#Tag · bonfire.cafe

"Cornell Tech has received more than $7 million from Schmidt Sciences and NASA to upgrade arXiv, an open-access research repository of more than 2.8 million articles. "

It's great news for arXiv. I'm just wondering why they want actually move to the cloud:

"finish migrating to cloud infrastructure"

NREN are providing access for educational networks, I hope the cloud mentioned there is operated by a NREN operator.

🔗 https://news.cornell.edu/stories/2025/11/7m-grant-nasa-schmidt-sciences-upgrade-arxiv

#arxiv #opendata #openscience #nren #science

Alexandre Dulaunoy

@a@paperbay.org · 5 days ago

"Cornell Tech has received more than $7 million from Schmidt Sciences and NASA to upgrade arXiv, an open-access research repository of more than 2.8 million articles. "

It's great news for arXiv. I'm just wondering why they want actually move to the cloud:

"finish migrating to cloud infrastructure"

NREN are providing access for educational networks, I hope the cloud mentioned there is operated by a NREN operator.

🔗 https://news.cornell.edu/stories/2025/11/7m-grant-nasa-schmidt-sciences-upgrade-arxiv

#arxiv #opendata #openscience #nren #science

petersuber

@petersuber@fediscience.org · 3 weeks ago

"Databricks co-founder argues US must go #OpenSource to beat China in AI."
https://techcrunch.com/2025/11/14/databricks-co-founder-argues-us-must-go-open-source-to-beat-china-in-ai/

"Major #AI labs, including OpenAI, Meta, and Anthropic, continue to innovate significantly, yet their innovations remain largely proprietary… #Konwinski argued that for ideas to truly flourish, they need to be freely exchanged and discussed with the larger academic community. He pointed out that generative AI emerged as a direct result of the #Transformer architecture, a pivotal training technique introduced in a freely available research paper [ #OpenAccess in #arXiv]."

Hacker News

@h4ckernews@mastodon.social · 3 weeks ago

LeJEPA

https://arxiv.org/abs/2511.08544

#HackerNews #LeJEPA #LeJEPA2023 #arXiv #Research #Technology #Innovation

Davide Eynard (+mala) boosted

Fabrizio Musacchio

@pixeltracker@sigmoid.social · 4 weeks ago

#arXiv will no longer accept #CcomputerScience review papers unless they’re already accepted elsewhere. The reason: too many low-quality and #AI-generated submissions.

🌍 https://www.nature.com/articles/d41586-025-03664-7

The arXiv repository will no longer accept review or position papers in computer science unless they have already been peer reviewed.Credit: Sharaf Maksumov/Alamy

Fabrizio Musacchio

@pixeltracker@sigmoid.social · 4 weeks ago

#arXiv will no longer accept #CcomputerScience review papers unless they’re already accepted elsewhere. The reason: too many low-quality and #AI-generated submissions.

🌍 https://www.nature.com/articles/d41586-025-03664-7

Hacker News

@h4ckernews@mastodon.social · 4 weeks ago

The Principles of Diffusion Models

https://arxiv.org/abs/2510.21890

#HackerNews #DiffusionModels #Principles #AI #Research #MachineLearning #Arxiv

arXiv.org

The Principles of Diffusion Models

This monograph presents the core principles that have guided the development of diffusion models, tracing their origins and showing how diverse formulations arise from shared mathematical ideas. Diffusion modeling starts by defining a forward process that gradually corrupts data into noise, linking the data distribution to a simple prior through a continuum of intermediate distributions. The goal is to learn a reverse process that transforms noise back into data while recovering the same intermediates. We describe three complementary views. The variational view, inspired by variational autoencoders, sees diffusion as learning to remove noise step by step. The score-based view, rooted in energy-based modeling, learns the gradient of the evolving data distribution, indicating how to nudge samples toward more likely regions. The flow-based view, related to normalizing flows, treats generation as following a smooth path that moves samples from noise to data under a learned velocity field. These perspectives share a common backbone: a time-dependent velocity field whose flow transports a simple prior to the data. Sampling then amounts to solving a differential equation that evolves noise into data along a continuous trajectory. On this foundation, the monograph discusses guidance for controllable generation, efficient numerical solvers, and diffusion-motivated flow-map models that learn direct mappings between arbitrary times. It provides a conceptual and mathematically grounded understanding of diffusion models for readers with basic deep-learning knowledge.

theruran 💻 🌐 :cereal_killer: and 1 other boosted

Hendrik Weimer

@hweimer@fediscience.org · last month

Given the drama surrounding the arXiv for its dependence on Google and its recent ban on review articles in computer science, I have a question:

Do we still need the arXiv?

Wouldn't a federated network of institutional repositories be a much better option?

#arXiv #academicchatter #physics #math #cs #ActivityPub

petersuber

@petersuber@fediscience.org · 2 months ago

Update. In response to this problem (previous post, this thread), some publishers are desk-rejecting papers based on open health datasets. The problem is not the quality of the data, but the absence of additional work to validate findings.

