Tootfinder

Russia keeps hitting Ukraine's training centers — and Ukraine doesn't have an answer: https://benborges.xyz/2025/10/18/russia-keeps-hitting-ukraines-training.html

„Only in the world of AI algorithm training can you claim that you were torrenting 2,400 porn videos for personal use and have that seem like the lesser of two evils.“ 😂
https://www.vice.com/en/article/meta-says-the-2400-ad…

Meta Says the 2,400 ‘Adult Movies’ They Torrented Were for Personal Use, Not Training AI
Two adult entertainment companies say Meta’s appetite for information has gotten a little kinky... and now they're suing.

Can an octopus learn to play a piano? This guy, Mattias Krantz, tried to find out. He rescues a random octopus from a fish market and then tries to work out how to teach it to play. You can skip the training to 15:20 to hear their first duet, but the training is interesting too. https://youtu.be/PcWnQ7fYzwI?si=BMFFE7

Israel-based Second Nature, whose conversational AI helps train sales teams, raised a $22M Series B led by Sienna VC, taking its total raised to $80M (Meir Orbach/CTech)
https://www.calcalistech.com/ctechnews/article/bj0cciralg

Zoom-backed Second Nature raises $22 million Series B to expand AI sales training platform
The Israeli startup uses conversational AI to transform how companies coach sales teams. 

🇺🇦 Auf radioeins läuft...
Doja Cat:
🎵 Aaahh Men!
#NowPlaying #DojaCat
https://thataquafinagirl.bandcamp.com/track/aaahh-men-x-training-season-doja-cat-x-dua-lipa-mashup
https://open.spotify.com/track/2qGvxAKXbajPFRb9ybIMOD
Schon gewusst❓
Wenn Du einen Beitrag des Bots ⭐ FAVORISIERST, erfährt das nur der Bot - nicht Deine FollowerInnen.
Wenn Du einen Song weitererempfehlen möchtest, dann 🔁 TEILE ihn❗

Russian missile strikes Ukrainian training ground, casualties reported, military says: https://benborges.xyz/2025/10/16/russian-missile-strikes-ukrainian-training.html

Standing atop a seven-ton military truck with his wife, Vance watched the largest Marine exercise in a decade on Oct. 18,
featuring F-18 and F-35 flyovers,
parachute landings,
Navy Seals swimming ashore,
offshore destroyers and amphibious ships,
simulated village explosions,
and MH-60 helicopters dropping additional Seals over the water.
-- all to distract attention from the massive "No Kings" protests happening all across the country.

Highway shuts down as Marines’ show sparks clash with Newsom
The Marines said in an earlier statement that the event would take place "on approved training ranges and comport with established safety protocols."

OpenAI launches the OpenAI Academy for News Organizations, which offers journalists AI training, open-source projects, and practical use cases (OpenAI)
https://openai.com/index/openai-academy-for-news-organizations

Short-term, the money may be in implantable visual prostheses for the blind. Longer-term, the money may be in training & support for noninvasive visual prostheses for the blind https://www.artificialvision.com/neuralink.htm Microsoft Windows vs Linux, Neuralink Blindsight vs The vOIC…

I’m in a work training for a agentic AI we will soon have access to. One use case that was suggested is researching and summarizing HR policy on a topic or for a question.
Oof. I expect it won’t be long before I read a news article where someone brings a wrongful termination suit based on either end of such a AI query.
#OfficeWorkerGripes

KI-Training: Normalsprachlicher Nutzungsvorbehalt reicht OLG Hamburg nicht
Das OLG in Hamburg hat im Fall der Klage eines Fotografen gegen den KI-Modell-Trainingsverein Laion die Berufung zurückgewiesen, eine Revision aber zugelassen.

Laminar: A Scalable Asynchronous RL Post-Training Framework
Guangming Sheng, Yuxuan Tong, Borui Wan, Wang Zhang, Chaobo Jia, Xibin Wu, Yuqi Wu, Xiang Li, Chi Zhang, Yanghua Peng, Haibin Lin, Xin Liu, Chuan Wu
https://arxiv.org/abs/2510.12633

Laminar: A Scalable Asynchronous RL Post-Training Framework
Reinforcement learning (RL) post-training for Large Language Models (LLMs) is now scaling to large clusters and running for extended durations to enhance model reasoning performance. However, the scalability of existing RL frameworks is limited, as extreme long-tail skewness in RL trajectory generation causes severe GPU underutilization. Current asynchronous RL systems attempt to mitigate this, but they rely on global weight synchronization between the actor and all rollouts, which creates a ri…

ViCO: A Training Strategy towards Semantic Aware Dynamic High-Resolution
Long Cui, Weiyun Wang, Jie Shao, Zichen Wen, Gen Luo, Linfeng Zhang, Yanting Zhang, Yu Qiao, Wenhai Wang
https://arxiv.org/abs/2510.12793

ViCO: A Training Strategy towards Semantic Aware Dynamic High-Resolution
Existing Multimodal Large Language Models (MLLMs) suffer from increased inference costs due to the additional vision tokens introduced by image inputs. In this work, we propose Visual Consistency Learning (ViCO), a novel training algorithm that enables the model to represent images of varying semantic complexities using different numbers of vision tokens. The key idea behind our method is to employ multiple MLP connectors, each with a different image compression ratio, to downsample the vision …

GOAT: A Training Framework for Goal-Oriented Agent with Tools
Hyunji Min, Sangwon Jung, Junyoung Sung, Dosung Lee, Leekyeung Han, Paul Hongsuck Seo
https://arxiv.org/abs/2510.12218

GOAT: A Training Framework for Goal-Oriented Agent with Tools
Large language models (LLMs) have recently been extended beyond traditional text generation to serve as interactive agents capable of using external tools based on user intent. However, current LLM agents still show limited ability to handle goal-oriented queries, which require decomposing a high-level objective into multiple interdependent API calls with correct planning and execution. Current approaches mainly rely on zero-shot evaluation due to the absence of training data. While proprietary…

Sources: Colts' Jones expected back for camp https://www.espn.com/nfl/story/_/id/47295871/sources-daniel-jones-expected-cleared-training-camp

GeoPipe: a Geo-distributed LLM Training Framework with enhanced Pipeline Parallelism in a Lossless RDMA-enabled Datacenter Optical Transport Network
Jun Dai, Xiaorun Wang, Kexiong Fang, Zheng Yang, Yuefeng Ji, Jiawei Zhang
https://arxiv.org/abs/2510.12064

