important for reviewers to be alert and mention suspicious cases in their reviews, and for the Chairs to look closely at some papers.

Author reports of plagiarism

During the paper review process, only reviewers are in a position to check for plagiarism. Once published, however, the IEEE suggests that offended authors :

"... Once you have collected all the necessary material, you should submit your case to the editor of the publication (or the sponsoring IEEE Society of the publication) in which the misconduct occurred. ..."

10Oct2018 A mailto link for the conference will be posted later.

Publications bans

The IEEE maintains confidential lists of authors who are subject to an "IEEE publications ban" for a predetermined period (usually a year or two). These lists are NOT available to the public. The authors involved will have been notified independently of the whole process regarding their papers. For IJCNN2017 Anchorage, there were at least 5 cases of authors possibly in an IEEE publications ban period, but non were confirmed by the IEEE (it's easy to mistake author identitites with this, another warning about jumping over cliffs!).

IEEE FAQs

What is CrossCheck? From an Elsevier webpage :

"... Cases of suspected plagiarism are rarely limited to one journal or publisher. Software solutions, therefore, require cooperation between (ideally) all publishing houses. In 2008, Crossref and the STM publishing community came together to develop CrossCheck, a service that helps editors to verify the originality of papers. CrossCheck is powered by the Ithenticate software from iParadigms, known in the academic community as providers of Turnitin.

Over 200 CrossRef members, including Elsevier, collaborate by donating full-text journal articles and book chapters to create a unique database of over 50 million articles. Note that even this database is not entirely exhaustive: research published by non-participating publishers or before the digital era may be absent. Elsevier's contribution consists of 10 million articles and 7000 books and is ever-increasing as all newly published articles are added.

For a searchable list of all participating publishers, please visit: www.crossref.org/crosscheck_members.html
For a searchable list of all journals in the CrossCheck database, please visit: www.ithenticate.com/search ..."

iThenticate itself uses the term "Crossref", which seems to be the same as Cross Check? ... strange.

"... Crossref Similarity Check powered by iThenticate is an initiative between iThenticate and Crossref®, a not-for-profit membership organization for scholarly publishing working to make content easy to find, cite, link, and assess. With thousands of journals sharing published works with the iThenticate database, the Crossref Similarity Check service offers publishers a way to compare documents against the largest comparison database of scientific, technical and medical content in the world. ..."

What is iThenticate? From the iThenticate website :

"... iThenticate is the leading provider of professional plagiarism detection and prevention technology used worldwide by scholarly publishers and research institutions to ensure the originality of written work before publication. iThenticate helps editors, authors and researchers prevent misconduct by comparing manuscripts against its database of over 60 billion web pages and 155 million content items, including 49 million works from 800 scholarly publisher participants of Crossref Similarity Check powered by iThenticate software. iThenticate is developed by Turnitin, the leader in plagiarism and originality checking for educational institutions worldwide. The company is headquartered in Oakland, California with an international office in Newcastle, United Kingdom. ..."

The "Authors' Guide Blogs" (see links in manu near top of page) are in an early draft form, and are being run on a trial basis. Comments are moderated by the Publications Chair before posting, so expect delays of at least a day or so before they appear. Note that emails have been edited - usually by [omitting salutations, endings], but also by omitting material not relevant to the "theme" under which the emails are placed below.

REMINDERS : This Authors' Guide is specific to the IEEE conference paper system, and for the most part links to, or repeats, information that is COPYRIGHTED by the IEEE. Keep in mind that the IEEE processes something like 1,800 conferences per year, and delays exceeding a week are to be expected, with longer delays during high-activity parts of the year. So don't wait too long before starting each process!
Author preparations for the conference include (roughly in squence) :

Initial submission of your paper to the IEEE paper review system.
Receipt of an email paper acceptance notification from the Program Co-Chairs via the IEEE paper review system.
Payment of the conference registration fee. At least one of the co-authors must be registered for the conference to present the paper.
IEEE approval of a signed IEEE electronic Copyright form (eCf) (online submission)
Submission and approval of a properly revised final paper.
Travel visa approval, for citizens of countries required to do that by Hungary.
?What else have I forgotten? ...

Directory of available files for this webpage

Many thanks to our Sponsors & Exhibitors

Ongoing support : [IEEE, IEEE-CIS] for [CEC, FUZZ, IJCNN]; INNS for IJCNN; [IET, EPS] for CEC

2020 IEEE World Congress on Computational Intelligence, Glasgow, Scotland.....19-24 July 2020

2019 International Joint Conference on Neural Networks, Budapest, HUNGARY.....14-19 July 2019

Silver(mailto)	Silver (mailto)	Bronze (mailto)	Bronze (mailto)