Tootfinder

MAGA's outrage over Bad Bunny could be good for Democrats if they can learn the right lesson - LGBTQ Nation
https://www.lgbtqnation.com/2025/10/magas-outrage-over-bad-bunny-could-be-good-for-democrats-if-they-can-learn-the-right-lesson/

Surgical Knowledge Rewrite in Compact LLMs: An 'Unlearn-then-Learn' Strategy with ($IA^3$) for Localized Factual Modulation and Catastrophic Forgetting Mitigation
Stanley Ngugi
https://arxiv.org/abs/2508.07075

Surgical Knowledge Rewrite in Compact LLMs: An 'Unlearn-then-Learn' Strategy with ($IA^3$) for Localized Factual Modulation and Catastrophic Forgetting Mitigation
Large Language Models (LLMs) struggle with dynamic knowledge updates, especially when new information conflicts with deeply embedded facts. Such conflicting factual edits often lead to two critical issues: resistance to adopting the new fact and severe catastrophic forgetting of unrelated knowledge. This paper introduces and evaluates a novel "unlearn-then-learn" strategy for precise knowledge editing in LLMs, leveraging the parameter-efficient fine-tuning (PEFT) technique, Infused Adapter by I…

Ein Podcast von Senioren:innen für Senior:innen
#Podcast #Senioren #Alter
…

On Condensation Management - Learn how you can best avoid problems with condensation, damp and mould in winter. #mould #condensation #ventilation -

On Condensation Management
Learn how you can best avoid problems with condensation, damp and mould in winter. #mould #condensation #ventilation

For those interested in local and self-hosting, gonna be running our monthly #libre/#FOSS catchup at https://meeting.iridescent.nz tonight at 20:00 NZST talking about migrating our *many* web services built with Doc…

Machines Learn Number Fields, But How? The Case of Galois Groups
Kyu-Hwan Lee, Seewoo Lee
https://arxiv.org/abs/2508.06670 https://arxiv.org/pdf/2508.06670…

Machines Learn Number Fields, But How? The Case of Galois Groups
By applying interpretable machine learning methods such as decision trees, we study how simple models can classify the Galois groups of Galois extensions over $\mathbb{Q}$ of degrees 4, 6, 8, 9, and 10, using Dedekind zeta coefficients. Our interpretation of the machine learning results allows us to understand how the distribution of zeta coefficients depends on the Galois group, and to prove new criteria for classifying the Galois groups of these extensions. Combined with previous results, thi…

Are you worried your #dotnet #security could be more secure? Join us for a #livestream on August 21st, 2025, to discuss FAPI 2.0, its relation to

What Can I Do Right Now to Defend Democracy and Fight Fascism?
https://www.peoplefor.org/what-can-i-do-right-now-defend-democracy-and-fight-fascism

What Can I Do Right Now to Defend Democracy and Fight Fascism?
Learn about ways you can take action right now to combat the spread of fascism in the United States.

I feel stupid for asking this… but is there still no good way to checkout and later pull a pull request branch from another repo?
We've got `gh pr checkout NUMBER` but the only way I've found to fetch updates is to delete the local branch and start over.
I really don't want to play adlib with manually adding remotes etc – is this still the best we can do in 2025??
EDIT: it’s solved! 🎉

Tomáš Janoušek (@liskin@genserver.social)
@hynek @carlton Drop fetch.all=true and that fixes it.You may need to learn to use git remote update instead of git fetch though.And I’d consider filing this as a bug upstream, because as you said ...

Base Models Know How to Reason, Thinking Models Learn When
Constantin Venhoff, Iv\'an Arcuschin, Philip Torr, Arthur Conmy, Neel Nanda
https://arxiv.org/abs/2510.07364 https…

Base Models Know How to Reason, Thinking Models Learn When
Why do thinking language models like DeepSeek R1 outperform their base counterparts? Despite consistent performance gains, it remains unclear to what extent thinking models learn entirely new reasoning capabilities or repurpose pre-existing base model ones. In this work, we propose a hybrid model where we activate reasoning mechanisms in base models at the right time to elicit thinking-model-level reasoning chains, implying that thinking models exploit already existing capabilities. To ground o…

https://fightchatcontrol.eu/
This seems useful.

I do wonder if an aspiring Cybersecurity Associate should know enough about opsec to not leave a permanent public record of them asking how to cram knowledge of entire of DNS in three days for a job interview, especially when they have chosen to have the world's most searchable identity.
https://list…

Great way to learn history.
#History #Education

Past Loves - A History Of The Greatest Love Stories
Podcast · Holly · Join our host Holly, with a range of guests from the world of heritage and academia, as we take a trip into the most gut-wrenching and heart-warming depths of history to explore affection, infatuation and attachment across time. As we ask, what is the best love story of all time?

A great advantage that speakers of English as a second language have, is the fact that all the little idioms and phrases and metaphors get explained to us and we can just learn them out of a book 😁

Fun podcast, but I love reading A Way with Words at #WordOrigins #Language
...Bab…

Mox Nix (episode #1662)
You might be surprised to learn that a “hoosier” isn’t necessarily from Indiana. Around the St. Louis, Missouri, area, the term hoosier has a whole other meaning. And: Scotland is the home of the…

Rail service is in Canada is beyond half-ass. When booking a ticket between Toronto and Montreal - the country's two largest cities, their urban areas representing something like a third of Canada's population - you learn that, "We are currently experiencing delays of 15-45 minutes on certain Corridor trains due to unexpected speed restrictions imposed by CN, the railway infrastructure owner." It's almost as if government after government after government couldn't b…

How to report setlists to PROs after shows and why you should
#MusicianTips #LiveMusic

How to report setlists to PROs after shows and why you should
Learn how to report setlists to PROs and maximize your earnings as a musician in the complex world of music rights.

