conv.

All stories
ScienceQuiet 45h · day 8

TMLR editor tests desk-rejected authors on their own papers, most fail

Nihar B. Shah, editor-in-chief of Transactions on Machine Learning Research, interviewed authors of ten rejected papers to see if they understood their own work.

What to know

  • TMLR editor-in-chief Nihar B. Shah interviewed authors of ten already desk-rejected papers to test whether they understood their own technical content.
  • Results: 3 authors dodged the interview, 3 could not answer basic questions, 3 answered basic questions but struggled with technical ones, and only 1 succeeded.
  • Shah frames the exercise as relevant to concerns about AI-generated or heavily AI-assisted submissions and author accountability for correctness and integrity.
  • The post has spread across Mastodon and Hacker News, drawing reactions ranging from praise for the accountability check to alarm at how poorly most authors performed.
many voices

Shah's interviews are a needed and admirable accountability check against AI-generated or AI-assisted papers.

  • “What a hero: Nihar B Shah interviewed authors of papers that would have been desk-rejected to see if they understood their own papers.”

    ionica@mathstodon.xyz · Mastodon ↗
some voices

The finding that authors can't explain their own papers is genuinely alarming for the integrity of academic publishing.

some voices

The breakdown of results shows publication itself is a weaker signal of genuine researcher contribution than assumed.

  • “The value of publication as a signal of researcher contribution is much weaker when authors cannot explain or defend their paper.”

    marick@mstdn.social · Mastodon ↗

Nihar B. Shah Editor-in-chief, Transactions on Machine Learning Research (TMLR)

How it unfolded 5 developments, newest first · click a bar or a number to jump posts

Peak 8 pieces in two hours at Sep 19, 10 AM; 30 pieces over 8 days (1 article · 8 posts · 21 comments) Sep 15, 8 PM — 1 piece · 1 article — Mastodon 1Sep 15, 10 PM — quietSep 16, 12 AM — quietSep 16, 2 AM — quietSep 16, 4 AM — quietSep 16, 6 AM — quietSep 16, 8 AM — quietSep 16, 10 AM — quietSep 16, 12 PM — quietSep 16, 2 PM — quietSep 16, 4 PM — quietSep 16, 6 PM — 1 piece · 1 post — Hacker News 1Sep 16, 8 PM — quietSep 16, 10 PM — quietSep 17, 12 AM — quietSep 17, 2 AM — 1 piece · 1 post — Hacker News 1Sep 17, 4 AM — quietSep 17, 6 AM — quietSep 17, 8 AM — 1 piece · 1 post — Mastodon 1Sep 17, 10 AM — quietSep 17, 12 PM — quietSep 17, 2 PM — 1 piece · 1 post — Mastodon 1Sep 17, 4 PM — 1 piece · 1 post — Mastodon 1Sep 17, 6 PM — quietSep 17, 8 PM — quietSep 17, 10 PM — quietSep 18, 12 AM — quietSep 18, 2 AM — 1 piece · 1 post — Hacker News 1Sep 18, 4 AM — quietSep 18, 6 AM — quietSep 18, 8 AM — quietSep 18, 10 AM — quietSep 18, 12 PM — quietSep 18, 2 PM — quietSep 18, 4 PM — quietSep 18, 6 PM — quietSep 18, 8 PM — quietSep 18, 10 PM — quietSep 19, 12 AM — quietSep 19, 2 AM — quietSep 19, 4 AM — quietSep 19, 6 AM — quietSep 19, 8 AM — 1 piece · 1 comment — Hacker News 1Sep 19, 10 AM — 8 pieces · 8 comments — Hacker News 8Sep 19, 12 PM — 3 pieces · 3 comments — Hacker News 3Sep 19, 2 PM — 1 piece · 1 post — Mastodon 1Sep 19, 4 PM — quietSep 19, 6 PM — quietSep 19, 8 PM — quietSep 19, 10 PM — quietSep 20, 12 AM — 2 pieces · 2 comments — Hacker News 2Sep 20, 2 AM — 1 piece · 1 comment — Hacker News 1Sep 20, 4 AM — 5 pieces · 5 comments — Hacker News 5Sep 20, 6 AM — quietSep 20, 8 AM — 1 piece · 1 comment — Hacker News 1Sep 20, 10 AM — quietSep 20, 12 PM — quietSep 20, 2 PM — quietSep 20, 4 PM — quietSep 20, 6 PM — quietSep 20, 8 PM — quietSep 20, 10 PM — quietSep 21, 12 AM — quietSep 21, 2 AM — quietSep 21, 4 AM — quietSep 21, 6 AM — quietSep 21, 8 AM — quietSep 21, 10 AM — quietSep 21, 12 PM — quietSep 21, 2 PM — quietSep 21, 4 PM — quietSep 21, 6 PM — quietSep 21, 8 PM — quietSep 21, 10 PM — quietSep 22, 12 AM — quietSep 22, 2 AM — quietSep 22, 4 AM — 1 piece · 1 post — Mastodon 1Sep 22, 6 AM — quietSep 22, 8 AM — quietSep 22, 10 AM — quietSep 22, 12 PM — quietSep 22, 2 PM — quietSep 22, 4 PM — quietSep 22, 6 PM — quietSep 22, 8 PM — quietSep 22, 10 PM — quietYesterday, 12 AM — quietYesterday, 2 AM — quietYesterday, 4 AM — quietYesterday, 6 AM — quietYesterday, 8 AM — quietYesterday, 10 AM — quietYesterday, 12 PM — quietYesterday, 2 PM — quietYesterday, 4 PM — quietYesterday, 6 PM — quietYesterday, 8 PM — quietYesterday, 10 PM — quietToday, 12 AM — quiet 123–45
Sep 16Sep 17Sep 18Sep 19Sep 20Sep 21Sep 22now · 2:55 AM ET
  1. 5