GeoPipe: a Geo-distributed LLM Training Framework with enhanced Pipeline Parallelism in a Lossless RDMA-enabled Datacenter Optical Transport Network
The proliferation of Large Language Models (LLMs) with exponentially growing parameters is making cross-data center (DC) training an inevitable trend. However, viable strategies for extending single-DC training frameworks to multi-DC environments remain underdeveloped. We experimentally demonstrate, for the first time, a high-performance geo-distributed LLMs training framework across multiple DCs interconnected by a lossless, remote direct memory access (RDMA) enabled Datacenter Optical Transpo…

Stabilizing MoE Reinforcement Learning by Aligning Training and Inference Routers
Wenhan Ma, Hailin Zhang, Liang Zhao, Yifan Song, Yudong Wang, Zhifang Sui, Fuli Luo
https://arxiv.org/abs/2510.11370

Stabilizing MoE Reinforcement Learning by Aligning Training and Inference Routers
Reinforcement learning (RL) has emerged as a crucial approach for enhancing the capabilities of large language models. However, in Mixture-of-Experts (MoE) models, the routing mechanism often introduces instability, even leading to catastrophic RL training collapse. We analyze the training-inference consistency of MoE models and identify a notable discrepancy in routing behaviors between the two phases. Moreover, even under identical conditions, the routing framework can yield divergent expert …

SentinelLabs' Dakota Cary linked Yu Yang and Qiu Daibing, two alleged members of the Chinese state hacking group, to participants of the 2012 Cisco Networking Academy Cup.
https://www.theregister.com/2025/12/11/salt_typhoon_cisco_training/

Researcher claims Salt Typhoon spies attended Cisco training scheme
: Skills gained later fed Beijing's cyber operations, according to SentinelLabs expert

Sensor Tower: ChatGPT had 73M DAUs in India as of last week, up 607% YoY and more than double the US, boosted by its free plans, far above Gemini's 17M DAUs (Munsif Vengattil/Reuters)
https://www.reuters.com/world/india/with-free…

Drawing from a training with the Climate Psychology Alliance, I've been reflecting on the role of changemakers in our world.
Before, I saw activists as Sisyphus, straining to push the world uphill.
But what if we saw them as keeping the world steady?
https://www.brichapman.com/p/what-is-…

The Elephants Remember: On Activism, Memory, and Holding the World Steady
Not all activism is about pushing forward. Sometimes it’s about remembering, holding, and steadying what’s worth saving. A reflection on emergence, tenderness, and what the elephants can teach us about endurance.

Took a few minutes over lunch to iron out the last of the bugs in the happy path of PCIe link training.
There's no timeouts or fallback if there's problems but if the other side is happy, it will train up to L0 and then sit there ignoring all incoming traffic.
After a while, the link partner gets mad that it hasn't seen a single DLLP from me and drops the link. I don't implement recovery yet so things go downhill from there.

DCP: Addressing Input Dynamism In Long-Context Training via Dynamic Context Parallelism
Chenyu Jiang, Zhenkun Cai, Ye Tian, Zhen Jia, Yida Wang, Chuan Wu
https://arxiv.org/abs/2510.10620

DCP: Addressing Input Dynamism In Long-Context Training via Dynamic Context Parallelism
Context parallelism has emerged as a key technique to support long-context training, a growing trend in generative AI for modern large models. However, existing context parallel methods rely on static parallelization configurations that overlook the dynamic nature of training data, specifically, the variability in sequence lengths and token relationships (i.e., attention patterns) across samples. As a result, these methods often suffer from unnecessary communication overhead and imbalanced comp…

Head Coach Jim Harbaugh on Preparing for Dak Prescott and the Cowboys Offense https://www.chargers.com/video/jim-harbaugh-press-conference-wednesday-week-16-2025

🇺🇦 #NowPlaying on KEXP's #SonicReducer
Camouflage:
🎵 Powers of Horror / Future Soldier Training Program
#Camouflage
https://cam0uflage.bandcamp.com/track/powers-of-horror-future-soldier-training-program

Standing atop a seven-ton military truck with his wife, Vance watched the largest Marine exercise in a decade on Oct. 18,
featuring F-18 and F-35 flyovers,
parachute landings,
Navy Seals swimming ashore,
offshore destroyers and amphibious ships,
simulated village explosions,
and MH-60 helicopters dropping additional Seals over the water.
-- all to distract attention from the massive "No Kings" protests happening all across the country.

Highway shuts down as Marines’ show sparks clash with Newsom
The Marines said in an earlier statement that the event would take place "on approved training ranges and comport with established safety protocols."

Ich glaube, mein nächster Bike-Computer aka Fahrrad-Navi wird dieses von #Coros. Quasi unendliche Nutzung.
#dura

COROS DURA | Solar-GPS-Fahrradcomputer für Abenteuer und Ausdauertraining
Der COROS DURA ist dein idealer Begleiter für Training und Abenteuer auf dem Rad – mit bis zu 120 Stunden Akkulaufzeit, Solaraufladung, Turn-by-Turn-Navigation und einer durchdachten Trainingsplattform von Anfang bis Ziel.

The Holistic Horse Podcast
A podcast for those who want to learn all areas of horse care, training and management from a holistic perspective...
Great Australian Pods Podcast Directory: https://www.greataustralianpods.com/the-holistic-horse/

BrainForm: a Serious Game for BCI Training and Data Collection
Michele Romani, Devis Zanoni, Elisabetta Farella, Luca Turchet
https://arxiv.org/abs/2510.10169 https://

BrainForm: a Serious Game for BCI Training and Data Collection
$\textit{BrainForm}$ is a gamified Brain-Computer Interface (BCI) training system designed for scalable data collection using consumer hardware and a minimal setup. We investigated (1) how users develop BCI control skills across repeated sessions and (2) perceptual and performance effects of two visual stimulation textures. Game Experience Questionnaire (GEQ) scores for Flow}, Positive Affect, Competence and Challenge were strongly positive, indicating sustained engagement. A within-subject stu…

RE: https://mastodon.social/@nixCraft/115554484108496189
I still cannot comprehend how anyone could honestly consider a statistical model created from training data to be “intelligent”. Don't you remember the times when people knowingly smiled at anyo…

Long-term visual-to-tactile stimulation induces functional reorganization of thalamic pathways to achieve visual perception https://www.sciencedirect.com/science/article/pii/S105381192500655X using haptic sensory substitution;

User interface pretending to be human: fraud.
Statistical sentence generation described as “intelligence”: fraud.
Telling people, often children, to commit dangerous or criminal acts: all sorts of crimes.
Scientific papers generated, non-existing cases quoted at court, homework faked, artists ripped-off: fraud and copyright infringement.
Training based on unlicensed works: copyright infringement.
Data centers with tax incentives and special energy pricing: theft.
Promises to investors of AGI: fraud.
Government bailouts with taxpayer money: theft and fraud.
—
Media and politicians: This is innovation, don’t get left behind.

