Strumień wydarzeń
Wszystko, co pojawiło się w monitorowanych źródłach – od najnowszych.
-
CLOSURE: Cross-Channel Closure for Multivariate Time Series Anomaly Detection
[Sztuczna inteligencja] Autorzy: Fan Xu, Hao Wu, Zhi Song i 5 więcej Multivariate time series produced by industrial installations and server clusters are under continuous monitoring, and anomaly detection is the first line of defense before a local fault escalates into a system-wide failure. In the absence of labels, the dominant paradigm trains a reconstruction model on normal data and scores the residual to reach a verdict. Recent work has steadily enriched the expressiveness of the recovery map through Transformers, diffusion models, and cross-scale associations. A residual, however, only certifies that an observation can be recovered; it does not certify that the variables still explain one another as they do under normal operation. A coordinated yet unfamiliar operating transition inflates the residual without breaking any relationship, whereas a sensor drifting away from its peers may leave the residual silent. Methods that fold inter-variable dep…
-
Multi-Objective Reinforcement Learning Using Routed Ensembles and Trajectory Attribution
[Uczenie maszynowe] Autorzy: Zhenshuo Zhang, Ruoxuan Xiong, Hongyang R. Zhang Multi-objective reinforcement learning is a fundamental problem for modern AI, with applications spanning post-training, reasoning models, and robotic control. A central challenge is to learn policies that balance competing objectives, such as helpfulness versus honesty in safety and alignment, or speed versus energy in robotics. The optimal policy can vary sharply with objective-weighting combinations, and the extent of this variation depends crucially on the degree of conflict among objectives, making per-preference policy learning sample-inefficient. In this paper, we introduce objective-conflict coefficients and derive near-optimal regret bounds for online settings using this measure, thereby quantifying the impact of misalignment on learning efficiency. Then, we design routed ensemble policy learning, an algorithm that maintains a small set of basis policies and dynamically routes …
-
A Category-Selective Encryption Framework for Privacy Protection of UAV Images
[Kryptografia i bezpieczeństwo] Autorzy: Ting Yan, Minghui Shang, Haolang Sun i 3 więcej Imagery captured by unmanned aerial vehicle (UAV) systems increasingly raises privacy concerns, as UAV aerial surveillance routinely captures sensitive objects such as pedestrians and vehicles alongside ordinary scene content. Existing privacy-preserving approaches typically employ class-agnostic saliency detection to indiscriminately obfuscate all visually salient regions, leading to unnecessary information loss and hindering downstream scene analysis. This paper proposes a category-selective encryption framework for privacy-preserving UAV surveillance. The core advance over prior class-agnostic approaches is semantic-level selectivity: the system first recognizes what is present in the scene via an 80-class detector, then encrypts only user-specified sensitive categories while leaving non-sensitive objects and background intact for downstream analytics. The framework supports two comp…
-
The Watchtower Imperative: Lifespan-Calibrated Evaluation for Frontier AI
[cs.CY] Autorzy: Casey M. Williams, Jennifer Victoria Scurrell The gap between AI capability deployment and safety assurance is structural and cannot be closed through internal evaluations alone. Analyzing documented harms (2017-2026) across the human lifespan reveals that disparate incidents --- from voice assistants endangering children to AI companions implicated in adolescent suicide and automated care denials for the elderly --- reflect a single failure: evaluating systems against an implicit adult user model. We introduce five developmental moderators structured into a Lifespan Vulnerability Matrix across six life stages. The matrix reveals a nonmonotonic risk distribution: vulnerability peaks at both ends of the lifespan, governance coverage is asymmetrical, and authority to convert model errors into real-world impact is scrutinized least. We then propose a set of lifespan design principles generalizing existing child-protective instruments,…
-
Can JEV Support Clinical Trial Criteria Assessment?
[Uczenie maszynowe] Autorzy: Mengxian He, Ziyue Gao, Chun Li i 4 więcej Clinical trial prescreening requires distinguishing unmet conditions from insufficient evidence in patient records. Given patient records, criterion questions, and predefined answer options, we retrospectively compared JEV, a non-generative decision model returning native option probabilities, with schema-constrained Qwen Flash across 2,048 test patient-condition pairs from TrialGPT and question-answer datasets derived from MIMIC-III and MIMIC-IV, all evaluated under strict Zero Data Retention (ZDR) policies to ensure clinical data privacy. Both models assessed conditions individually or in batches sharing the same patient context, with performance measured against reference labels. In batch mode, JEV and Qwen achieved similar macro-F1 scores on TrialGPT (0.652 vs. 0.660) and MIMIC-IV (0.807 vs. 0.814), whereas JEV substantially outperformed Qwen on MIMIC-III (0.768 vs. 0.534; $Δ$ = +0.…
-
Audit Where Hacks Pay: Budgeted Strong-Verifier Audits Against Reward Hacking in GRPO
[Uczenie maszynowe] Autorzy: Shurun Li, Michael Hua Wang, Haibo Zeng i 1 więcej Reinforcement learning with verifiable rewards trains on cheap checks, such as unit tests, that a policy can learn to game, while a stronger check is too expensive to run on every rollout. We study how to spend a limited strong-verifier budget in GRPO, separating which weakly passing rollouts to audit from what reward a caught hack receives. At a fixed policy state with centered advantages, the expected update depends on a design only through each rollout's marginal audit probability and correction; reflecting a caught hack's reward about the failure reward negates the advantages of an all-hack group; and auditing whole groups with the fewest weak passes first (low-pass budgeted audits, LB) minimizes the advantage mass left on hacks among pass-count designs with a fixed correction when a label's value falls with the pass count. We tested this analysis in actual LLM training: from each r…
