Tootfinder

Hello North Sea. #denmark #enbee

@… https://www.traficom.fi/feed/rss/fi

Hey, just a reminder that Thursday 18th there’s the “IPTC Photo Metadata Conference 2025”, a free Zoom event open to all. I’m a speaker (focusing on #C2PA). It looks like it should be interesting for anyone interested in the tangle of issues about media provenance, #genAI, publishing workflow, social media, and…

Photo Metadata Conference 2025 - IPTC
IPTC is the global standards body of the news media. We provide the technical foundation for the news ecosystem.

Courtship with consequences.

The UK has ramped up the use of facial recognition, AI, and internet regulation to address crime and other issues, stoking concerns of surveillance overreach (New York Times)
https://www.nytimes.com/2025/09/17/t…

Has Britain Gone Too Far With Its Digital Controls?
British authorities have ramped up the use of facial recognition, artificial intelligence and internet regulation to address crime and other issues, stoking concerns of surveillance overreach.

Ich habe vorhin ungefähr 30 #Störche auf einem frisch beackerten Feld bei #Appenheim in #Rheinhessen gesehen. Als ich Kind war, gab es hier überhaupt keine Störche. Ich wüsste zu gerne, wo die a…

berkstan_web: Webgraph (Berkeley-Stanford)
The web graph of Berkeley and Stanford Universities (berkeley.edu and stanford.edu), as collected in 2002. Nodes represent pages and directed edges represent hyperlinks between them.
This network has 685231 nodes and 7600595 edges.
Tags: Informational, Web graph, Unweighted
https://

Und natürlich gab es auch heute ein #Kundenglücklichmachungsposteinlieferungsgeräusch, wenn auch anders und später als sonst üblich.

Heute vor 35 Jahren: Am 16. August 1990 führte #China am Teststandort Lop Nor (Provinz Xinjiang) einen unterirdischen #Atomtest durch, Nr. 43. Er diente der Waffenentwicklung (JL-1), die Sprengkraft lag bei 189 Kt TNT.

Right now it does do a limited form of isolation by putting everything in a temp directory, but of course this just makes sure you declare your header files. It doesn’t do anything to stop you referencing stuff by absolute path, or using the wrong compiler version, or depending on the system time, or hitting the network, or…

Today, cherry oat muffins, egg bacon tea coffee. More fringe tonight, and then a wild ride culminating in a week away for work starting Sunday. #TogetherBreakfast https://photos.app.goo.gl/TTToGZHVGkidU3MJ8

»👨‍💻️ Rust & Co. – sichere moderne Programmiersprachen im Vergleich
(CC2tv Folge 406)«
Für einige ist dieses Video sicherlich sehr trocken über moderne Programmiersprachen doch für mich angenehm ruhig erklärt so wie Informativ obwohl ich schon ein paar Dinge kenne was erklärt wird. Es werden die klaren Strukturierung so wie weitere Vorteile von Rust erläutert.
📺

This article does not in fact mention anything Reality Winner says she shares with Donald Trump.
https://www.nytimes.com/2025/09/10/books/review/i-am-not-your-enemy-reality-winner.html

I think the Warriors might just edge this 🏈 #rugbyunion #rugby

Hätte Zeit, Musse und ein Hüngerchen.
Aber im Bordbistro ist mal wieder etwas kaputt und man bekommt nur Chips oder andere Snacks.
Und dann wundert man sich, dass die Leute lieber Futter vom Bahnhof anschleppen? 🙄🤷‍♂️ #TGVLyria

I resemble this remark . . .
https://beige.party/@mayaisloading/115384801259480525

#potd Anna Mabo & die Buben @ Porgy & Bess, presenting their new record "Mittelschwere Extase". Very nice gig. But my mobile phone camera is really crap. https://domm.plix.at/potd/2025-10-15.h

Replaced article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[7/7]:
- Next Edit Prediction: Learning to Predict Code Edits from Context and Interaction History
Ruofan Lu, Yintong Huo, Meng Zhang, Yichen Li, Michael R. Lyu

Researchers say they used generative AI algorithms to design novel antibiotics targeting drug-resistant Neisseria gonorrhoeae and MRSA (Anne Trafton/MIT News)
https://news.mit.edu/2025/using-generative-ai-researchers-design-compoun…

Using generative AI, researchers design compounds that can kill drug-resistant bacteria
With help from artificial intelligence, MIT researchers designed novel antibiotics that can combat a drug-resistant form of Neisseria gonorrhoeae and multi-drug-resistant Staphylococcus aureus (MRSA).