The very excellent “what happened last week” newsletter by @… focuses on “the A in AI stands for African” this week, namely on how Chinese AI companies are turning Kenya with its chronic unemployment of 67 percent into a hot spot of cheap AI labor.
More and more Kenyan students and recent graduates are hired to label thousands of videos a day through opaq…

The hidden Kenyan workers training China’s AI models
An unemployment crisis has created fertile ground for companies to step in with opaque systems built on WhatsApp groups, middlemen, and bargain-basement wages.

RE: #Microsoft Identity training!

Gradient-Guided Furthest Point Sampling for Robust Training Set Selection
Morris Trestman, Stefan Gugler, Felix A. Faber, O. A. von Lilienfeld
https://arxiv.org/abs/2510.08906 h…

Gradient-Guided Furthest Point Sampling for Robust Training Set Selection
Smart training set selections procedures enable the reduction of data needs and improves predictive robustness in machine learning problems relevant to chemistry. We introduce Gradient Guided Furthest Point Sampling (GGFPS), a simple extension of Furthest Point Sampling (FPS) that leverages molecular force norms to guide efficient sampling of configurational spaces of molecules. Numerical evidence is presented for a toy-system (Styblinski-Tang function) as well as for molecular dynamics traject…

Web Crawler Restrictions, AI Training Datasets \& Political Biases
Paul Bouchaud (ISC-PIF, m\'edialab), Pedro Ramaciotti (ISC-PIF, m\'edialab)
https://arxiv.org/abs/2510.09031

Web Crawler Restrictions, AI Training Datasets \& Political Biases
Large language models rely on web-scraped text for training; concurrently, content creators are increasingly blocking AI crawlers to retain control over their data. We analyze crawler restrictions across the top one million most-visited websites since 2023 and examine their potential downstream effects on training data composition. Our analysis reveals growing restrictions, with blocking patterns varying by website popularity and content type. A quarter of the top thousand websites restrict AI …

Advancing End-to-End Pixel Space Generative Modeling via Self-supervised Pre-training
Jiachen Lei, Keli Liu, Julius Berner, Haiming Yu, Hongkai Zheng, Jiahong Wu, Xiangxiang Chu
https://arxiv.org/abs/2510.12586

Advancing End-to-End Pixel Space Generative Modeling via Self-supervised Pre-training
Pixel-space generative models are often more difficult to train and generally underperform compared to their latent-space counterparts, leaving a persistent performance and efficiency gap. In this paper, we introduce a novel two-stage training framework that closes this gap for pixel-space diffusion and consistency models. In the first stage, we pre-train encoders to capture meaningful semantics from clean images while aligning them with points along the same deterministic sampling trajectory, …

VM-UNSSOR: Unsupervised Neural Speech Separation Enhanced by Higher-SNR Virtual Microphone Arrays
Shulin He, Zhong-Qiu Wang
https://arxiv.org/abs/2510.08914 https://

VM-UNSSOR: Unsupervised Neural Speech Separation Enhanced by Higher-SNR Virtual Microphone Arrays
Blind speech separation (BSS) aims to recover multiple speech sources from multi-channel, multi-speaker mixtures under unknown array geometry and room impulse responses. In unsupervised setup where clean target speech is not available for model training, UNSSOR proposes a mixture consistency (MC) loss for training deep neural networks (DNN) on over-determined training mixtures to realize unsupervised speech separation. However, when the number of microphones of the training mixtures decreases, …

Sharing this trick is my contribution to improve the quality of life of every corporate and government employee:
https://mastodon.cloud/@lrz/115350971038126826

A major southern California highway was being shut down while the U.S. Marine Corps stages a demonstration set to involve live fire on Oct. 18,
pitting the state's governor against the federal government yet again.
Interstate 5 will be shut down from Harbor Drive to Basilone Road,
a stretch of the main artery over 15 miles,
from 11 a.m. to 3 p.m. local time, the California Highway Patrol announced the morning of Oct. 18.
The news comes after days of back-an…

Highway shuts down as Marines’ show sparks clash with Newsom
The Marines said in an earlier statement that the event would take place "on approved training ranges and comport with established safety protocols."

Getting Your Indices in a Row: Full-Text Search for LLM Training Data for Real World
Ines Altemir Marinas, Anastasiia Kucherenko, Alexander Sternfeld, Andrei Kucharavy
https://arxiv.org/abs/2510.09471 …

Getting Your Indices in a Row: Full-Text Search for LLM Training Data for Real World
The performance of Large Language Models (LLMs) is determined by their training data. Despite the proliferation of open-weight LLMs, access to LLM training data has remained limited. Even for fully open LLMs, the scale of the data makes it all but inscrutable to the general scientific community, despite potentially containing critical data scraped from the internet. In this paper, we present the full-text indexing pipeline for the Apertus LLM training data. Leveraging Elasticsearch parallel i…

Parallel zum Schulschachkongress bieten wir den Schülern und Schülerinnen aller Schulen des Rhein-Kreises-Neuss ein einmaliges kostenfreies Training an. Weitere Informationen und die Anmeldung unter: https://www.deutsche-schachjugend.de/termine/2025/traini…

A headline emphasising the government's support for Britain's High Streets.
> Thousands to get free digital training so everyone has the chance to shop around for cheaper deals online
https://www.gov.uk/government…

Thousands to get free digital training so everyone has the chance to shop around for cheaper deals online
Free support to get people online is rolling out in communities in every region of the country.