I mean sure, "AI" might make that somehow easier to pull off, but "politicians holding signs that could be manipulated" has been the source of hilarity since Photoshop v1 released.
And every generation of politicians and political campaign managers seems to have to re-learn this lesson. 😅
«Swedish Prime Minister Pulls AI Campaign Tool After It Was Used to Ask Hitler for Support»
https://www.404media.co/swedish-prime-minister-pulls-ai-campaign-tool-after-it-was-used-to-ask-hitler-for-support/

Discover the horrifying OKC-Spatz case, spanning over 27 years of legal battles among 3 EU countries! Your donations can truly impact the victims finish the case in France. Learn more at: https://chardonbleus.org/donate

So now we have a patient registry for a Neuralink Blindsight brain implant at https://neuralink.com/trials/visual-prosthesis/ and a patient registry for a Science Corporation retinal implant at

Visual Prosthesis | Neuralink
Learn more about our future visual prosthesis clinical trials.

On Spacetherm Aerogel Thermal Insulation - Learn about the highest-performing robust wall insulation for where space is at a premium: small homes in expensive cities. #aerogel #IWI #insulation -

On Spacetherm Aerogel Thermal Insulation
Learn about the highest-performing robust wall insulation for where space is at a premium: small homes in expensive cities. #aerogel #IWI #insulation

Challenge: "Name 20 living female authors you admire, 1 per day"
Day 18: Kim Småge
Sorry, but you'll have to learn some Norwegian if you want to read this author, as she hasn't been translated into English yet - but that means you could read Ibsen too...
"Containerkvinnan" was a real knockout. The bodies of two women are found in a container in Trondheim's harbour, and from here proceeds the whodunnit. It's a whole series and I keep mean…

openAI's "Why language models hallucinate" is a pretty general argument: any model trained on pass/fail tests will learn to guess/bullshit, which is sometimes right, vs saying "IDK", which is always wrong.
I think this can be generalized beyond AI: humans sometimes have that problem. market/competitive systems too. it's sort of an offshoot of Goodhart's law.
the paper doesn't make that leap. its conclusion is "we should train models with t…

Should I switch my SMTP server to Mailcow? Kinda want to learn that software but sounds like it uses a lot more resources than my regular dovecot/postfix server.

Correo enviado a todos los europarlamentarios contando argumentos contra #chatcontrol mediante https://fightchatcontrol.eu/#contact-tool

End of an era. RIP Guppe :(

From David Suzuki
Want climate action? You’re part of a silenced majority
Most people worldwide want action on climate change. A significant majority would even be willing to give up some of their income to help. You wouldn’t learn that from most mainstream media or politicians.

Questions for those people who use @… - and with RSS or without: how do you handle UPDATES in a blog post?
like: when you add something in an old post - is there a way of doing it, so it's on top again? just changing the date? also that it's correctly in RSS?
happy to learn about how you do it

LLMs Learn to Deceive Unintentionally: Emergent Misalignment in Dishonesty from Misaligned Samples to Biased Human-AI Interactions
XuHao Hu, Peng Wang, Xiaoya Lu, Dongrui Liu, Xuanjing Huang, Jing Shao
https://arxiv.org/abs/2510.08211

LLMs Learn to Deceive Unintentionally: Emergent Misalignment in Dishonesty from Misaligned Samples to Biased Human-AI Interactions
Previous research has shown that LLMs finetuned on malicious or incorrect completions within narrow domains (e.g., insecure code or incorrect medical advice) can become broadly misaligned to exhibit harmful behaviors, which is called emergent misalignment. In this work, we investigate whether this phenomenon can extend beyond safety behaviors to a broader spectrum of dishonesty and deception under high-stakes scenarios (e.g., lying under pressure and deceptive behavior). To explore this, we fin…

"We think of generating source code from a prompt as an AI-powered feature of modern IDEs, but the general problem has a rich history in research efforts and domain-specific programming systems." At Berlin Buzzwords 2025, William Benton discussed the history of program synthesis, its relationship to the history of AI, and the lessons we can learn from it today.
Watch the full session: #bbuzz 2026: https://tickets.plainschwarz.com/bbuzz26/c/8Hvk0ZvJA/

Ugly mushroom for an ugly week in US politics. We can learn from fungi and mushrooms; they recycle and renew. Rot and destruction create opportunities, for better or worse. We can rebuild what was lost better than it was before. Or we can let the people that destroyed the system build a worse one in its place. It's our choice. #fungifriday

#Marfan conference in Atlanta. It's been a busy summer!
Probably a…

Do buy a Baofeng and learn how to use it. Don't get your ham radio ticket. Nobody you'd want to talk to is on there, and if you're prepping for SHTF, the FCC is not going to be the last surviving arm of the government. Just keep your finger off the TX switch until then.

I really need to learn how to paint darker skin tones better...