I have been working on a build tool. Again.
This one… I think it’s got legs.
Here it is, concatenating a bunch of text files together:
https://codeberg.org/ooble/um/src/branch/main/tests/cat/oh.um
And here, compiling a simple C program, w…

Did I forget in yesterday’s rush to share Marie’s biscuits, egg bacon tea and coffee? Yes! Did we go to Fringe fest last night? Also yes! #TogetherBreakfast https://photos.app.goo.gl/XZKzuBT6doHWY3Wc9

Au revoir, Saint-Malo! 🥰
#MeFàStM

Wordle 1,550 2/6
⬜🟨⬜🟩🟨
🟩🟩🟩🟩🟩

#potd Today GüDi and I did a bunch of fixed rope routes in Höllental/Rax, followed by a very refreshing swim in Schwarza (16°) http://domm.plix.at/potd/2025-08-15.html

So Hallowe'en, where do you stand?

Replaced article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[6/7]:
- EasyEdit2: An Easy-to-use Steering Framework for Editing Large Language Models
Xu, Wang, Xu, Xu, Wang, Deng, Yao, Zheng, Chen, Zhang

@… Ooooof. I hope you recover quick.

Everyone is somewhere else at the moment so I made breakfast and then ate myself. One of us can’t eat today at all (colonoscopy). Fun times. Rustic sourdough cornbread (cheddar and poblano), egg and bacon, tea.
#TogetherBreakfast https://photos.app.goo.gl/NQ3mAkQ4QGrmjNpt9

Wordle 1,549 4/6
⬜🟩🟨⬜⬜
🟩🟩⬜⬜⬜
🟩🟩⬜🟨🟨
🟩🟩🟩🟩🟩

@… I have convinced myself there isn’t one.
You may now proceed with the game.

Today Marie made biscuits and then without eating, left for her colonoscopy. (Kudos for that!) egg and bacon, tea and coffee. Just work today! #TogetherBreakfast https://photos.app.goo.gl/nVCFenc6eLPE3KPZ9

Replaced article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[5/7]:
- Marigold-DC: Zero-Shot Monocular Depth Completion with Guided Diffusion
Viola, Qu, Metzger, Ke, Becker, Schindler, Obukhov

I wish J. K. Rowling a bag full of nails for breakfast.

Replaced article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[4/7]:
- Piecewise Deterministic Markov Processes for Bayesian Neural Networks
Ethan Goan, Dimitri Perrin, Kerrie Mengersen, Clinton Fookes

@… Fuck you I’m gonna be staring at this all day now

@… I’m afraid that at least as far as I’m concerned, we’re at the “poke it and see what happens” stage of debugging.

Replaced article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[3/7]:
- Quantized Neural Networks for Microcontrollers: A Comprehensive Review of Methods, Platforms, and...
Hamza A. Abushahla, Dara Varam, Ariel J. N. Panopio, Mohamed I. AlHajri

@… Bah. Try setting `LD_LIBRARY_PATH` too?
Don’t forget to preserve the existing one. Something like:
export LD_LIBRARY_PATH=“${pkg}:$LD_LIBRARY_PATH”

Replaced article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[2/7]:
- Convergence Analysis of Asynchronous Federated Learning with Gradient Compression for Non-Convex ...
Diying Yang, Yingwei Hou, Weigang Wu

@… Oh, add `export` to your shell hook?

@… That looks legit! I take it you verified that bearlibterminal’s got a .dylib?

Replaced article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[1/7]:
- Extended UCB Policies for Multi-armed Bandit Problems
Keqin Liu, Tianshuo Zheng, Zhi-Hua Zhou

@… Are you using Nix to provision GHC C deps, then Cabal to get Haskell deps? Or using cabal2nix?

Crosslisted article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[6/6]:
- Quantum Noise Tomography with Physics-Informed Neural Networks
Antonin Sulc
http…

I forgot to mention why I was thinking about Solid. It’s because I was scrolling through the NLnet-sponsored projects and saw a lot of them.
https://nlnet.nl/project/current.html

Crosslisted article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[5/6]:
- PeruMedQA: Benchmarking Large Language Models (LLMs) on Peruvian Medical Exams -- Dataset Constru...
Rodrigo M. Carrillo-Larco, Jesus Lov\'on Melgarejo, Manuel Castillo-Cara, Gusseppe Bravo-Rocca

@… Hahahaha, that’s awesome.
I imagine there are plenty of people around (including my partner) who work in multiple fluent languages, and their “native” one doesn’t get special treatment any more.

@… Oh, I guess at this point you have multiple fluent languages, so it makes sense. Perhaps “one of the languages I’m fluent in” would work better than “my native language”.
Either way, thanks for the clarification!

Crosslisted article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[4/6]:
- Predictable Compression Failures: Why Language Models Actually Hallucinate
Leon Chlon, Ahmed Karim, Maggie Chlon

@… @… I usually recall e.g. German as if it were English, but I recall Hindi films I watched when I was a child in Hindi, even though I don’t know what the words mean (and have forgotten the subtitles).