Training Dynamics Impact Post-Training Quantization Robustness
Albert Catalan-Tatjer, Niccol\`o Ajroldi, Jonas Geiping
https://arxiv.org/abs/2510.06213 https://

Training Dynamics Impact Post-Training Quantization Robustness
While post-training quantization is widely adopted for efficient deployment of large language models, the mechanisms underlying quantization robustness remain unclear. We conduct a comprehensive analysis of quantization degradation across open-source language model training trajectories up to 32B parameters and 15T training tokens to accurately assess the relationship between training dynamics and quantization performance. Our key finding is that quantization errors in large-scale training runs…

Next Interest Flow: A Generative Pre-training Paradigm for Recommender Systems by Modeling All-domain Movelines
Chen Gao, Zixin Zhao, Lv Shao, Tong Liu
https://arxiv.org/abs/2510.11317

Next Interest Flow: A Generative Pre-training Paradigm for Recommender Systems by Modeling All-domain Movelines
Click-Through Rate (CTR) prediction, a cornerstone of modern recommender systems, has been dominated by discriminative models that react to past user behavior rather than proactively modeling user intent. Existing generative paradigms attempt to address this but suffer from critical limitations: Large Language Model (LLM) based methods create a semantic mismatch by forcing e-commerce signals into a linguistic space, while ID-based generation is constrained by item memorization and cold-start is…

Training Models to Detect Successive Robot Errors from Human Reactions
Shannon Liu, Maria Teresa Parreira, Wendy Ju
https://arxiv.org/abs/2510.09080 https://

Training Models to Detect Successive Robot Errors from Human Reactions
As robots become more integrated into society, detecting robot errors is essential for effective human-robot interaction (HRI). When a robot fails repeatedly, how can it know when to change its behavior? Humans naturally respond to robot errors through verbal and nonverbal cues that intensify over successive failures-from confusion and subtle speech changes to visible frustration and impatience. While prior work shows that human reactions can indicate robot failures, few studies examine how the…

Understanding the Efficacy of Phishing Training in Practice
"Combined with the bulk of empirical evidence from other studies involving
real-world, controlled experiments, our results suggests that organizations should not expect large anti-phishing benefits from either annual security awareness training or embedded phishing as commonly deployed today."
In addition, the overall cost on third-party organisations doing incident response should not exclude the impact of false-posi…

VideoTIM5 tactile display from ABTIM https://abtim.com/vt5-vt3 "new Optacon", around € 9000, or € 11000 including training https://www.

Cocoon: A System Architecture for Differentially Private Training with Correlated Noises
Donghwan Kim, Xin Gu, Jinho Baek, Timothy Lo, Younghoon Min, Kwangsik Shin, Jongryool Kim, Jongse Park, Kiwan Maeng
https://arxiv.org/abs/2510.07304

Cocoon: A System Architecture for Differentially Private Training with Correlated Noises
Machine learning (ML) models memorize and leak training data, causing serious privacy issues to data owners. Training algorithms with differential privacy (DP), such as DP-SGD, have been gaining attention as a solution. However, DP-SGD adds a noise at each training iteration, which degrades the accuracy of the trained model. To improve accuracy, a new family of approaches adds carefully designed correlated noises, so that noises cancel out each other across iterations. We performed an extensive…

Interviews with copywriters on generative AI's impact: their work used for training, layoffs, wages and rates in free fall, freelancers losing clients, and more (Brian Merchant/Blood in the Machine)
https://www.bloodinthemachine.com/p/i-was-forced-to-use-ai-until-the…

"I was forced to use AI until the day I was laid off." Copywriters reveal how AI has decimated their industry
Copywriters were one of the first to have their jobs targeted by AI firms. These are their stories, three years into the AI era.

PromptFlow: Training Prompts Like Neural Networks
Jingyi Wang, Hongyuan Zhu, Ye Niu, Yunhui Deng
https://arxiv.org/abs/2510.12246 https://arxiv.org/pdf/251…

PromptFlow: Training Prompts Like Neural Networks
Large Language Models (LLMs) have demonstrated profound impact on Natural Language Processing (NLP) tasks. However, their effective deployment across diverse domains often require domain-specific adaptation strategies, as generic models may underperform when faced with specialized data distributions. Recent advances in prompt engineering (PE) offer a promising alternative to extensive retraining by refining input instructions to align LLM outputs with task objectives. This paradigm has emerged …

Universal Music lässt Musik zum Training von KI verwenden
Universal Music öffnet die Tür für das Training Künstlicher Intelligenz mit Werken seiner Künstler. Eine Klage gegen den Dienst Udio wird beigelegt.
https…

Architecture Induces Structural Invariant Manifolds of Neural Network Training Dynamics
Jiajie Zhao, Tao Luo, Yaoyu Zhang
https://arxiv.org/abs/2510.09564 https://

Architecture Induces Structural Invariant Manifolds of Neural Network Training Dynamics
While architecture is recognized as key to the performance of deep neural networks, its precise effect on training dynamics has been unclear due to the confounding influence of data and loss functions. This paper proposed an analytic framework based on the geometric control theory to characterize the dynamical properties intrinsic to a model's parameterization. We prove that the Structural Invariant Manifolds (SIMs) of an analytic model $F(\mathbfθ)(\mathbf{x})$--submanifolds that confine grad…

ACE-G: Improving Generalization of Scene Coordinate Regression Through Query Pre-Training
Leonard Bruns, Axel Barroso-Laguna, Tommaso Cavallari, \'Aron Monszpart, Sowmya Munukutla, Victor Adrian Prisacariu, Eric Brachmann
https://arxiv.org/abs/2510.11605

ACE-G: Improving Generalization of Scene Coordinate Regression Through Query Pre-Training
Scene coordinate regression (SCR) has established itself as a promising learning-based approach to visual relocalization. After mere minutes of scene-specific training, SCR models estimate camera poses of query images with high accuracy. Still, SCR methods fall short of the generalization capabilities of more classical feature-matching approaches. When imaging conditions of query images, such as lighting or viewpoint, are too different from the training views, SCR models fail. Failing to genera…

Automated Evolutionary Optimization for Resource-Efficient Neural Network Training
Ilia Revin, Leon Strelkov, Vadim A. Potemkin, Ivan Kireev, Andrey Savchenko
https://arxiv.org/abs/2510.09566