Two reports:

1. "Journals and publishers crack down on research from open health data sets," Science, Oct 8, 2025.
https://www.science.org/content/article/journals-and-publishers-crack-down-research-open-health-data-sets

2. "AI: Journals are automatically rejecting public health dataset papers to combat paper mills," BMJ, Oct 15, 2025.
https://www.bmj.com/content/391/bmj.r2170
( #paywalled)

#AI #Data #Frontiers #Medicine #PaperMills #PLOS #OpenData

petersuber

@petersuber@fediscience.org replied · last month

Update. Here's how #arXiv is dealing with a similar problem in computer science.
https://blog.arxiv.org/2025/10/31/attention-authors-updated-practice-for-review-articles-and-position-papers-in-arxiv-cs-category/

"Before being considered for submission to arXiv’s #CS category, review articles and position papers must now be accepted at a journal or a conference and complete successful peer review…In the past few years, arXiv has been flooded with papers. Generative #AI / #LLMs have added to this flood by making papers – especially papers not introducing new research results – fast and easy to write. While categories across arXiv have all seen a major increase in submissions, it’s particularly pronounced in arXiv’s CS category."

#PeerReview #Postprints #Preprints #ScholComm

https://blog.arxiv.org

Attention Authors: Updated Practice for Review Articles and Position Papers in arXiv CS Category

Florian 'floe' Echtler boosted

Sebastian Büttrich @ ITU

@sebastian@social.itu.dk · last month

#AISlop killing open submission -

https://blog.arxiv.org/2025/10/31/attention-authors-updated-practice-for-review-articles-and-position-papers-in-arxiv-cs-category/

"arXiv’s computer science (CS) category has updated its moderation practice with respect to review (or survey) articles and position papers. Before being considered for submission to arXiv’s CS category, review articles and position papers must now be accepted at a journal or a conference and complete successful peer review."

#AI #thereIsNoAI #arXiv

https://blog.arxiv.org

Attention Authors: Updated Practice for Review Articles and Position Papers in arXiv CS Category

Sebastian Büttrich @ ITU

@sebastian@social.itu.dk · last month

#AISlop killing open submission -

https://blog.arxiv.org/2025/10/31/attention-authors-updated-practice-for-review-articles-and-position-papers-in-arxiv-cs-category/

#AI #thereIsNoAI #arXiv

https://blog.arxiv.org

Attention Authors: Updated Practice for Review Articles and Position Papers in arXiv CS Category

Hendrik Weimer

@hweimer@fediscience.org · last month

Given the drama surrounding the arXiv for its dependence on Google and its recent ban on review articles in computer science, I have a question:

Do we still need the arXiv?

Wouldn't a federated network of institutional repositories be a much better option?

#arXiv #academicchatter #physics #math #cs #ActivityPub

Wladimir Mufty

@wlaatje@social.edu.nl · last month

Why is the #arXiv CS category making this change?

“In the past few years, arXiv has been flooded with papers. Generative AI / large language models have added to this flood by making papers – especially papers not introducing new research results – fast and easy to write.”

https://blog.arxiv.org/2025/10/31/attention-authors-updated-practice-for-review-articles-and-position-papers-in-arxiv-cs-category/

Hacker News

@h4ckernews@mastodon.social · last month

arXiv No Longer Accepts Computer Science Position or Review Papers Due to LLMs

https://blog.arxiv.org/2025/10/31/attention-authors-updated-practice-for-review-articles-and-position-papers-in-arxiv-cs-category/

#HackerNews #arXiv #LLMs #ComputerScience #ResearchUpdates #AcademicPublishing

Hacker News

@h4ckernews@mastodon.social · last month

Beyond Smoothed Analysis: Analyzing the Simplex Method by the Book

https://arxiv.org/abs/2510.21613

#HackerNews #BeyondSmoothedAnalysis #SimplexMethod #AlgorithmResearch #Arxiv #Mathematics

arXiv.org

Beyond Smoothed Analysis: Analyzing the Simplex Method by the Book

Narrowing the gap between theory and practice is a longstanding goal of the algorithm analysis community. To further progress our understanding of how algorithms work in practice, we propose a new algorithm analysis framework that we call by the book analysis. In contrast to earlier frameworks, by the book analysis not only models an algorithm's input data, but also the algorithm itself. Results from by the book analysis are meant to correspond well with established knowledge of an algorithm's practical behavior, as they are meant to be grounded in observations from implementations, input modeling best practices, and measurements on practical benchmark instances. We apply our framework to the simplex method, an algorithm which is beloved for its excellent performance in practice and notorious for its high running time under worst-case analysis. The simplex method similarly showcased the state of the art framework smoothed analysis (Spielman and Teng, STOC'01). We explain how our framework overcomes several weaknesses of smoothed analysis and we prove that under input scaling assumptions, feasibility tolerances and other design principles used by simplex method implementations, the simplex method indeed attains a polynomial running time.

Chris Walter boosted

j_bertolotti

@j_bertolotti@mathstodon.xyz · 4 months ago

New paper on #ArXiv!
When (coherent) light propagates in a multimode fibre, each mode accumulates a slightly different phase, so the output looks like a random speckle patter.
But it is NOT random!
This non-randomness has a number of consequences, among them the fact that the you get predictable patterns in the apparently random output. We show that a focused input will always result in a ring of excess intensity at the same radius of the input (and suggest how this might be useful).

#Optics #OpticalFibre