ResAD: Normalized Residual Trajectory Modeling for End-to-End Autonomous Driving
Zhiyu Zheng, Shaoyu Chen, Haoran Yin, Xinbang Zhang, Jialv Zou, Xinggang Wang, Qian Zhang, Lefei Zhang
https://arxiv.org/abs/2510.08562

ResAD: Normalized Residual Trajectory Modeling for End-to-End Autonomous Driving
End-to-end autonomous driving (E2EAD) systems, which learn to predict future trajectories directly from sensor data, are fundamentally challenged by the inherent spatio-temporal imbalance of trajectory data. This imbalance creates a significant optimization burden, causing models to learn spurious correlations instead of causal inference, while also prioritizing uncertain, distant predictions, thereby compromising immediate safety. To address these issues, we propose ResAD, a novel Normalized R…

It's the end of #BannedBooksWeek, but not the end of book bans. Local involvement with your library is the best way to fight back and defend everyone's access to knowledge. Check out For The People's excellent new Libraries 101 video to learn how!

Towards Balanced Behavior Cloning from Imbalanced Datasets
Sagar Parekh, Heramb Nemlekar, Dylan P. Losey
https://arxiv.org/abs/2508.06319 https://arxiv.org…

Towards Balanced Behavior Cloning from Imbalanced Datasets
Robots should be able to learn complex behaviors from human demonstrations. In practice, these human-provided datasets are inevitably imbalanced: i.e., the human demonstrates some subtasks more frequently than others. State-of-the-art methods default to treating each element of the human's dataset as equally important. So if -- for instance -- the majority of the human's data focuses on reaching a goal, and only a few state-action pairs move to avoid an obstacle, the learning algorithm will pla…

How to Watch Panthers vs. Cowboys: Time, TV Channel and Live Stream https://www.foxsports.com/articles/nfl/how-to-watch-panthers-vs-cowboys-time-tv-channel-and-live-stream-2025-week-6

On Spacetherm Aerogel Thermal Insulation - Learn about the highest-performing robust wall insulation for where space is at a premium: small homes in expensive cities. #aerogel #IWI #insulation -

On Spacetherm Aerogel Thermal Insulation
Learn about the highest-performing robust wall insulation for where space is at a premium: small homes in expensive cities. #aerogel #IWI #insulation

SBS Learn English
Ease into the English language and Australian culture...
Great Australian Pods Podcast Directory: #GreatAusPods

What Builds Effective In-Context Examples for Code Generation?
Dongze Li, Songqiang Chen, Jialun Cao, Shing-Chi Cheung
https://arxiv.org/abs/2508.06414 https://

What Builds Effective In-Context Examples for Code Generation?
In-Context Learning (ICL) has emerged as a promising solution to enhance the code generation capabilities of Large Language Models (LLMs), which incorporates code examples inside the prompt to let LLMs learn from demonstrations. However, despite the substantial effectiveness of the code example-based ICL approach, the specific features (e.g., identifier naming styles, code formatting, solution insight) within the ICL-provided code examples that significantly contribute to the ICL's effectivenes…

Why learn good time management, when you can achieve the same with abject self-exploitation? #academicChatter

How to Watch Raiders vs. Titans: Time, TV Channel and Live Stream https://www.foxsports.com/articles/nfl/how-to-watch-raiders-vs-titans-time-tv-channel-and-live-stream-2025-week-6

Generative quantum advantage for classical and quantum problems
Hsin-Yuan Huang, Michael Broughton, Norhan Eassa, Hartmut Neven, Ryan Babbush, Jarrod R. McClean
https://arxiv.org/abs/2509.09033

Generative quantum advantage for classical and quantum problems
Recent breakthroughs in generative machine learning, powered by massive computational resources, have demonstrated unprecedented human-like capabilities. While beyond-classical quantum experiments can generate samples from classically intractable distributions, their complexity has thwarted all efforts toward efficient learning. This challenge has hindered demonstrations of generative quantum advantage: the ability of quantum computers to learn and generate desired outputs substantially better …

Attend online events and trainings 
If you can’t make it out to an
in-person event,
you can find a wide variety of events, trainings, and more online.
Visit the Mobilize feed to see what events are happening and learn how you can join in. 
https://www.mobilize.us/peoplefor/

Volunteer Opportunities, Events, and Petitions Near Me · People For the American Way on Mobilize
Find events, petitions, volunteer opportunities, fundraisers and more with People For the American Way.

Automatic Semantic Alignment of Flow Pattern Representations for Exploration with Large Language Models
Weihan Zhang, Jun Tao
https://arxiv.org/abs/2508.06300 https://

Automatic Semantic Alignment of Flow Pattern Representations for Exploration with Large Language Models
Explorative flow visualization allows domain experts to analyze complex flow structures by interactively investigating flow patterns. However, traditional visual interfaces often rely on specialized graphical representations and interactions, which require additional effort to learn and use. Natural language interaction offers a more intuitive alternative, but teaching machines to recognize diverse scientific concepts and extract corresponding structures from flow data poses a significant chall…

Really enjoyed @…’s thoughtful and entertaining talk at #smashingconf today! Amazing to see how philosophical ideas influenced the design of iA’s apps and what designers can and could learn from philosophy – and vice versa. A lot of food for thought!…

Trump's Abuses of Power Show We Didn't Learn Much From the Nixon Era (Jill Lawrence/The Bulwark)
https://www.thebulwark.com/p/trump-nixon-abuses-of-power-becoming-katharine-graham-review
http://www.memeorandum.com/251010/p33#a251010p33

Please do your part! This page makes it easy, just fill in the details, and send the email. I did it, so should you! This same shit is happening all over again. Wtf?
#ChatControl