Automated Evolutionary Optimization for Resource-Efficient Neural Network Training
There are many critical challenges in optimizing neural network models, including distributed computing, compression techniques, and efficient training, regardless of their application to specific tasks. Solving such problems is crucial because the need for scalable and resource-efficient models is increasing. To address these challenges, we have developed a new automated machine learning (AutoML) framework, Parameter Efficient Training with Robust Automation (PETRA). It applies evolutionary op…

MaP: A Unified Framework for Reliable Evaluation of Pre-training Dynamics
Jiapeng Wang, Changxin Tian, Kunlong Chen, Ziqi Liu, Jiaxin Mao, Wayne Xin Zhao, Zhiqiang Zhang, Jun Zhou
https://arxiv.org/abs/2510.09295

MaP: A Unified Framework for Reliable Evaluation of Pre-training Dynamics
Reliable evaluation is fundamental to the progress of Large Language Models (LLMs), yet the evaluation process during pre-training is plagued by significant instability that obscures true learning dynamics. In this work, we systematically diagnose this instability, attributing it to two distinct sources: \textit{Parameter Instability} from training stochasticity and \textit{Evaluation Instability} from noisy measurement protocols. To counteract both sources of noise, we introduce \textbf{MaP}, …

Andrej Karpathy unveils nanochat, a full-stack training and inference implementation of an LLM in a single, dependency-minimal codebase (Andrej Karpathy/@karpathy)
https://x.com/karpathy/status/1977755427569111362

Andrej Karpathy (@karpathy) on X
Excited to release new repo: nanochat! (it's among the most unhinged I've written). Unlike my earlier similar repo nanoGPT which only covered pretraining, nanochat is a minimal, from scratch, full-stack training/inference pipeline of a simple ChatGPT clone in a single,

EmboMatrix: A Scalable Training-Ground for Embodied Decision-Making
Zixing Lei, Sheng Yin, Yichen Xiong, Yuanzhuo Ding, Wenhao Huang, Yuxi Wei, Qingyao Xu, Yiming Li, Weixin Li, Yunhong Wang, Siheng Chen
https://arxiv.org/abs/2510.12072

EmboMatrix: A Scalable Training-Ground for Embodied Decision-Making
Embodied decision-making enables agents to translate high-level goals into executable actions through continuous interactions within the physical world, forming a cornerstone of general-purpose embodied intelligence. Large language models (LLMs), with their general decision-making capabilities, offer a promising path to realize this potential; however, LLMs trained solely on language lack exposure to physical environments, limiting their true embodied understanding. To bridge this gap, we propo…

PCIe any% speedrun update: LTSSM is in full swing, I've made it all the way to Configuration.LinkWidth.Accept (and improved the ngscopeclient decode for link training in the process).

🇺🇦 Auf radioeins läuft...
Doja Cat:
🎵 Aaahh Men!
#NowPlaying #DojaCat
https://thataquafinagirl.bandcamp.com/track/aaahh-men-x-training-season-doja-cat-x-dua-lipa-mashup
https://open.spotify.com/track/2qGvxAKXbajPFRb9ybIMOD

Unión del Barrio has been doing “know your rights” workshops
and training people on how to spot ICE and report to the community when they do see them.
Cardona said it is important to document ICE raids and kidnappings to help people understand what they have been seeing and the way ICE goes about these operations.
Advocates have also been organizing patrols and responding to reported ICE activity.
Cardona said there have been many instances where community members in L…

Racial Profiling Is ICE’s New Norm. Activists Are Mobilizing in Response.
Arizonans long endured grim racial profiling under SB 1070. Now the Supreme Court has greenlighted its use nationwide.

An Explorative Study on Distributed Computing Techniques in Training and Inference of Large Language Models
Sheikh Azizul Hakim, Saem Hasan
https://arxiv.org/abs/2510.11211 http…

An Explorative Study on Distributed Computing Techniques in Training and Inference of Large Language Models
Large language models (LLM) are advanced AI systems trained on extensive textual data, leveraging deep learning techniques to understand and generate human-like language. Today's LLMs with billions of parameters are so huge that hardly any single computing node can train, fine-tune, or infer from them. Therefore, several distributed computing techniques are being introduced in the literature to properly utilize LLMs. We have explored the application of distributed computing techniques in LLMs f…

Für KI-Training: US-App zahlt Geld für abgehörte Telefonate
15 bis 30 Cent pro Minute zahlt die Neon App dafür, dass sie Telefongespräche mitschneidet. Die Aufnahmen sollen für KI-Training weiterverkauft werden.

A Framework for Low-Effort Training Data Generation for Urban Semantic Segmentation
Denis Zavadski, Damjan Kal\v{s}an, Tim K\"uchler, Haebom Lee, Stefan Roth, Carsten Rother
https://arxiv.org/abs/2510.11567

A Framework for Low-Effort Training Data Generation for Urban Semantic Segmentation
Synthetic datasets are widely used for training urban scene recognition models, but even highly realistic renderings show a noticeable gap to real imagery. This gap is particularly pronounced when adapting to a specific target domain, such as Cityscapes, where differences in architecture, vegetation, object appearance, and camera characteristics limit downstream performance. Closing this gap with more detailed 3D modelling would require expensive asset and scene design, defeating the purpose of…

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training
Yunxiang Zhang, Muhammad Khalifa, Lechen Zhang, Xin Liu, Ayoung Lee, Xinliang Frederick Zhang, Farima Fatahi Bayat, Lu Wang
https://arxiv.org/abs/2510.09354

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training
Large reasoning models exhibit long chain-of-thought reasoning with strategies such as backtracking and self-correction, though recent studies suggest that these abilities typically require additional training. We first investigate whether such behaviors can be elicited without any training. To this end, we propose a decoding-time approach, ThinkLogit, which utilizes logit arithmetic to tune a target large non-reasoning model for long reasoning using a substantially smaller reasoning model as t…

QB Kyler Murray on Cardinals' plans for new 2028 training facility: 'Hopefully I'm here to see it' https://www.nfl.com/news/qb-kyler-murray-on-cardinals-plans-for-new-2028-training-facility-hopefully-i-m-here-t…

Interviews with copywriters on generative AI's impact: job losses to AI, work used for training, falling wages and rates, freelancers losing clients, and more (Brian Merchant/Blood in the Machine)
https://www.bloodinthemachine.com/p/i-was-forced-to-use-ai-until-the

"I was forced to use AI until the day I was laid off." Copywriters reveal how AI has decimated their industry
Copywriters were one of the first to have their jobs targeted by AI firms. These are their stories, three years into the AI era.