Crosslisted article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[3/6]:
- Physics-informed neural network solves minimal surfaces in curved spacetime
Koji Hashimoto, Koichi Kyo, Masaki Murata, Gakuto Ogiwara, Norihiro Tanahashi

@… Let’s try that again.
https://mastodon.functional.computer/@samir/114869348120796747

Those of you who frequently watch TV shows or movies in languages you don’t understand (e.g. with subtitles):
When you recall dialogue in the show/film, do you remember it in the actual language you heard, or in your native language?

Crosslisted article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[2/6]:
- Trial-Level Time-frequency EEG Desynchronization as a Neural Marker of Pain
D. A. Blanco-Mora, A. Dierolf, J. Gon\c{c}alves, M. van Der Meulen

@… Goddamnit now I have to redo the poll. You’re bang on.

Crosslisted article(s) found for cs.LG. https://arxiv.org/list/cs.LG/new
[1/6]:
- Learning Stackable and Skippable LEGO Bricks for Efficient, Reconfigurable, and Variable-Resoluti...
Zheng, Wang, Yuan, Ning, He, You, Yang, Zhou

@… @… I hope so too. 😊

@… @… You folks are making it hard to not be there!

Dynamic Relational Priming Improves Transformer in Multivariate Time Series
Hunjae Lee, Corey Clark
https://arxiv.org/abs/2509.12196 https://arxiv.org/pdf/…

Dynamic Relational Priming Improves Transformer in Multivariate Time Series
Standard attention mechanisms in transformers employ static token representations that remain unchanged across all pair-wise computations in each layer. This limits their representational alignment with the potentially diverse relational dynamics of each token-pair interaction. While they excel in domains with relatively homogeneous relationships, standard attention's static relational learning struggles to capture the diverse, heterogeneous inter-channel dependencies of multivariate time serie…

@… Urgh, that sucks!
I will boost. Let me know if I can do anything else.

Event2Vec: A Geometric Approach to Learning Composable Representations of Event Sequences
Antonin Sulc
https://arxiv.org/abs/2509.12188 https://arxiv.org/p…

Event2Vec: A Geometric Approach to Learning Composable Representations of Event Sequences
The study of neural representations, both in biological and artificial systems, is increasingly revealing the importance of geometric and topological structures. Inspired by this, we introduce Event2Vec, a novel framework for learning representations of discrete event sequences. Our model leverages a simple, additive recurrent structure to learn composable, interpretable embeddings. We provide a theoretical analysis demonstrating that, under specific training objectives, our model's learned rep…

@… Ah, I know the feeling.
Out of curiosity, does the source matter? My partner is Romanian and deals much better with sheep/goat cheese, or very fresh cheese such as mozzarella.

@… @… Looks interesting, I should check it out!
I really would love to abandon Kubernetes and run into the hills, but everyone wants it… I find it very frustrating.

All that structure matches does not glitter
Maya M. Martirossyan, Thomas Egg, Philipp Hoellmer, George Karypis, Mark Transtrum, Adrian Roitberg, Mingjie Liu, Richard G. Hennig, Ellad B. Tadmor, Stefano Martiniani
https://arxiv.org/abs/2509.12178

All that structure matches does not glitter
Generative models for materials, especially inorganic crystals, hold potential to transform the theoretical prediction of novel compounds and structures. Advancement in this field depends critically on robust benchmarks and minimal, information-rich datasets that enable meaningful model evaluation. This paper critically examines common datasets and reported metrics for a crystal structure prediction task$\unicode{x2014}$generating the most likely structures given the chemical composition of a m…

From Autoencoders to CycleGAN: Robust Unpaired Face Manipulation via Adversarial Learning
Collin Guo
https://arxiv.org/abs/2509.12176 https://arxiv.org/pdf…

From Autoencoders to CycleGAN: Robust Unpaired Face Manipulation via Adversarial Learning
Human face synthesis and manipulation are increasingly important in entertainment and AI, with a growing demand for highly realistic, identity-preserving images even when only unpaired, unaligned datasets are available. We study unpaired face manipulation via adversarial learning, moving from autoencoder baselines to a robust, guided CycleGAN framework. While autoencoders capture coarse identity, they often miss fine details. Our approach integrates spectral normalization for stable training, i…

Learning Neural Networks by Neuron Pursuit
Akshay Kumar, Jarvis Haupt
https://arxiv.org/abs/2509.12154 https://arxiv.org/pdf/2509.12154

Learning Neural Networks by Neuron Pursuit
The first part of this paper studies the evolution of gradient flow for homogeneous neural networks near a class of saddle points exhibiting a sparsity structure. The choice of these saddle points is motivated from previous works on homogeneous networks, which identified the first saddle point encountered by gradient flow after escaping the origin. It is shown here that, when initialized sufficiently close to such saddle points, gradient flow remains near the saddle point for a sufficiently lon…

