• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
  • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Posts Accuse XM Paper of Copying Prior IMLE Work

    Side-by-side excerpts highlight similar mode forcing sections in XM and CHIMLE papers.

    SZ
    JH
    YD
    20 Sources, 60d ago, first seen 60d ago

    TLDR

    Susan Zhang posted excerpts comparing sections on mode forcing across papers and noted that the XM paper cited a lengthy preprint instead of the 2023 work that introduced the term. You Jiacheng replied that the XM paper's discussion of IMLE as a special case appears incorrect. Pasquale Minervini retweeted criticism that authors rebranded IMLE under a new name rather than crediting the original. Yilun Du replied that the draft refers to a different concept than the earlier mode forcing paper and that the team learned of IMLE late in the project. The posts treat the visible overlaps as the basis for the claims.

    Combined views

    476.8K

    20 Sources, first seen 60d ago

    Combined views

    476.8K

    20 Sources, first seen 60d ago

    2.1K likes
    2.1K likes
    60 comments
    729 saves
    54 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    60 comments
    729 saves
    54 reposts
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    20 Sources

    @YouJiachengRT @ZhengyangGeng: That's simple! 👏 XM: K guesses, one gradient. A clean step toward end-to-end generative modeling! Also reminds me of…
    @kastnerkyleRT @AlexiGlad: @YouJiacheng (copy pasting from other tweet) yep we note this in Sec E.3, IMLE is basically a specific instance of end-to-en…
    @cloneofsimoKinda crazy how calm and gentle he is reacting to this situation. Huge respect to this. If I were in his shoes i wouldve been absolutely furious ☠
    @suchenzangso... the explorative-modeling-third-pretraining-axis thing was effectively plagiarism? 👀
    @PMinerviniRT @bozavlado: They could have said, "We are revisiting IMLE, and it really works well at scale". Instead, they invent a new name for the…
    @du_yilun@suchenzang The above thread from @KL_Div mischaracterizes our paper, @AlexiGlad will clarify in a thread soon. We became aware of IMLE roughly 80% through project completion and did not see the follow up papers. We believe the papers are actually very different.
    @yuntiandengRT @thoma_gu: IMLE is always inspiring and insightful!
    @adityakusupatiRT @adityakusupati: @AlexiGlad @NepalAadim I like Alexi's work in general, and I am not an expert in this line of work. But having a pape…
    @jbhuang0604ML twitter this week

    20 Sources

    @YouJiachengRT @ZhengyangGeng: That's simple! 👏 XM: K guesses, one gradient. A clean step toward end-to-end generative modeling! Also reminds me of…
    @kastnerkyleRT @AlexiGlad: @YouJiacheng (copy pasting from other tweet) yep we note this in Sec E.3, IMLE is basically a specific instance of end-to-en…
    @cloneofsimoKinda crazy how calm and gentle he is reacting to this situation. Huge respect to this. If I were in his shoes i wouldve been absolutely furious ☠
    @suchenzangso... the explorative-modeling-third-pretraining-axis thing was effectively plagiarism? 👀
    @PMinerviniRT @bozavlado: They could have said, "We are revisiting IMLE, and it really works well at scale". Instead, they invent a new name for the…
    @du_yilun@suchenzang The above thread from @KL_Div mischaracterizes our paper, @AlexiGlad will clarify in a thread soon. We became aware of IMLE roughly 80% through project completion and did not see the follow up papers. We believe the papers are actually very different.
    @yuntiandengRT @thoma_gu: IMLE is always inspiring and insightful!
    @adityakusupatiRT @adityakusupati: @AlexiGlad @NepalAadim I like Alexi's work in general, and I am not an expert in this line of work. But having a pape…
    @jbhuang0604ML twitter this week