ThinkPilot: Steering Reasoning Models via Automated Think-prefixes Optimization
Sunzhu Li, Zhiyu Lin, Shuling Yang, Jiale Zhao, Wei Chen
https://arxiv.org/abs/2510.12063 https:/…

ThinkPilot: Steering Reasoning Models via Automated Think-prefixes Optimization
Large Reasoning Models (LRMs) are powerful, but they still suffer from inefficient and off-target reasoning. Currently, training-free methods are limited to either rigid heuristics or descriptive, non-actionable analyses. In this paper, we introduce ThinkPilot, a training-free framework that automatically optimizes LRMs reasoning. It uses an evolutionary process to generate think-prefixes, which are instructions that evolve driven by a taxonomy of reasoning behaviors to guide models toward supe…

DiT360: High-Fidelity Panoramic Image Generation via Hybrid Training
Haoran Feng, Dizhe Zhang, Xiangtai Li, Bo Du, Lu Qi
https://arxiv.org/abs/2510.11712 https://

DiT360: High-Fidelity Panoramic Image Generation via Hybrid Training
In this work, we propose DiT360, a DiT-based framework that performs hybrid training on perspective and panoramic data for panoramic image generation. For the issues of maintaining geometric fidelity and photorealism in generation quality, we attribute the main reason to the lack of large-scale, high-quality, real-world panoramic data, where such a data-centric view differs from prior methods that focus on model design. Basically, DiT360 has several key modules for inter-domain transformation a…

New military training center for Ukrainian, NATO troops opens in Poland: https://benborges.xyz/2025/10/03/new-military-training-center-for.html

Early Detection and Reduction of Memorisation for Domain Adaptation and Instruction Tuning
Dean L. Slack, Noura Al Moubayed
https://arxiv.org/abs/2510.11372 https://

Early Detection and Reduction of Memorisation for Domain Adaptation and Instruction Tuning
Although large language models excel across many tasks, they can memorise training data and thereby expose private or copyrighted text. Most defences target the pre-training stage, leaving memorisation during fine-tuning, especially for domain adaptation and instruction tuning, poorly understood. We fine-tune Pythia, Llama3, and Mistral models spanning 1.4B-70B parameters on common evaluation datasets and track verbatim memorisation throughout training. We find that memorisation increases drama…

Representation-Based Exploration for Language Models: From Test-Time to Post-Training
Jens Tuyls, Dylan J. Foster, Akshay Krishnamurthy, Jordan T. Ash
https://arxiv.org/abs/2510.11686

Representation-Based Exploration for Language Models: From Test-Time to Post-Training
Reinforcement learning (RL) promises to expand the capabilities of language models, but it is unclear if current RL techniques promote the discovery of novel behaviors, or simply sharpen those already present in the base model. In this paper, we investigate the value of deliberate exploration -- explicitly incentivizing the model to discover novel and diverse behaviors -- and aim to understand how the knowledge in pre-trained models can guide this search. Our main finding is that exploration wi…

PCIe-from-scratch update: GTP is configured and sending a hardcoded 8b10b junk pattern to the SoC.
On the RX side, I'm able to correctly identify and parse TS1/TS2 training sets.
Next step will be to refactor a bit, then implement the actual LTSSM and start sending training sets.

EU will DSGVO schleifen – für KI und Cookie-Banner
Der von der EU-Kommission geplante "digitale Omnibus" würde bestehende Datenschutzrechte aufweichen. Es geht etwa um Cookies und das Training von KI-Systemen.
h…

The Allen Institute of AI launches Bolmo 7B and Bolmo 1B, claiming they are "the first fully open byte-level language models", built on its Olmo 3 models (Emilia David/VentureBeat)
https://venturebeat.com/ai/bolmos-architecture-unlock…

LayerSync: Self-aligning Intermediate Layers
Yasaman Haghighi, Bastien van Delft, Mariam Hassan, Alexandre Alahi
https://arxiv.org/abs/2510.12581 https://a…

LayerSync: Self-aligning Intermediate Layers
We propose LayerSync, a domain-agnostic approach for improving the generation quality and the training efficiency of diffusion models. Prior studies have highlighted the connection between the quality of generation and the representations learned by diffusion models, showing that external guidance on model intermediate representations accelerates training. We reconceptualize this paradigm by regularizing diffusion models with their own intermediate representations. Building on the observation t…

Interactive Training: Feedback-Driven Neural Network Optimization
Wentao Zhang, Yang Young Lu, Yuntian Deng
https://arxiv.org/abs/2510.02297 https://arxiv.…

Interactive Training: Feedback-Driven Neural Network Optimization
Traditional neural network training typically follows fixed, predefined optimization recipes, lacking the flexibility to dynamically respond to instabilities or emerging training issues. In this paper, we introduce Interactive Training, an open-source framework that enables real-time, feedback-driven intervention during neural network training by human experts or automated AI agents. At its core, Interactive Training uses a control server to mediate communication between users or agents and the…

Varjo erneuert Premium-XR-Headsets für hochrealistische Flugsimulatoren
Varjo aktualisiert seine Mixed-Reality-Headsets für Simulatoren. Verbesserte Kameras und Eye-Tracking sollen das Training genauer und effizienter machen.

Mid-Training of Large Language Models: A Survey
Kaixiang Mo, Yuxin Shi, Weiwei Weng, Zhiqiang Zhou, Shuman Liu, Haibo Zhang, Anxiang Zeng
https://arxiv.org/abs/2510.06826 https:…

Mid-Training of Large Language Models: A Survey
Large language models (LLMs) are typically developed through large-scale pre-training followed by task-specific fine-tuning. Recent advances highlight the importance of an intermediate mid-training stage, where models undergo multiple annealing-style phases that refine data quality, adapt optimization schedules, and extend context length. This stage mitigates diminishing returns from noisy tokens, stabilizes convergence, and expands model capability in late training. Its effectiveness can be ex…

Data or Language Supervision: What Makes CLIP Better than DINO?
Yiming Liu, Yuhui Zhang, Dhruba Ghosh, Ludwig Schmidt, Serena Yeung-Levy
https://arxiv.org/abs/2510.11835 https:/…

Data or Language Supervision: What Makes CLIP Better than DINO?
CLIP outperforms self-supervised models like DINO as vision encoders for vision-language models (VLMs), but it remains unclear whether this advantage stems from CLIP's language supervision or its much larger training data. To disentangle these factors, we pre-train CLIP and DINO under controlled settings -- using the same architecture, dataset, and training configuration -- achieving similar ImageNet accuracy. Embedding analysis shows that CLIP captures high-level semantics (e.g., object catego…

Anthropic now requires users to accept new privacy terms by October 8, including choosing whether new chats and coding sessions can be used to train AI models (Reece Rogers/Wired)
https://www.wired.com/story/anthropic-using-claude-chats-for-training-how-…

Anthropic Will Use Claude Chats for Training Data. Here’s How to Opt Out
Anthropic is starting to train its models on new Claude chats. If you’re using the bot and don’t want your chats used as training data, here’s how to opt out.