Do machine learning climate models work in changing climate dynamics?
Maria Conchita Agana Navarro, Geng Li, Theo Wolf, Mar\'ia P\'erez-Ortiz
https://arxiv.org/abs/2509.12147

Do machine learning climate models work in changing climate dynamics?
Climate change is accelerating the frequency and severity of unprecedented events, deviating from established patterns. Predicting these out-of-distribution (OOD) events is critical for assessing risks and guiding climate adaptation. While machine learning (ML) models have shown promise in providing precise, high-speed climate predictions, their ability to generalize under distribution shifts remains a significant limitation that has been underexplored in climate contexts. This research systema…

$K$-Level Policy Gradients for Multi-Agent Reinforcement Learning
Aryaman Reddi, Gabriele Tiboni, Jan Peters, Carlo D'Eramo
https://arxiv.org/abs/2509.12117 https://

$K$-Level Policy Gradients for Multi-Agent Reinforcement Learning
Actor-critic algorithms for deep multi-agent reinforcement learning (MARL) typically employ a policy update that responds to the current strategies of other agents. While being straightforward, this approach does not account for the updates of other agents at the same update step, resulting in miscoordination. In this paper, we introduce the $K$-Level Policy Gradient (KPG), a method that recursively updates each agent against the updated policies of other agents, speeding up the discovery of ef…

Draw a Portrait of Your Graph Data: An Instance-Level Profiling Framework for Graph-Structured Data
Tianqi Zhao, Russa Biswas, Megha Khosla
https://arxiv.org/abs/2509.12094 http…

Draw a Portrait of Your Graph Data: An Instance-Level Profiling Framework for Graph-Structured Data
Graph machine learning models often achieve similar overall performance yet behave differently at the node level, failing on different subsets of nodes with varying reliability. Standard evaluation metrics such as accuracy obscure these fine grained differences, making it difficult to diagnose when and where models fail. We introduce NodePro, a node profiling framework that enables fine-grained diagnosis of model behavior by assigning interpretable profile scores to individual nodes. These scor…

Deceptive Risk Minimization: Out-of-Distribution Generalization by Deceiving Distribution Shift Detectors
Anirudha Majumdar
https://arxiv.org/abs/2509.12081 https://

Deceptive Risk Minimization: Out-of-Distribution Generalization by Deceiving Distribution Shift Detectors
This paper proposes deception as a mechanism for out-of-distribution (OOD) generalization: by learning data representations that make training data appear independent and identically distributed (iid) to an observer, we can identify stable features that eliminate spurious correlations and generalize to unseen domains. We refer to this principle as deceptive risk minimization (DRM) and instantiate it with a practical differentiable objective that simultaneously learns features that eliminate dis…

A Time-Series Foundation Model by Universal Delay Embedding
Zijian Wang, Peng Tao, Jifan Shi, Rui Bao, Rui Liu, Luonan Chen
https://arxiv.org/abs/2509.12080 https://

A Time-Series Foundation Model by Universal Delay Embedding
This study introduces Universal Delay Embedding (UDE), a pretrained foundation model designed to revolutionize time-series forecasting through principled integration of delay embedding representation and Koopman operator prediction. Leveraging Takens' embedding theorem, UDE as a dynamical representation of observed data constructs two-dimensional subspace patches from Hankel matrices, theoretically preserving dynamical and topological properties of underlying dynamical systems. Such patches are…

Early Detection of Branched Broomrape (Phelipanche ramosa) Infestation in Tomato Crops Using Leaf Spectral Analysis and Machine Learning
Mohammadreza Narimani, Alireza Pourreza, Ali Moghimi, Parastoo Farajpoor, Hamid Jafarbiglu, Mohsen B. Mesgaran
https://arxiv.org/abs/2509.12074

Early Detection of Branched Broomrape (Phelipanche ramosa) Infestation in Tomato Crops Using Leaf Spectral Analysis and Machine Learning
Branched broomrape (Phelipanche ramosa) is a chlorophyll-deficient parasitic weed that threatens tomato production by extracting nutrients from the host. We investigate early detection using leaf-level spectral reflectance (400-2500 nm) and ensemble machine learning. In a field experiment in Woodland, California, we tracked 300 tomato plants across growth stages defined by growing degree days (GDD). Leaf reflectance was acquired with a portable spectrometer and preprocessed (band denoising, 1 n…

Foundational theory for optimal decision tree problems. II. Optimal hypersurface decision tree algorithm
Xi He
https://arxiv.org/abs/2509.12057 https://arx…