    Hacker News discussion of the TMLR post continues

    The Medium post circulates further on Hacker News under the title "Asking Authors About Their Own Papers (TMLR)", with additional submissions and a small comment thread.

    • This is the journal's policy on LLM use by authors [1]:> LLMs may be used as general-purpose assistive tools. Whichever tools are used, authors are fully responsible for content on which they are listed as (co-) authors. This includes, but is not limited to, content generated by LLMs that could be construed as plagiarism or scientific misconduct…

      WCSTombsHacker News4d agoview on Hacker News ↗
    2 more of the top 3 · 22 posts in this stretch
    • telescoper@mastodon.social

      What happens if an Editor asks an author about their use of AI? https:// medium.com/@TmlrOrg/asking-aut hors-about-their-own-papers-3d2e04e5dee0

      telescoper@mastodon.socialMastodon1d agoview on Mastodon ↗
    • I answered requests to be a peer reviewer. (I'm not sure why I was selected, I don't have many publications or credentials.) I saw a lot of papers with hallucinated references. I also remember one paper that described a methodology that I don't think the authors really performed, I think it was just academic fraud where they pretended to have…

      logicalleeHacker News4d agoview on Hacker News ↗
    all of them →
  2. 4

    Hachyderm user reacts with alarm to the article's rationale

    A hachyderm.io post quotes the article's justification—linking author accountability requirements to concerns about AI-generated submissions—and reacts with dismay.

    “jfc…”
    — inthehands@hachyderm.io
    • ❝Most journals and conferences require authors to ensure correctness and integrity of their papers and take responsibility for them. This is particularly relevant given potential AI generated or heavily AI-assisted submissions. When authors could not answer questions about the technical parts, and sometimes even basic questions about the paper, it…

      inthehands@hachyderm.ioMastodon6d ago77▲view on Mastodon ↗
  3. 3

    Mathstodon user hails Shah's interviews under #FuckAI

    A mathstodon.xyz post praises Shah for interviewing authors of already desk-rejected papers to check their comprehension, tagging the post #FuckAI.

    “What a hero: Nihar B Shah interviewed authors of papers that would have been desk-rejected to see if they understood their own papers.”
    — ionica@mathstodon.xyz
    • ionica@mathstodon.xyz

      What a hero: Nihar B Shah interviewed authors of papers that would have been desk-rejected to see if they understood their own papers. https:// medium.com/@TmlrOrg/asking-aut hors-about-their-own-papers-3d2e04e5dee0 # FuckAI

      ionica@mathstodon.xyzMastodon6d ago73▲view on Mastodon ↗
  4. 2

    Commenter summarizes the interview outcomes

    A Mastodon post breaks down the results of Shah's ten interviews: three authors dodged the interview, three could not answer basic questions, three answered basic questions but struggled with technical ones, and one succeeded.

    “The value of publication as a signal of researcher contribution is much weaker when authors cannot explain or defend their paper.”
    — marick@mstdn.social (quoting the article)
    • marick@mstdn.social

      Editor-in-chief of Transactions on Machine Learning Research wanted to interview authors of ten (already desk-rejected) papers about the content of their own paper. The expected ensues. 3 dodged the interview. 3 could not answer basic questions. 3 could answer basic questions, but struggled with technical ones. 1 success "The value of publication…

      marick@mstdn.socialMastodon6d agoview on Mastodon ↗
  5. 1

    Nihar B. Shah publishes account of author interviews

    TMLR editor-in-chief Nihar B. Shah posts "Asking Authors About Their Own Papers" on Medium, describing how he interviewed authors of ten desk-rejected papers to test whether they understood the technical content of their own work, citing concern over AI-generated or heavily AI-assisted submissions.

    1. first by Mastodon, 8d ago

What people are saying 18 voices from 1 site · best of 25 · verbatim