Understanding Ice Crystal Habit Diversity with Self-Supervised Learning
Joseph Ko, Hariprasath Govindarajan, Fredrik Lindsten, Vanessa Przybylo, Kara Sulia, Marcus van Lier-Walqui, Kara Lamb
https://arxiv.org/abs/2509.07688

Understanding Ice Crystal Habit Diversity with Self-Supervised Learning
Ice-containing clouds strongly impact climate, but they are hard to model due to ice crystal habit (i.e., shape) diversity. We use self-supervised learning (SSL) to learn latent representations of crystals from ice crystal imagery. By pre-training a vision transformer with many cloud particle images, we learn robust representations of crystal morphology, which can be used for various science-driven tasks. Our key contributions include (1) validating that our SSL approach can be used to learn me…

"I think the root of homelessness is greed... England has a population of 60 million people and about the same homeless population as San Francisco, which is a city of about 850,000 people."
Kevin's book is in my queue. May move it up. He's speaking at the Roxie on Sunday!

For longtime Chronicle homelessness reporter Kevin Fagan, the business was always personal
Learn about Kevin Fagan’s powerful stories of homelessness and recovery that come from connections made in a life-long reporting career.

iRobot co-founder Rodney Brooks details why humanoid robots won't learn human-level dexterity from current methods, how to make them safe for humans, and more (Rodney Brooks)
https://rodneybrooks.com/why-todays-humanoids-wont-learn-dexterity/

DeCodec: Rethinking Audio Codecs as Universal Disentangled Representation Learners
Xiaoxue Luo, Jinwei Huang, Runyan Yang, Yingying Gao, Junlan Feng, Chao Deng, Shilei Zhang
https://arxiv.org/abs/2509.09201

DeCodec: Rethinking Audio Codecs as Universal Disentangled Representation Learners
Universal audio codecs learn entangled representations across audio types, whereas some specific codecs offer decoupled representations but are limited to speech. Real-world audio, however, often contains mixed speech and background sounds, and downstream tasks require selective access to these components. Therefore, we rethink the audio codec as a universal disentangled representation learner to enable controllable feature selection across different audio tasks. To this end, we introduce DeCod…

#Food #FoodampDining Stop Guessing How Much Coffee to Use. The 'Golden Ratio' Is the Answer

Want a Perfect Pot of Coffee? Experts Suggest the 'Golden Ratio.' Here's How it Works
I spoke with a coffee educator to learn about the "Golden Ratio," grind size and other factors that make or break your home brew.

Event: Games for degrowth
Come to our AGM, help design games, and learn about our work Steady State Manchester Notice of Annual General Meeting Date: 17 September, 2025 Time: 7:00pm Venue: Methodist Central Hall, Oldham Street, Manchester, M1 1JQ Map: Part 1. Formal business. All may observe but only paid-up members may participate. We aim to keep this part of the meeting as brief as possible.]

Event: Games for degrowth
Come to our AGM, help design games, and learn about our work Steady State Manchester Notice of Annual General Meeting Date: 17 September, 2025 Time: 7:00pm Venue: Methodist Central Hall, Oldham Str…

Every year someone on #GBBO tries to construct some kind of fancy architectural showstopper thing using a dough that isn't gingerbread. Every year they say "this is risky, it might crack" and you know what, every year someone on GBBO sees their fancy architectural showstopper crack apart.
Now I'm no baker, but I can at least learn from history. Use gingerbread, people!

MHRV (Mechanical Heat-Recovery Ventilation) and HR25H Review - Is MHRV right for you? Learn from our years with the Vent-Axia humidistat HR25. #frugal #cutCarbon -

MHRV (Mechanical Heat-Recovery Ventilation) and HR25H Review
Is MHRV right for you? Learn from our years with the Vent-Axia humidistat HR25. #frugal #cutCarbon

Cowboys won't learn from the Micah Parsons debacle because Jerry Jones doesn't want to learn from it

https://www.cbssports.com/nfl/news/cowboys

Event: Games for degrowth
Come to our AGM, help design games, and learn about our work Steady State Manchester Notice of Annual General Meeting Date: 17 September, 2025 Time: 7:00pm Venue: Methodist Central Hall, Oldham Street, Manchester, M1 1JQ Map: Part 1. Formal business. All may observe but only paid-up members may participate. We aim to keep this part of the meeting as brief as possible.]

Event: Games for degrowth
Come to our AGM, help design games, and learn about our work Steady State Manchester Notice of Annual General Meeting Date: 17 September, 2025 Time: 7:00pm Venue: Methodist Central Hall, Oldham Str…

Hands off the EU’s environmental laws! #Europe

#petition Hands off the EU’s environmental laws!
The European Commission plans to weaken laws designed to protect the environment – both at home and worldwide. Details will be announced in the fall through an “Omnibus package.” The Commission is currently gathering feedback from the public to learn what people think. Get involved and make your voice heard!