Foundational theory for optimal decision tree problems. II. Optimal hypersurface decision tree algorithm
Decision trees are a ubiquitous model for classification and regression tasks due to their interpretability and efficiency. However, solving the optimal decision tree (ODT) problem remains a challenging combinatorial optimization task. Even for the simplest splitting rules--axis-parallel hyperplanes--it is NP-hard to optimize. In Part I of this series, we rigorously defined the proper decision tree model through four axioms and, based on these, introduced four formal definitions of the ODT prob…

Hi-DARTS: Hierarchical Dynamically Adapting Reinforcement Trading System
Hoon Sagong, Heesu Kim, Hanbeen Hong
https://arxiv.org/abs/2509.12048 https://arxi…

Hi-DARTS: Hierarchical Dynamically Adapting Reinforcement Trading System
Conventional autonomous trading systems struggle to balance computational efficiency and market responsiveness due to their fixed operating frequency. We propose Hi-DARTS, a hierarchical multi-agent reinforcement learning framework that addresses this trade-off. Hi-DARTS utilizes a meta-agent to analyze market volatility and dynamically activate specialized Time Frame Agents for high-frequency or low-frequency trading as needed. During back-testing on AAPL stock from January 2024 to May 2025, H…

Travel Time and Weather-Aware Traffic Forecasting in a Conformal Graph Neural Network Framework
Mayur Patil, Qadeer Ahmed, Shawn Midlam-Mohler
https://arxiv.org/abs/2509.12043 h…

Travel Time and Weather-Aware Traffic Forecasting in a Conformal Graph Neural Network Framework
Traffic flow forecasting is essential for managing congestion, improving safety, and optimizing various transportation systems. However, it remains a prevailing challenge due to the stochastic nature of urban traffic and environmental factors. Better predictions require models capable of accommodating the traffic variability influenced by multiple dynamic and complex interdependent factors. In this work, we propose a Graph Neural Network (GNN) framework to address the stochasticity by leveragin…

Imitation Learning as Return Distribution Matching
Filippo Lazzati, Alberto Maria Metelli
https://arxiv.org/abs/2509.12026 https://arxiv.org/pdf/2509.12026…

Imitation Learning as Return Distribution Matching
We study the problem of training a risk-sensitive reinforcement learning (RL) agent through imitation learning (IL). Unlike standard IL, our goal is not only to train an agent that matches the expert's expected return (i.e., its average performance) but also its risk attitude (i.e., other features of the return distribution, such as variance). We propose a general formulation of the risk-sensitive IL problem in which the objective is to match the expert's return distribution in Wasserstein dist…

Learning non-Markovian Dynamical Systems with Signature-based Encoders
Eliott Pradeleix, R\'emy Hosseinkhan-Boucher, Alena Shilova, Onofrio Semeraro, Lionel Mathelin
https://arxiv.org/abs/2509.12022

Learning non-Markovian Dynamical Systems with Signature-based Encoders
Neural ordinary differential equations offer an effective framework for modeling dynamical systems by learning a continuous-time vector field. However, they rely on the Markovian assumption - that future states depend only on the current state - which is often untrue in real-world scenarios where the dynamics may depend on the history of past states. This limitation becomes especially evident in settings involving the continuous control of complex systems with delays and memory effects. To capt…

AMQ: Enabling AutoML for Mixed-precision Weight-Only Quantization of Large Language Models
Sangjun Lee, Seung-taek Woo, Jungyu Jin, Changhun Lee, Eunhyeok Park
https://arxiv.org/abs/2509.12019

AMQ: Enabling AutoML for Mixed-precision Weight-Only Quantization of Large Language Models
To enable broader deployment of Large Language Models (LLMs), it is essential to identify the best-performing model under strict memory constraints. We present AMQ, Automated Mixed-Precision Weight-Only Quantization, a framework that assigns layer-wise quantization bit-widths to optimally balance model quality and memory usage. However, the combinatorial search space, with over 10^{100} possible configurations, makes conventional black-box optimization infeasible. AMQ overcomes this challenge t…

Generalizing Behavior via Inverse Reinforcement Learning with Closed-Form Reward Centroids
Filippo Lazzati, Alberto Maria Metelli
https://arxiv.org/abs/2509.12010 https://

Generalizing Behavior via Inverse Reinforcement Learning with Closed-Form Reward Centroids
We study the problem of generalizing an expert agent's behavior, provided through demonstrations, to new environments and/or additional constraints. Inverse Reinforcement Learning (IRL) offers a promising solution by seeking to recover the expert's underlying reward function, which, if used for planning in the new settings, would reproduce the desired behavior. However, IRL is inherently ill-posed: multiple reward functions, forming the so-called feasible set, can explain the same observed beha…