Training-Free Time Series Classification via In-Context Reasoning with LLM Agents
Songyuan Sui, Zihang Xu, Yu-Neng Chuang, Kwei-Herng Lai, Xia Hu
https://arxiv.org/abs/2510.05950

Training-Free Time Series Classification via In-Context Reasoning with LLM Agents
Time series classification (TSC) spans diverse application scenarios, yet labeled data are often scarce, making task-specific training costly and inflexible. Recent reasoning-oriented large language models (LLMs) show promise in understanding temporal patterns, but purely zero-shot usage remains suboptimal. We propose FETA, a multi-agent framework for training-free TSC via exemplar-based in-context reasoning. FETA decomposes a multivariate series into channel-wise subproblems, retrieves a few s…

From Condensation to Rank Collapse: A Two-Stage Analysis of Transformer Training Dynamics
Zheng-An Chen, Tao Luo
https://arxiv.org/abs/2510.06954 https://a…

From Condensation to Rank Collapse: A Two-Stage Analysis of Transformer Training Dynamics
Although transformer-based models have shown exceptional empirical performance, the fundamental principles governing their training dynamics are inadequately characterized beyond configuration-specific studies. Inspired by empirical evidence showing improved reasoning capabilities under small initialization scales in language models, we employ the gradient flow analytical framework established in [Zhou et al. NeurIPS 2022] to systematically investigate linearized Transformer training dynamics. …

A political candidate in the New York City suburbs went for a night swim in the Atlantic Ocean this past spring -- and never returned.
Petros Krommidas’s phone, keys and clothes were found on the sands at Long Beach on Long Island.
The 29-year-old former Ivy League rower, who was training for a triathlon, had parked his car just off the picturesque wooden boardwalk.
As the months passed, local Democrats attempted to field a replacement to run for the seat in the Nassau Coun…

MetaVLA: Unified Meta Co-training For Efficient Embodied Adaption
Chen Li, Zhantao Yang, Han Zhang, Fangyi Chen, Chenchen Zhu, Anudeepsekhar Bolimera, Marios Savvides
https://arxiv.org/abs/2510.05580

MetaVLA: Unified Meta Co-training For Efficient Embodied Adaption
Vision-Language-Action (VLA) models show promise in embodied reasoning, yet remain far from true generalists-they often require task-specific fine-tuning, and generalize poorly to unseen tasks. We propose MetaVLA, a unified, backbone-agnostic post-training framework for efficient and scalable alignment. MetaVLA introduces Context-Aware Meta Co-Training, which consolidates diverse target tasks into a single fine-tuning stage while leveraging structurally diverse auxiliary tasks to improve in-dom…

Microsofts Gaming Copilot macht Screenshots – aber nicht für KI-Training
Auf Vorwürfe, dass PC-Spieler mit der neuen Funktion überwacht würden, hat Microsoft nun reagiert. Die Antworten sind jedoch nicht ganz eindeutig.

Peloton unveils its Cross Training Series, including a $1,695 Bike and $6,695 Tread Plus, and AI-based Peloton IQ to track workouts on new and old machines (Victoria Song/The Verge)
https://www.theverge.com/tech/789282/peloton-cross-training-…

Peloton ushers in a new era with revamped hardware and AI
Peloton just announced a complete product lineup refresh, adding swivel displays, upgraded specs, and the Peloton IQ AI into every model.

Cross-Receiver Generalization for RF Fingerprint Identification via Feature Disentanglement and Adversarial Training
Yuhao Pan, Xiucheng Wang, Nan Cheng, Wenchao Xu
https://arxiv.org/abs/2510.09405

Cross-Receiver Generalization for RF Fingerprint Identification via Feature Disentanglement and Adversarial Training
Radio frequency fingerprint identification (RFFI) is a critical technique for wireless network security, leveraging intrinsic hardware-level imperfections introduced during device manufacturing to enable precise transmitter identification. While deep neural networks have shown remarkable capability in extracting discriminative features, their real-world deployment is hindered by receiver-induced variability. In practice, RF fingerprint signals comprise transmitter-specific features as well as c…

The only remaining criminal case against Donald Trump has been revived
after the head of Georgia’s prosecutor’s council appointed himself to replace Fani Willis,
the Fulton county district attorney, who was removed from the election interference case in September.
Pete Skandalakis,
a Republican and the executive director of the prosecuting attorneys’ council of Georgia,
the state body that provides legal training and is often charged to mitigate prosecutorial co…

NaViL: Rethinking Scaling Properties of Native Multimodal Large Language Models under Data Constraints
Changyao Tian, Hao Li, Gen Luo, Xizhou Zhu, Weijie Su, Hanming Deng, Jinguo Zhu, Jie Shao, Ziran Zhu, Yunpeng Liu, Lewei Lu, Wenhai Wang, Hongsheng Li, Jifeng Dai
https://arxiv.org/abs/2510.08565 …

NaViL: Rethinking Scaling Properties of Native Multimodal Large Language Models under Data Constraints
Compositional training has been the de-facto paradigm in existing Multimodal Large Language Models (MLLMs), where pre-trained vision encoders are connected with pre-trained LLMs through continuous multimodal pre-training. However, the multimodal scaling property of this paradigm remains difficult to explore due to the separated training. In this paper, we focus on the native training of MLLMs in an end-to-end manner and systematically study its design space and scaling property under a practica…

Information-Theoretic Policy Pre-Training with Empowerment
Moritz Schneider, Robert Krug, Narunas Vaskevicius, Luigi Palmieri, Michael Volpp, Joschka Boedecker
https://arxiv.org/abs/2510.05996