Propagation Tree Is Not Deep: Adaptive Graph Contrastive Learning Approach for Rumor Detection
Chaoqun Cui, Caiyan Jia
https://arxiv.org/abs/2508.07201 https://

Propagation Tree Is Not Deep: Adaptive Graph Contrastive Learning Approach for Rumor Detection
Rumor detection on social media has become increasingly important. Most existing graph-based models presume rumor propagation trees (RPTs) have deep structures and learn sequential stance features along branches. However, through statistical analysis on real-world datasets, we find RPTs exhibit wide structures, with most nodes being shallow 1-level replies. To focus learning on intensive substructures, we propose Rumor Adaptive Graph Contrastive Learning (RAGCL) method with adaptive view augmen…

I'm usually not into reality shows. This one is great though. I guess it's the mix of London and farming for me. He comes across very genuine.
https://www.theguardian.com/tv-and-radio/2025/aug/10/mc-fekky-l…

Looking to Learn: Token-wise Dynamic Gating for Low-Resource Vision-Language Modelling
Bianca-Mihaela Ganescu, Suchir Salhan, Andrew Caines, Paula Buttery
https://arxiv.org/abs/2510.08470

Looking to Learn: Token-wise Dynamic Gating for Low-Resource Vision-Language Modelling
Training vision-language models on cognitively-plausible amounts of data requires rethinking how models integrate multimodal information. Within the constraints of the Vision track for the BabyLM Challenge 2025, we propose a lightweight decoder-based architecture with (1) token-wise dynamic gating for adaptive fusion of linguistic and visual cues, (2) feature modulation and channel attention to maximise the utility of limited visual information and (3) auxiliary contrastive objectives for visua…

Testing the Cosmic Distance Duality Relation with Neural Kernel Gaussian Process Regression
Xin Luo, Nan Liang
https://arxiv.org/abs/2508.07040 https://arx…

Testing the Cosmic Distance Duality Relation with Neural Kernel Gaussian Process Regression
In this work, we test the cosmic distance duality relation (CDDR) by combining Pantheon+ Type Ia supernova (SNe Ia) data and DESI DR2 baryon acoustic oscillation (BAO) measurements. To resolve the redshift mismatch between the two datasets, we develop a new method called Neural Kernel Gaussian Process Regression (NKGPR), which uses two neural networks to simultaneously learn the mean and kernel functions of a Gaussian process. This approach improves upon traditional Gaussian process regression …

Hierarchical Placement Learning for Network Slice Provisioning
Jesutofunmi Ajayi, Antonio Di Maio, Torsten Braun
https://arxiv.org/abs/2508.06432 https://a…

Hierarchical Placement Learning for Network Slice Provisioning
In this work, we aim to address the challenge of slice provisioning in edge-based mobile networks. We propose a solution that learns a service function chain placement policy for Network Slice Requests, to maximize the request acceptance rate, while minimizing the average node resource utilization. To do this, we consider a Hierarchical Multi-Armed Bandit problem and propose a two-level hierarchical bandit solution which aims to learn a scalable placement policy that optimizes the stated object…

Let's see what we learn from this case of "Big corporation buys up important open source infrastructure because it's organized as a company". #arduino

SF Supervisor Matt Dorsey was one of 15 BCDC Commissioners who voted to take away weekday access to the Richmond-San Rafael Bridge bike path.
Dorsey said cyclists need to learn to compromise more. So much for his bragging about using bikeshare 🚮
https://richmond…

Richmond bridge bike lane access to be reduced starting this fall
The change to Thursday afternoons through Sundays, starting in October, lets transportation officials test the impact on traffic congestion.

If you're interested in productivity and would like to learn more about the psychology of getting stuff done, I'm running a *free* webinar on how coaching supports productivity next Weds at 12pm BST.
https://www.eventbrite.co.uk/e/1267475706959?aff=oddtdtcreator

In late 2020, I wrote an odd little musing of thoughts. Just re-read it today, and I like these lines:
"Love must have compassion and empathy at its beck and call.
As uncomfortable as truth can so often be, it remains true without any believers, without any champions arguing its case.
Pain is, heartrendingly, unavoidable. Misery is an option we can choose willfully by thinking we know everything and have nothing to learn."

Midas Touch but for Pain
some people have the ability to diminish all they touch

Two Sides of the Same Optimization Coin: Model Degradation and Representation Collapse in Graph Foundation Models
Xunkai Li, Daohan Su, Sicheng Liu, Ru Zhang, Rong-Hua Li, Guoren Wang
https://arxiv.org/abs/2509.08401

Two Sides of the Same Optimization Coin: Model Degradation and Representation Collapse in Graph Foundation Models
Graph foundation models, inspired by the success of LLMs, are designed to learn the optimal embedding from multi-domain TAGs for the downstream cross-task generalization capability. During our investigation, graph VQ-MAE stands out among the increasingly diverse landscape of GFM architectures. This is attributed to its ability to jointly encode topology and textual attributes from multiple domains into discrete embedding spaces with clear semantic boundaries. Despite its potential, domain gener…

Maximally Useful and Minimally Redundant: The Key to Self Supervised Learning for Imbalanced Data
Yash Kumar Sharma, Vineet Nair, Wilson Naik
https://arxiv.org/abs/2509.08469 ht…

Maximally Useful and Minimally Redundant: The Key to Self Supervised Learning for Imbalanced Data
The robustness of contrastive self-supervised learning (CSSL) for imbalanced datasets is largely unexplored. CSSL usually makes use of \emph{multi-view} assumptions to learn discriminatory features via similar and dissimilar data samples. CSSL works well on balanced datasets, but does not generalize well for imbalanced datasets. In a very recent paper, as part of future work, Yann LeCun pointed out that the self-supervised multiview framework can be extended to cases involving \emph{more than t…

Replaced article(s) found for cs.RO. https://arxiv.org/list/cs.RO/new
[1/3]:
- Learn to Teach: Sample-Efficient Privileged Learning for Humanoid Locomotion over Diverse Terrains
Feiyang Wu, Xavier Nal, Jaehwi Jang, Wei Zhu, Zhaoyuan Gu, Anqi Wu, Ye Zhao

Something someone just said to me:
I need to learn how to use AI cause I don't have time to do anything.
Yeah, maybe it's not abstaining from LLMs that's the problem.