Learning from Uncertain Similarity and Unlabeled Data
Meng Wei, Zhongnian Li, Peng Ying, Xinzheng Xu
https://arxiv.org/abs/2509.11984 https://arxiv.org/pdf…

Learning from Uncertain Similarity and Unlabeled Data
Existing similarity-based weakly supervised learning approaches often rely on precise similarity annotations between data pairs, which may inadvertently expose sensitive label information and raise privacy risks. To mitigate this issue, we propose Uncertain Similarity and Unlabeled Learning (USimUL), a novel framework where each similarity pair is embedded with an uncertainty component to reduce label leakage. In this paper, we propose an unbiased risk estimator that learns from uncertain simil…

Low-rank Orthogonalization for Large-scale Matrix Optimization with Applications to Foundation Model Training
Chuan He, Zhanwang Deng, Zhaosong Lu
https://arxiv.org/abs/2509.11983

Low-rank Orthogonalization for Large-scale Matrix Optimization with Applications to Foundation Model Training
Neural network (NN) training is inherently a large-scale matrix optimization problem, yet the matrix structure of NN parameters has long been overlooked. Recently, the optimizer Muon \cite{jordanmuon}, which explicitly exploits this structure, has gained significant attention for its strong performance in foundation model training. A key component contributing to Muon's success is matrix orthogonalization. In this paper, we propose {\it low-rank orthogonalization}, which explicitly leverages th…

Examining the Relationship between Scientific Publishing Activity and Hype-Driven Financial Bubbles: A Comparison of the Dot-Com and AI Eras
Aksheytha Chelikavada, Casey C. Bennett
https://arxiv.org/abs/2509.11982

Examining the Relationship between Scientific Publishing Activity and Hype-Driven Financial Bubbles: A Comparison of the Dot-Com and AI Eras
Financial bubbles often arrive without much warning, but create long-lasting economic effects. For example, during the dot-com bubble, innovative technologies created market disruptions through excitement for a promised bright future. Such technologies originated from research where scientists had developed them for years prior to their entry into the markets. That raises a question on the possibility of analyzing scientific publishing data (e.g. citation networks) leading up to a bubble for si…

MillStone: How Open-Minded Are LLMs?
Harold Triedman, Vitaly Shmatikov
https://arxiv.org/abs/2509.11967 https://arxiv.org/pdf/2509.11967

MillStone: How Open-Minded Are LLMs?
Large language models equipped with Web search, information retrieval tools, and other agentic capabilities are beginning to supplant traditional search engines. As users start to rely on LLMs for information on many topics, including controversial and debatable issues, it is important to understand how the stances and opinions expressed in LLM outputs are influenced by the documents they use as their information sources. In this paper, we present MillStone, the first benchmark that aims to s…

Deep operator network for surrogate modeling of poroelasticity with random permeability fields
Sangjoon Park, Yeonjong Shin, Jinhyun Choo
https://arxiv.org/abs/2509.11966 https:…

Deep operator network for surrogate modeling of poroelasticity with random permeability fields
Poroelasticity -- coupled fluid flow and elastic deformation in porous media -- often involves spatially variable permeability, especially in subsurface systems. In such cases, simulations with random permeability fields are widely used for probabilistic analysis, uncertainty quantification, and inverse problems. These simulations require repeated forward solves that are often prohibitively expensive, motivating the development of efficient surrogate models. However, efficient surrogate modelin…

TabStruct: Measuring Structural Fidelity of Tabular Data
Xiangjian Jiang, Nikola Simidjievski, Mateja Jamnik
https://arxiv.org/abs/2509.11950 https://arxiv…

TabStruct: Measuring Structural Fidelity of Tabular Data
Evaluating tabular generators remains a challenging problem, as the unique causal structural prior of heterogeneous tabular data does not lend itself to intuitive human inspection. Recent work has introduced structural fidelity as a tabular-specific evaluation dimension to assess whether synthetic data complies with the causal structures of real data. However, existing benchmarks often neglect the interplay between structural fidelity and conventional evaluation dimensions, thus failing to prov…

Transparent and Fair Profiling in Employment Services: Evidence from Switzerland
Tim R\"az
https://arxiv.org/abs/2509.11847 https://arxiv.org/pdf/2509…

Transparent and Fair Profiling in Employment Services: Evidence from Switzerland
Long-term unemployment (LTU) is a challenge for both jobseekers and public employment services. Statistical profiling tools are increasingly used to predict LTU risk. Some profiling tools are opaque, black-box machine learning models, which raise issues of transparency and fairness. This paper investigates whether interpretable models could serve as an alternative, using administrative data from Switzerland. Traditional statistical, interpretable, and black-box models are compared in terms of p…

FedDAF: Federated Domain Adaptation Using Model Functional Distance
Mrinmay Sen, Ankita Das, Sidhant Nair, C Krishna Mohan
https://arxiv.org/abs/2509.11819 https://