Information-Theoretic Policy Pre-Training with Empowerment
Empowerment, an information-theoretic measure of an agent's potential influence on its environment, has emerged as a powerful intrinsic motivation and exploration framework for reinforcement learning (RL). Besides for unsupervised RL and skill learning algorithms, the specific use of empowerment as a pre-training signal has received limited attention in the literature. We show that empowerment can be used as a pre-training signal for data-efficient downstream task adaptation. For this we extend…

Beyond Turn Limits: Training Deep Search Agents with Dynamic Context Window
Qiaoyu Tang, Hao Xiang, Le Yu, Bowen Yu, Yaojie Lu, Xianpei Han, Le Sun, WenJuan Zhang, Pengbo Wang, Shixuan Liu, Zhenru Zhang, Jianhong Tu, Hongyu Lin, Junyang Lin
https://arxiv.org/abs/2510.08276

Beyond Turn Limits: Training Deep Search Agents with Dynamic Context Window
While recent advances in reasoning models have demonstrated cognitive behaviors through reinforcement learning, existing approaches struggle to invoke deep reasoning capabilities in multi-turn agents with long-horizon interactions. We propose DeepMiner, a novel framework that elicits such abilities by introducing high-difficulty training tasks and dynamic context window. DeepMiner presents a reverse construction method to generate complex but verifiable question-answer pairs from authentic web …

Hugging Face details how it used its new tool, Skills, to fine tune LLMs using Claude, including for writing scripts, submitting jobs to cloud GPUs, and more (Hugging Face)
https://huggingface.co/blog/hf-skills-training

We Got Claude to Fine-Tune an Open Source LLM
We’re on a journey to advance and democratize artificial intelligence through open source and open science.

Out-of-Distribution Detection from Small Training Sets using Bayesian Neural Network Classifiers
Kevin Raina, Tanya Schmah
https://arxiv.org/abs/2510.06025 https://

Out-of-Distribution Detection from Small Training Sets using Bayesian Neural Network Classifiers
Out-of-Distribution (OOD) detection is critical to AI reliability and safety, yet in many practical settings, only a limited amount of training data is available. Bayesian Neural Networks (BNNs) are a promising class of model on which to base OOD detection, because they explicitly represent epistemic (i.e. model) uncertainty. In the small training data regime, BNNs are especially valuable because they can incorporate prior model information. We introduce a new family of Bayesian posthoc OOD sco…

Localist LLMs -- A Mathematical Framework for Dynamic Locality Control
Joachim Diederich
https://arxiv.org/abs/2510.09338 https://arxiv.org/pdf/2510.09338

Localist LLMs -- A Mathematical Framework for Dynamic Locality Control
We present a novel framework for training large language models with continuously adjustable internal representations that span the full spectrum from localist (interpretable, rule-based) to distributed (generalizable, efficient) encodings. The key innovation is a locality dial, a tunable parameter that dynamically controls the degree of localization during both training and inference without requiring model retraining. This is achieved through group sparsity penalties on attention mechanisms, …

Chip giants' efforts to turn Phoenix into a US hub may hinge on training local workers; an estimated 115K local chip jobs are set to be created in four years (Peter S. Goodman/New York Times)
https://www.nytimes.com/2025/12/04/business/tsmc-arizona-workers-…

A Growing U.S. Tech Hub Needs Workers. Colleges Try to Keep Up.
The success of efforts to turn Phoenix into a dominant center of semiconductor manufacturing may hinge on efforts to train local workers.

Draft, Verify, and Improve: Toward Training-Aware Speculative Decoding
Shrenik Bhansali, Larry Heck
https://arxiv.org/abs/2510.05421 https://arxiv.org/pdf/…

Draft, Verify, and Improve: Toward Training-Aware Speculative Decoding
Autoregressive (AR) decoding is a major latency bottleneck for large language models. Speculative decoding (SD) accelerates AR by letting a drafter propose multi-token blocks that a verifier accepts or rejects. However, many SD systems require heavy offline training or extra components. These choices raise data/compute cost and can yield brittle drafters under distribution drift. We introduce \emph{Draft, Verify, \& Improve (DVI)}, a training-aware self-speculative framework that combines infer…

Micro1, which helps AI labs find experts for data annotation, says it has crossed $100M in annualized revenue and fielded investment offers at a $2.5B valuation (Anna Tong/Forbes)
http://www.forbes.com/sites/annatong/2025

This 24 Year Old Built A Multibillion-Dollar AI Training Empire In Eight Months
Ali Ansari’s decision to turn micro1’s AI recruitment assistant into a data labeling business spiked the company’s valuation from $80 million to $2.5 billion.

Anker paid users of its Eufy security cameras $2 per video of staged or real package and car thefts to train its AI systems from December 2024 to February 2025 (Lorenzo Franceschi-Bicchierai/TechCrunch)
https://techcrunch.com/2025/10/04/anke

Anker offered Eufy camera owners $2 per video for AI training | TechCrunch
Hundreds of Eufy customers have donated hundreds of thousands of videos to train the company’s AI systems.

Fisher Information, Training and Bias in Fourier Regression Models
Lorenzo Pastori, Veronika Eyring, Mierk Schwabe
https://arxiv.org/abs/2510.06945 https://

Fisher Information, Training and Bias in Fourier Regression Models
Motivated by the growing interest in quantum machine learning, in particular quantum neural networks (QNNs), we study how recently introduced evaluation metrics based on the Fisher information matrix (FIM) are effective for predicting their training and prediction performance. We exploit the equivalence between a broad class of QNNs and Fourier models, and study the interplay between the \emph{effective dimension} and the \emph{bias} of a model towards a given task, investigating how these affe…

Unified Molecule Pre-training with Flexible 2D and 3D Modalities: Single and Paired Modality Integration
Tengwei Song, Min Wu, Yuan Fang
https://arxiv.org/abs/2510.07035 https:/…

Unified Molecule Pre-training with Flexible 2D and 3D Modalities: Single and Paired Modality Integration
Molecular representation learning plays a crucial role in advancing applications such as drug discovery and material design. Existing work leverages 2D and 3D modalities of molecular information for pre-training, aiming to capture comprehensive structural and geometric insights. However, these methods require paired 2D and 3D molecular data to train the model effectively and prevent it from collapsing into a single modality, posing limitations in scenarios where a certain modality is unavailabl…