-
Quantitative Generalization Bounds for Multi-Layer Transformers: A Few Long Strings can Beat Many Short Ones
[Uczenie maszynowe] Autorzy: Marius Henry, Michael Rizvi-Martel, Michael Hahn Transformer-based language models exhibit an astonishing capacity to generalize to longer and more difficult problem instances than those seen in training. However, theoretical understanding of this phenomenon still lags behind. Unlike most machine learning models, the generalisation abilities of transformers inherently require considering two axes. On one hand, one must examine in-distribution statistical generalisation, which concerns the ability of a model to broaden a function from training samples to novel samples of the same distribution. On the other hand, given the sequential nature of language data, one must also consider length generalisation, which reflects whether a model trained on inputs of a given length can extrapolate to longer ones.These two notions have mostly been studied separately, leaving open how they interact. In this work, we give a joint treatment of both axes…
-
Concerning behavior from a popular open-weight AI model
314 punktów · 123 komentarze · autor matthieu_bl · cohere.com
-
Explainable Multimodal AI Approach to Reduce Unnecessary Breast Biopsies
[Sztuczna inteligencja] Autorzy: Sadam Hussain, Mansoor Ali, Daly Betzabeth Avendaño-Avalos i 1 więcej Recently, there has been rise in multimodal vision-language models for computer-aided diagnosis (CAD), yet their application in breast diagnosis, particularly for predicting biopsy outcomes remains limited. This study presents a multimodal vision-language approach, incorporating Grad-CAM as an Explainable AI (XAI) technique, to predict benign tumors in patients recommended for biopsy, especially those classified as BI-RADS 4 and 5. A novel multimodal framework was devel- oped using Vision Transformer (ViT) for imaging feature extraction and DeBERTa-V3-Large for textual embeddings. The model was trained and evaluated on a comprehensive in-house dataset comprising 1,632 patients, each with four mammography views and a corresponding radiology report. The features of both modalities were extracted independently and fused by late fusion before classification. Grad-CAM was used to generate vi…
-
Show HN: Hob - Open-source, on device alternative to Claude Cowork and ChatGPT
157 punktów · 35 komentarzy · autor pancomplex · github.com
-
CSS-Tricks has a new owner
526 punktów · 109 komentarzy · autor edent · css-tricks.com
-
Microsoft offers mass buyouts for roughly a quarter of its workforce
997 punktów · 1015 komentarzy · autor doctorzook · www.seattletimes.com
-
Solar power is becoming bigger than nuclear
459 punktów · 647 komentarzy · autor geox · www.bbc.com
-
Sebastian Raschka's from-scratch video course on building reasoning models
736 punktów · 102 komentarze · autor sgiannopoulos · magazine.sebastianraschka.com
-
Magnituda 4,6 — 154 km NW of Port-Olry, Vanuatu
Głębokość 10 km.
-
DOJ says presidents should be able to sell secrets on eBay
192 punktów · 90 komentarzy · autor _____k · freedom.press
-
OLED iPad Mini
158 punktów · 3 komentarze · autor etothet · www.apple.com
-
New iPad mini
222 punktów · 539 komentarzy · autor soheilpro · www.apple.com
-
Anthropic S-1
1 636 punktów · 1558 komentarzy · autor SiNTEx · www.sec.gov
-
Fixing Tool-Call Accuracy in vLLM
162 punktów · 68 komentarzy · autor coffeemug · www.ubicloud.com
-
Brown student maps the hidden global labor powering AI
179 punktów · 171 komentarzy · autor cdrnsf · www.brown.edu
-
Peter Thiel, the tech oligarch who believes in the Antichrist
159 punktów · 214 komentarzy · autor geox · english.elpais.com
-
Magnituda 5,2 — 205 km NW of Port-Olry, Vanuatu
Głębokość 10 km.
-
Zdjęcie dnia: NASA Science
The Great Nebula in Orion, an immense, nearby starbirth region, is one of the most famous of all astronomical nebulas. Here, glowing gas surrounds hot young stars at the edge of an immense interstellar molecular cloud only 1500 light-years away. In the featured deep image in assigned colors highlighted by emission in hydrogen, wisps and sheets of dust and gas are particularly evident. The Great Nebula in Orion can be found with the unaided eye near the easily identifiable belt of three stars in the popular constellation Orion. In addition to housing a bright open cluster of stars known as the Trapezium, the Orion Nebula contains many stellar nurseries. These nurseries contain much hydrogen gas, hot young stars, proplyds, and stellar jets spewing material at high speeds. Also known as M42, the Orion Nebula spans about 40 light years and is located in the same spiral arm of our Galaxy as the Sun.Tomorrow's picture: open space
-
Ceny mieszkań 26.10.2026: mediana 11 700 zł/m²
Próba: 411 ofert od 17 deweloperów. Kraków: 15 600 zł/m² · Mińsk Mazowiecki: 12 700 zł/m² · Warszawa: 12 385 zł/m² · Jurczyce: 10 582 zł/m² · Dawidy Bankowe: 9 779 zł/m² · Bochnia: 9 500 zł/m² · Ostrów Wielkopolski: 9 000 zł/m² · Cieszyn: 8 671 zł/m² · Staszów: 7 950 zł/m² · Przasnysz: 7 628 zł/m² · Gołaszyn: 5 251 zł/m²
-
Magnituda 4,6 — 81 km SSE of Akutan, Alaska
Głębokość 35 km.
-
Magnituda 4,8 — 87 km SSE of Akutan, Alaska
Głębokość 3 km.
-
Magnituda 5,0 — 218 km NW of Port-Olry, Vanuatu
Głębokość 10 km.
-
Pharo 14
260 punktów · 71 komentarzy · autor xkriva11 · pharo.org
-
Magnituda 4,8 — South Sandwich Islands region
Głębokość 173 km. Zgłoszeń od świadków: 0.