Discover the horrifying OKC-Spatz case, spanning over 27 years of legal battles among 3 EU countries! Your donations can truly impact the victims finish the case in France. Learn more at: https://chardonsbleus.org/donate

What One Cannot, Two Can: Two-Layer Transformers Provably Represent Induction Heads on Any-Order Markov Chains
Chanakya Ekbote, Marco Bondaschi, Nived Rajaraman, Jason D. Lee, Michael Gastpar, Ashok Vardhan Makkuva, Paul Pu Liang
https://arxiv.org/abs/2508.07208

What One Cannot, Two Can: Two-Layer Transformers Provably Represent Induction Heads on Any-Order Markov Chains
In-context learning (ICL) is a hallmark capability of transformers, through which trained models learn to adapt to new tasks by leveraging information from the input context. Prior work has shown that ICL emerges in transformers due to the presence of special circuits called induction heads. Given the equivalence between induction heads and conditional k-grams, a recent line of work modeling sequential inputs as Markov processes has revealed the fundamental impact of model depth on its ICL capa…

Whither symbols in the era of advanced neural networks?
Thomas L. Griffiths, Brenden M. Lake, R. Thomas McCoy, Ellie Pavlick, Taylor W. Webb
https://arxiv.org/abs/2508.05776 htt…

Whither symbols in the era of advanced neural networks?
Some of the strongest evidence that human minds should be thought about in terms of symbolic systems has been the way they combine ideas, produce novelty, and learn quickly. We argue that modern neural networks -- and the artificial intelligence systems built upon them -- exhibit similar abilities. This undermines the argument that the cognitive processes and representations used by human minds are symbolic, although the fact that these neural networks are typically trained on data generated by…

Decoupling Clinical and Class-Agnostic Features for Reliable Few-Shot Adaptation under Shift
Umaima Rahman, Raza Imam, Mohammad Yaqub, Dwarikanath Mahapatra
https://arxiv.org/abs/2509.09397

Decoupling Clinical and Class-Agnostic Features for Reliable Few-Shot Adaptation under Shift
Medical vision-language models (VLMs) offer promise for clinical decision support, yet their reliability under distribution shifts remains a major concern for safe deployment. These models often learn task-agnostic correlations due to variability in imaging protocols and free-text reports, limiting their generalizability and increasing the risk of failure in real-world settings. We propose DRiFt, a structured feature decoupling framework that explicitly separates clinically relevant signals fro…

Whoopi Goldberg says get a tan and learn some Spanish; how about a Spanish song, something easy to learn, and something a large group of people could sing if they were, like, protesting?

In-situ Value-aligned Human-Robot Interactions with Physical Constraints
Hongtao Li, Ziyuan Jiao, Xiaofeng Liu, Hangxin Liu, Zilong Zheng
https://arxiv.org/abs/2508.07606 https:…

In-situ Value-aligned Human-Robot Interactions with Physical Constraints
Equipped with Large Language Models (LLMs), human-centered robots are now capable of performing a wide range of tasks that were previously deemed challenging or unattainable. However, merely completing tasks is insufficient for cognitive robots, who should learn and apply human preferences to future scenarios. In this work, we propose a framework that combines human preferences with physical constraints, requiring robots to complete tasks while considering both. Firstly, we developed a benchmar…

On Condensation Management - Learn how you can best avoid problems with condensation, damp and mould in winter. #mould #condensation #ventilation -

On Condensation Management
Learn how you can best avoid problems with condensation, damp and mould in winter. #mould #condensation #ventilation

Inside the NFL Academy: 6:30am gym sessions, school, and 68 students with a dream https://www.nytimes.com/athletic/6651852/2025/10/10/nfl-academy-loughborough-nfl-international-games/

Trump wants a Nobel Prize. He'll learn Friday if his campaign paid off. (Washington Post)
https://www.washingtonpost.com/politics/2025/10/09/trump-nobel-prize-friday/
http://www.memeorandum.com/251009/p18#a251009p18

Former Centers for Disease Control and Prevention Director
Susan Monarez is scheduled to testify about her high-profile firing from the agency at a hearing next Wednesday before the
Senate Committee on Health, Education, Labor and Pensions.
Chair Bill Cassidy (R-La.) said the hearing will allow lawmakers on the committee to directly learn about the circumstances surrounding Monarez’s ouster and the resignations of other top officials at the CDC.
Cassidy said his commit…

Group Distributionally Robust Machine Learning under Group Level Distributional Uncertainty
Xenia Konti, Yi Shen, Zifan Wang, Karl Henrik Johansson, Michael J. Pencina, Nicoleta J. Economou-Zavlanos, Michael M. Zavlanos
https://arxiv.org/abs/2509.08942

Group Distributionally Robust Machine Learning under Group Level Distributional Uncertainty
The performance of machine learning (ML) models critically depends on the quality and representativeness of the training data. In applications with multiple heterogeneous data generating sources, standard ML methods often learn spurious correlations that perform well on average but degrade performance for atypical or underrepresented groups. Prior work addresses this issue by optimizing the worst-group performance. However, these approaches typically assume that the underlying data distribution…

