[Hotstorage-chairs] HotStorage paper similarity

[Keith Smith]

Thanks Bryan,

>  if the papers' contents significantly overlap, there is an ethical case
to be made about duplicate submission.

To be clear -- I don't think the papers' content does not overlap in that
manner. The papers are each about a different topic but they use similar
approaches and feel like they were written using the same template.

One paper proposes that authors of KV store papers should measure and
report how much performance variance occurs if they change the data layout
(such as row store vs column store). Another paper proposes that authors
should instead report how much variance stems from the choice of compiler
or the implementation of key storage primitives such as compression,
hashing, etc.

All the papers frame this additional information as a concrete artifact
authors should provide -- a "contract" or "certificate". All of the papers
have a brief motivation, an explanation of the contract/certificate, an
incredibly brief section titled "Evidence" that provides some numbers but
no explanation of where they came from, and then 3 pages of analysis
explaining the supposed benefit of the approach. They all have appendixes
with (IMO) little useful information.

None of the papers I reviewed were any good, although that is obviously
just one reviewer's opinion.

I'm obviously annoyed because all the papers make more work for the
reviewers, but it is also frustrating because it seems like a poor use of
time. Even with AI, producing these papers took some thought and work.
Perhaps if the authors had focused that effort on a single paper/topic they
would have produced something more interesting.

Keith

On Tue, Jul 7, 2026 at 1:09 PM Bryan S. Kim <bkim01 at syr.edu> wrote:

> Dear Keith,
>
> Thank you for bringing this up, and we appreciate you going the extra mile
> to look through even the papers not assigned to you.
>
> Your suspicion about authorship is correct - they are all from the same
> authors. I cannot definitively say about AI usage, but if the papers'
> contents significantly overlap, there is an ethical case to be made about
> duplicate submission. The current CFP has some language about simultaneous
> submission to multiple venues (HS + other venue), but does not cover this
> specific case (similar papers to a single venue).
>
> Young-ri and I discussed this, and we will keep the papers submitted by
> these authors in mind. Should more than one of them be leaning toward
> acceptance (based on the assigned reviews), we will discuss this during the
> PC meeting over Zoom with all (if not most) of the involved reviewers for
> these papers.
>
> We will definitely keep the Steering Committee in the loop about this. You
> are right: the current ACM policy places the burden on the reviewers, and
> we, as members of the research community, need to be concerned about AI
> abuse cases that would otherwise degrade the quality of published research
> and waste reviewers' effort.
>
> Thank you again. Reviewers like you are what make our research community
> shine.
> -Bryan
>
>
> ------------------------------
> *From:* Hotstorage-chairs <hotstorage-chairs-bounces at fsl.cs.sunysb.edu>
> on behalf of Keith Smith via Hotstorage-chairs <
> hotstorage-chairs at fsl.cs.sunysb.edu>
> *Sent:* Monday, July 6, 2026 18:57
> *To:* chairs26 at hotstorage.org <chairs26 at hotstorage.org>
> *Subject:* [Hotstorage-chairs] HotStorage paper similarity
>
> Hi Young-ri, Bryan
>
> I just wanted to raise a potential concern about some of the submissions
> in case you haven't noticed or nobody else has brought it up.
>
> The last three papers that I reviewed -- papers 102
> <https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fhotstorage26.hotcrp.com%2fpaper%2f102&c=E,1,e6XcSBUh2lXx0YK7xzYXWOf5jvnoLgv4_enf1qUUTnniRhGQ48NSr0r5yfxM6O7E92fXfAVI6GhvbrVPFycvBV49giJV0ogqlvwPsW76ud38CEtdsaD4wBM,&typo=1>,
> 197
> <https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fhotstorage26.hotcrp.com%2fpaper%2f197&c=E,1,Ice3HQzEKOXOs_zEWcFWtqHCHMuIioeUVlpu6ITJkG9FHXYx8nUZkmQG-G5PB3oCBsBr0NNmWByOASD_YYrKGmV80Eizri1W2jcG14hMH_e7auWH4HuAu0ZBitA,&typo=1>,
> and 209
> <https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fhotstorage26.hotcrp.com%2fpaper%2f209&c=E,1,WlCyBqG6PmeFVNBh0hyrclBwnqxS7WDazVALqvL4Vco-mzOUe5dqrF60_dOim79QvLMT2U7e_bIJ3LYIwrWzQczfEdDe4_ZpYgvuJ0QZ0HQkdA,,&typo=1> --
> were strikingly similar. They all describe a similar concept applied to
> different areas. The papers have similar structure, similar language, and
> similar flaws. In addition to the papers I've reviewed, at least 4 other
> papers (based on a quick scan) seem to match the pattern (22
> <https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fhotstorage26.hotcrp.com%2fpaper%2f22&c=E,1,ZpulObutuYh9lsWZPK4XfQDFOMLzsAWwCyhYwYmy9C2wVrugWOphjMZ57zzbSuOf5bYPT-SPyf7spCImuCljctoPCdHj3OVbeEFGbfF9pDNc_tu7Ww,,&typo=1>,
> 40
> <https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fhotstorage26.hotcrp.com%2fpaper%2f40&c=E,1,wpe7fuwwN_cWaUpFpaJBIXSEk-VO0dfnI-KwyIQiImJ_SkGS5RWy8IyILGHxSSJhIoCuaLIjVeIElhGr5JVNDB4DYnpYgOwT2w8C56twfns8zGAvrFg,&typo=1>,
> 49
> <https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fhotstorage26.hotcrp.com%2fpaper%2f49&c=E,1,MWzCQxvJqcp6tRrlhdcU7rX1k2O5kSqQbb6gymqkIt8oBDXHyAmQcEyz5ImRzD3e7z_vRf5wvycXXFrL9OkgCqh_GBTAm12KbYWlP3xXCMX1Q3M,&typo=1>,
> and 117
> <https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fhotstorage26.hotcrp.com%2fpaper%2f117&c=E,1,ihjEoPGpc6SBxfHMXwy7W8dc1iiRj7zCpz62M9Qjddh5T43kgSpiPLYQ8B5MqQi3hRxboYEuF5SHvUi-LLIWYx8M14MO9aZQPUdlSbPyzZywznKnqubcjnoAS0Li&typo=1>).
> I can explain the similarities if you want.
>
> Given the strong similarities I suspect they are all by the same
> author(s). In this day and age, it seems likely they were largely AI
> generated, but that's hard to tell, and anyway the ACM author guidelines
> <https://www.acm.org/publications/policies/new-acm-policy-on-authorship> explicitly
> say authors can use AI to generate the text of their paper without
> disclosing that.
>
> So I don't have reason to believe that the author(s) here have violated
> the rules. And I'm not asking you all to do something about it (unless
> something here strikes you as illegitimate).
>
> Still I wanted to point this out, in case you haven't noticed this. And I
> would suggest that the Steering Committee might want to consider changes to
> the submission guidelines for future years. In particular, I'm concerned
> that it is too easy for authors to spam a workshop like HotStorage with AI
> generated submissions, since the ACM allows the undisclosed use of AI in
> writing papers, and a position paper does not need to have original
> research content (where AI use would have to be disclosed).
>
> Keith
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.fsl.cs.sunysb.edu/pipermail/hotstorage-chairs/attachments/20260707/7a73a422/attachment.htm>