FedDAF: Federated Domain Adaptation Using Model Functional Distance
Federated Domain Adaptation (FDA) is a federated learning (FL) approach that improves model performance at the target client by collaborating with source clients while preserving data privacy. FDA faces two primary challenges: domain shifts between source and target data and limited labeled data at the target. Most existing FDA methods focus on domain shifts, assuming ample target data, yet often neglect the combined challenges of both domain shifts and data scarcity. Moreover, approaches that …

Collapse of Irrelevant Representations (CIR) Ensures Robust and Non-Disruptive LLM Unlearning
Filip Sondej, Yushi Yang
https://arxiv.org/abs/2509.11816 https://

Collapse of Irrelevant Representations (CIR) Ensures Robust and Non-Disruptive LLM Unlearning
Current unlearning techniques and safety training consistently fail to remove dangerous knowledge from language models. We analyze the root causes and propose a highly selective technique which unlearns robustly and without disrupting general performance. We perform PCA on activations and module output gradients to identify subspaces containing common representations, and collapse them before calculating unlearning updates. This way we avoid unlearning general representations, and only target…

Cost-aware Stopping for Bayesian Optimization
Qian Xie, Linda Cai, Alexander Terenin, Peter I. Frazier, Ziv Scully
https://arxiv.org/abs/2507.12453 https:/…

Cost-aware Stopping for Bayesian Optimization
In automated machine learning, scientific discovery, and other applications of Bayesian optimization, deciding when to stop evaluating expensive black-box functions is an important practical consideration. While several adaptive stopping rules have been proposed, in the cost-aware setting they lack guarantees ensuring they stop before incurring excessive function evaluation costs. We propose a cost-aware stopping rule for Bayesian optimization that adapts to varying evaluation costs and is free…

A Bayesian Incentive Mechanism for Poison-Resilient Federated Learning
Daniel Commey, Rebecca A. Sarpong, Griffith S. Klogo, Winful Bagyl-Bac, Garth V. Crosby
https://arxiv.org/abs/2507.12439

A Bayesian Incentive Mechanism for Poison-Resilient Federated Learning
Federated learning (FL) enables collaborative model training across decentralized clients while preserving data privacy. However, its open-participation nature exposes it to data-poisoning attacks, in which malicious actors submit corrupted model updates to degrade the global model. Existing defenses are often reactive, relying on statistical aggregation rules that can be computationally expensive and that typically assume an honest majority. This paper introduces a proactive, economic defense:…

Targeted Deep Architectures: A TMLE-Based Framework for Robust Causal Inference in Neural Networks
Yi Li, David Mccoy, Nolan Gunter, Kaitlyn Lee, Alejandro Schuler, Mark van der Laan
https://arxiv.org/abs/2507.12435

Targeted Deep Architectures: A TMLE-Based Framework for Robust Causal Inference in Neural Networks
Modern deep neural networks are powerful predictive tools yet often lack valid inference for causal parameters, such as treatment effects or entire survival curves. While frameworks like Double Machine Learning (DML) and Targeted Maximum Likelihood Estimation (TMLE) can debias machine-learning fits, existing neural implementations either rely on "targeted losses" that do not guarantee solving the efficient influence function equation or computationally expensive post-hoc "fluctuations" for mult…

Mixture of Raytraced Experts
Andrea Perin, Giacomo Lagomarsini, Claudio Gallicchio, Giuseppe Nuti
https://arxiv.org/abs/2507.12419 https://

Mixture of Raytraced Experts
We introduce a Mixture of Raytraced Experts, a stacked Mixture of Experts (MoE) architecture which can dynamically select sequences of experts, producing computational graphs of variable width and depth. Existing MoE architectures generally require a fixed amount of computation for a given sample. Our approach, in contrast, yields predictions with increasing accuracy as the computation cycles through the experts' sequence. We train our model by iteratively sampling from a set of candidate exper…

NOCTA: Non-Greedy Objective Cost-Tradeoff Acquisition for Longitudinal Data
Dzung Dinh, Boqi Chen, Marc Niethammer, Junier Oliva
https://arxiv.org/abs/2507.12412

NOCTA: Non-Greedy Objective Cost-Tradeoff Acquisition for Longitudinal Data
In many critical applications, resource constraints limit the amount of information that can be gathered to make predictions. For example, in healthcare, patient data often spans diverse features ranging from lab tests to imaging studies. Each feature may carry different information and must be acquired at a respective cost of time, money, or risk to the patient. Moreover, temporal prediction tasks, where both instance features and labels evolve over time, introduce additional complexity in dec…

ROC-n-reroll: How verifier imperfection affects test-time scaling
Florian E. Dorner, Yatong Chen, Andr\'e F. Cruz, Fanny Yang
https://arxiv.org/abs/2507.12399