Spatial-ORMLLM: Improve Spatial Relation Understanding in the Operating Room with Multimodal Large Language Model
Peiqi He, Zhenhao Zhang, Yixiang Zhang, Xiongjun Zhao, Shaoliang Peng
https://arxiv.org/abs/2508.08199

Spatial-ORMLLM: Improve Spatial Relation Understanding in the Operating Room with Multimodal Large Language Model
Precise spatial modeling in the operating room (OR) is foundational to many clinical tasks, supporting intraoperative awareness, hazard avoidance, and surgical decision-making. While existing approaches leverage large-scale multimodal datasets for latent-space alignment to implicitly learn spatial relationships, they overlook the 3D capabilities of MLLMs. However, this approach raises two issues: (1) Operating rooms typically lack multiple video and audio sensors, making multimodal 3D data diff…

Thermal Imaging Survey: DIY vs Professional - Learn how a thermal imaging survey can spot real problems. What does it cost? Is a trained operator useful? #thermalImaging #cutCarbon -

Thermal Imaging Survey: DIY vs Professional
Learn how a thermal imaging survey can spot real problems. What does it cost? Is a trained operator useful? #thermalImaging #cutCarbon

Risk-Bounded Multi-Agent Visual Navigation via Dynamic Budget Allocation
Viraj Parimi, Brian C. Williams
https://arxiv.org/abs/2509.08157 https://arxiv.org…

Risk-Bounded Multi-Agent Visual Navigation via Dynamic Budget Allocation
Safe navigation is essential for autonomous systems operating in hazardous environments, especially when multiple agents must coordinate using just visual inputs over extended time horizons. Traditional planning methods excel at solving long-horizon tasks but rely on predefined distance metrics, while safe Reinforcement Learning (RL) can learn complex behaviors using high-dimensional inputs yet struggles with multi-agent, goal-conditioned scenarios. Recent work combined these paradigms by lever…

Fourier Learning Machines: Nonharmonic Fourier-Based Neural Networks for Scientific Machine Learning
Mominul Rubel, Adam Meyers, Gabriel Nicolosi
https://arxiv.org/abs/2509.08759

Fourier Learning Machines: Nonharmonic Fourier-Based Neural Networks for Scientific Machine Learning
We introduce the Fourier Learning Machine (FLM), a neural network (NN) architecture designed to represent a multidimensional nonharmonic Fourier series. The FLM uses a simple feedforward structure with cosine activation functions to learn the frequencies, amplitudes, and phase shifts of the series as trainable parameters. This design allows the model to create a problem-specific spectral basis adaptable to both periodic and nonperiodic functions. Unlike previous Fourier-inspired NN models, the …

CLIPin: A Non-contrastive Plug-in to CLIP for Multimodal Semantic Alignment
Shengzhu Yang, Jiawei Du, Shuai Lu, Weihang Zhang, Ningli Wang, Huiqi Li
https://arxiv.org/abs/2508.06434

CLIPin: A Non-contrastive Plug-in to CLIP for Multimodal Semantic Alignment
Large-scale natural image-text datasets, especially those automatically collected from the web, often suffer from loose semantic alignment due to weak supervision, while medical datasets tend to have high cross-modal correlation but low content diversity. These properties pose a common challenge for contrastive language-image pretraining (CLIP): they hinder the model's ability to learn robust and generalizable representations. In this work, we propose CLIPin, a unified non-contrastive plug-in t…

BeyondMimic: From Motion Tracking to Versatile Humanoid Control via Guided Diffusion
Takara E. Truong, Qiayuan Liao, Xiaoyu Huang, Guy Tevet, C. Karen Liu, Koushil Sreenath
https://arxiv.org/abs/2508.08241

BeyondMimic: From Motion Tracking to Versatile Humanoid Control via Guided Diffusion
Learning skills from human motions offers a promising path toward generalizable policies for whole-body humanoid control, yet two key cornerstones are missing: (1) a high-quality motion tracking framework that faithfully transforms large-scale kinematic references into robust and extremely dynamic motions on real hardware, and (2) a distillation approach that can effectively learn these motion primitives and compose them to solve downstream tasks. We address these gaps with BeyondMimic, the fir…

Towards Interpretable Deep Neural Networks for Tabular Data
Khawla Elhadri, J\"org Schl\"otterer, Christin Seifert
https://arxiv.org/abs/2509.08617 https://

Towards Interpretable Deep Neural Networks for Tabular Data
Tabular data is the foundation of many applications in fields such as finance and healthcare. Although DNNs tailored for tabular data achieve competitive predictive performance, they are blackboxes with little interpretability. We introduce XNNTab, a neural architecture that uses a sparse autoencoder (SAE) to learn a dictionary of monosemantic features within the latent space used for prediction. Using an automated method, we assign human-interpretable semantics to these features. This allows u…

Sticker-TTS: Learn to Utilize Historical Experience with a Sticker-driven Test-Time Scaling Framework
Jie Chen, Jinhao Jiang, Yingqian Min, Zican Dong, Shijie Wang, Wayne Xin Zhao, Ji-Rong Wen
https://arxiv.org/abs/2509.05007

Sticker-TTS: Learn to Utilize Historical Experience with a Sticker-driven Test-Time Scaling Framework
Large reasoning models (LRMs) have exhibited strong performance on complex reasoning tasks, with further gains achievable through increased computational budgets at inference. However, current test-time scaling methods predominantly rely on redundant sampling, ignoring the historical experience utilization, thereby limiting computational efficiency. To overcome this limitation, we propose Sticker-TTS, a novel test-time scaling framework that coordinates three collaborative LRMs to iteratively e…

Conv4Rec: A 1-by-1 Convolutional AutoEncoder for User Profiling through Joint Analysis of Implicit and Explicit Feedbacks
Antoine Ledent, Petr Kasalick\'y, Rodrigo Alves, Hady W. Lauw
https://arxiv.org/abs/2509.07499

Conv4Rec: A 1-by-1 Convolutional AutoEncoder for User Profiling through Joint Analysis of Implicit and Explicit Feedbacks
We introduce a new convolutional AutoEncoder architecture for user modelling and recommendation tasks with several improvements over the state of the art. Firstly, our model has the flexibility to learn a set of associations and combinations between different interaction types in a way that carries over to each user and item. Secondly, our model is able to learn jointly from both the explicit ratings and the implicit information in the sampling pattern (which we refer to as `implicit feedback')…

An Interpretable Deep Learning Model for General Insurance Pricing
Patrick J. Laub, Tu Pho, Bernard Wong
https://arxiv.org/abs/2509.08467 https://arxiv.org…

An Interpretable Deep Learning Model for General Insurance Pricing
This paper introduces the Actuarial Neural Additive Model, an inherently interpretable deep learning model for general insurance pricing that offers fully transparent and interpretable results while retaining the strong predictive power of neural networks. This model assigns a dedicated neural network (or subnetwork) to each individual covariate and pairwise interaction term to independently learn its impact on the modeled output while implementing various architectural constraints to allow for…

Agent Learning via Early Experience
Kai Zhang, Xiangchao Chen, Bo Liu, Tianci Xue, Zeyi Liao, Zhihan Liu, Xiyao Wang, Yuting Ning, Zhaorun Chen, Xiaohan Fu, Jian Xie, Yuxuan Sun, Boyu Gou, Qi Qi, Zihang Meng, Jianwei Yang, Ning Zhang, Xian Li, Ashish Shah, Dat Huynh, Hengduo Li, Zi Yang, Sara Cao, Lawrence Jang, Shuyan Zhou, Jiacheng Zhu, Huan Sun, Jason Weston, Yu Su, Yifan Wu

Agent Learning via Early Experience
A long-term goal of language agents is to learn and improve through their own experience, ultimately outperforming humans in complex, real-world tasks. However, training agents from experience data with reinforcement learning remains difficult in many environments, which either lack verifiable rewards (e.g., websites) or require inefficient long-horizon rollouts (e.g., multi-turn tool use). As a result, most current agents rely on supervised fine-tuning on expert data, which is challenging to s…

Super-Insulating Our Living Room (2010) - Learn how we retrofitted interior wall insulation for winter warmth and summer cool, and trimmed our carbon footprint. #aerogel #frugal #cutCarbon -

Super-Insulating Our Living Room (2010)
Learn how we retrofitted interior wall insulation for winter warmth and summer cool, and trimmed our carbon footprint. #aerogel #frugal #cutCarbon

Gaze on the Prize: Shaping Visual Attention with Return-Guided Contrastive Learning
Andrew Lee, Ian Chuang, Dechen Gao, Kai Fukazawa, Iman Soltani
https://arxiv.org/abs/2510.08442

Gaze on the Prize: Shaping Visual Attention with Return-Guided Contrastive Learning
Visual Reinforcement Learning (RL) agents must learn to act based on high-dimensional image data where only a small fraction of the pixels is task-relevant. This forces agents to waste exploration and computational resources on irrelevant features, leading to sample-inefficient and unstable learning. To address this, inspired by human visual foveation, we introduce Gaze on the Prize. This framework augments visual RL with a learnable foveal attention mechanism (Gaze), guided by a self-supervise…

Super-Insulating Our Living Room (2010) - Learn how we retrofitted interior wall insulation for winter warmth and summer cool, and trimmed our carbon footprint. #aerogel #frugal #cutCarbon -

Super-Insulating Our Living Room (2010)
Learn how we retrofitted interior wall insulation for winter warmth and summer cool, and trimmed our carbon footprint. #aerogel #frugal #cutCarbon

Shift Before You Learn: Enabling Low-Rank Representations in Reinforcement Learning
Bastien Dubail, Stefan Stojanovic, Alexandre Prouti\`ere
https://arxiv.org/abs/2509.05193 htt…

Shift Before You Learn: Enabling Low-Rank Representations in Reinforcement Learning
Low-rank structure is a common implicit assumption in many modern reinforcement learning (RL) algorithms. For instance, reward-free and goal-conditioned RL methods often presume that the successor measure admits a low-rank representation. In this work, we challenge this assumption by first remarking that the successor measure itself is not low-rank. Instead, we demonstrate that a low-rank structure naturally emerges in the shifted successor measure, which captures the system dynamics after bypa…

The Physics of Data and Tasks: Theories of Locality and Compositionality in Deep Learning
Alessandro Favero
https://arxiv.org/abs/2510.06106 https://arxiv.…

The Physics of Data and Tasks: Theories of Locality and Compositionality in Deep Learning
Deep neural networks have achieved remarkable success, yet our understanding of how they learn remains limited. These models can learn high-dimensional tasks, which is generally statistically intractable due to the curse of dimensionality. This apparent paradox suggests that learnable data must have an underlying latent structure. What is the nature of this structure? How do neural networks encode and exploit it, and how does it quantitatively impact performance - for instance, how does general…