ROC-n-reroll: How verifier imperfection affects test-time scaling
Test-time scaling aims to improve language model performance by leveraging additional compute during inference. While many works have empirically studied techniques like Best-of-N (BoN) and rejection sampling that make use of a verifier to enable test-time scaling, there is little theoretical understanding of how verifier imperfection affects performance. In this work, we address this gap. Specifically, we prove how instance-level accuracy of these methods is precisely characterized by the geom…

Trustworthy Tree-based Machine Learning by $MoS_2$ Flash-based Analog CAM with Inherent Soft Boundaries
Bo Wen, Guoyun Gao, Zhicheng Xu, Ruibin Mao, Xiaojuan Qi, X. Sharon Hu, Xunzhao Yin, Can Li
https://arxiv.org/abs/2507.12384

Trustworthy Tree-based Machine Learning by $MoS_2$ Flash-based Analog CAM with Inherent Soft Boundaries
The rapid advancement of artificial intelligence has raised concerns regarding its trustworthiness, especially in terms of interpretability and robustness. Tree-based models like Random Forest and XGBoost excel in interpretability and accuracy for tabular data, but scaling them remains computationally expensive due to poor data locality and high data dependence. Previous efforts to accelerate these models with analog content addressable memory (CAM) have struggled, due to the fact that the diff…

Improving Reinforcement Learning Sample-Efficiency using Local Approximation
Mohit Prashant, Arvind Easwaran
https://arxiv.org/abs/2507.12383 https://

Improving Reinforcement Learning Sample-Efficiency using Local Approximation
In this study, we derive Probably Approximately Correct (PAC) bounds on the asymptotic sample-complexity for RL within the infinite-horizon Markov Decision Process (MDP) setting that are sharper than those in existing literature. The premise of our study is twofold: firstly, the further two states are from each other, transition-wise, the less relevant the value of the first state is when learning the $ε$-optimal value of the second; secondly, the amount of 'effort', sample-complexity-wise, ex…

Heat Kernel Goes Topological
Maximilian Krahn, Vikas Garg
https://arxiv.org/abs/2507.12380 https://arxiv.org/pdf/2507.12380

Heat Kernel Goes Topological
Topological neural networks have emerged as powerful successors of graph neural networks. However, they typically involve higher-order message passing, which incurs significant computational expense. We circumvent this issue with a novel topological framework that introduces a Laplacian operator on combinatorial complexes (CCs), enabling efficient computation of heat kernels that serve as node descriptors. Our approach captures multiscale information and enables permutation-equivariant represen…

Nonlinear Concept Erasure: a Density Matching Approach
Antoine Saillenfest, Pirmin Lemberger
https://arxiv.org/abs/2507.12341 https://

Nonlinear Concept Erasure: a Density Matching Approach
Ensuring that neural models used in real-world applications cannot infer sensitive information, such as demographic attributes like gender or race, from text representations is a critical challenge when fairness is a concern. We address this issue through concept erasure, a process that removes information related to a specific concept from distributed representations while preserving as much of the remaining semantic information as possible. Our approach involves learning an orthogonal project…

Thought Purity: Defense Paradigm For Chain-of-Thought Attack
Zihao Xue, Zhen Bi, Long Ma, Zhenlin Hu, Yan Wang, Zhenfang Liu, Qing Sheng, Jie Xiao, Jungang Lou
https://arxiv.org/abs/2507.12314

Thought Purity: Defense Paradigm For Chain-of-Thought Attack
While reinforcement learning-trained Large Reasoning Models (LRMs, e.g., Deepseek-R1) demonstrate advanced reasoning capabilities in the evolving Large Language Models (LLMs) domain, their susceptibility to security threats remains a critical vulnerability. This weakness is particularly evident in Chain-of-Thought (CoT) generation processes, where adversarial methods like backdoor prompt attacks can systematically subvert the model's core reasoning mechanisms. The emerging Chain-of-Thought Atta…

PROL : Rehearsal Free Continual Learning in Streaming Data via Prompt Online Learning
M. Anwar Ma'sum, Mahardhika Pratama, Savitha Ramasamy, Lin Liu, Habibullah Habibullah, Ryszard Kowalczyk
https://arxiv.org/abs/2507.12305

PROL : Rehearsal Free Continual Learning in Streaming Data via Prompt Online Learning
The data privacy constraint in online continual learning (OCL), where the data can be seen only once, complicates the catastrophic forgetting problem in streaming data. A common approach applied by the current SOTAs in OCL is with the use of memory saving exemplars or features from previous classes to be replayed in the current task. On the other hand, the prompt-based approach performs excellently in continual learning but with the cost of a growing number of trainable parameters. The first ap…