The post Boosting the Western Balkans’ digital transformation: The key role of the private sector appeared first on Atlantic Council.

Some of the baby chicks were still chirping after they were raked in and buried alive beneath compost. That sentence is difficult to write. It’s even more difficult to imagine or witness. Yet it is one of the many realities Mercy For Animals documented during our latest undercover investigation at a USDA-certified organic chicken farm that supplies Costco. Our undercover investigation documented chickens beaten with sticks and repeatedly bludgeoned with a pipe, injured animals left to languish among the bodies of their dead flockmates, and birds collapsing under their own weight, who could no longer walk or even reach food and water at a North Carolina farm raising them for Costco. Our investigator also documented baby chicks crushed to death and buried alive. The newly released footage uncovers birds bred to grow unnaturally fast, biosecurity failures linked to potential public health risks, and conditions inconsistent with USDA organic requirements, even though these chickens were marketed to Costco shoppers as “certified organic.” I’ve seen undercover investigations for many years, and this footage was still difficult to watch. Not because these abuses are unprecedented, but because it continues to happen behind labels that millions of people trust. Animals and consumers deserve better. Most Costco members that I know—especially those who seek out the “organic” label—would likely be shocked to learn animals suffered this way in a supply chain connected to a company they trust and rely on. No animal should be beaten, crushed, buried alive, or left to die in agony because a company failed to prevent this cruelty. The farm Mercy For Animals investigated is estimated to confine up to 520,000 chickens at a time and produce roughly 3.2 million pounds of chicken each year for a dedicated Costco supplier. But this investigation is about more than one farm. It exposes a food system that too often prioritizes production and profit over animals’ well-being. Costco was warned. Costco promised. Yet the cruelty continued. This is the second time in five years that Mercy For Animals has exposed cruelty connected to Costco’s chicken supply chain. After our 2021 investigation drew national attention, Costco publicly committed to improving animal welfare. Yet our latest investigation found the company had not adopted meaningful, enforceable animal welfare standards. Across its supply chain, Costco has the opportunity to reduce suffering for more than 250 million chickens annually by implementing proven measures, including slower-growing breeds, improved living conditions, stronger oversight, and greater transparency. Together, we can hold Costco accountable. Costco now has another opportunity to do what it promised years ago: implement meaningful, enforceable animal welfare standards, increase transparency, and help prevent this cruelty from happening again. Your voice is animals’ chance for change. Tell Costco: No more baby chicks buried alive. Demand an end to the abuse in its chicken supply chain. Animals cannot wait for another empty promise, and neither will we. The post New Investigation Reveals What’s Behind Costco’s Organic Chicken appeared first on Mercy For Animals.

The post Why GCC integration matters more than ever appeared first on Atlantic Council.

The post How to reset US-Gulf security partnerships appeared first on Atlantic Council.

arXiv:2510.22899v2 Announce Type: replace Abstract: We investigate the role of network architecture in shaping the inductive biases of modern score-based generative models. To this end, we introduce the Score Anisotropy Directions (SADs), architecture-dependent directions that reveal how different networks preferentially capture data structure. Our analysis suggests that SADs form adaptive bases aligned with the architecture's output geometry, providing a principled way to predict generalization ability in score models prior to training. Through both synthetic data and standard image benchmarks, we demonstrate that SADs reliably capture fine-grained model behavior and correlate with downstream performance, as measured by Wasserstein metrics. Our work offers a new lens for explaining and predicting directional biases of generative models.

arXiv:2511.04949v2 Announce Type: replace-cross Abstract: Rapid advances in generative AI have led to increasingly realistic deepfakes, posing growing challenges for law enforcement and public trust. Existing passive deepfake detectors struggle to keep pace, largely due to their dependence on specific forgery artifacts, which limits their ability to generalize to new deepfake types. Proactive deepfake detection using watermarks has emerged to address the challenge of identifying high-quality synthetic media. However, these methods often struggle to balance robustness against benign distortions with sensitivity to malicious tampering. This paper introduces a novel deep learning framework that harnesses high-dimensional latent space representations and the Adversarial Reinforcement Learning (ARL) paradigm to develop a robust and adaptive watermarking approach. Specifically, we develop a learnable watermark embedder that operates in the latent space, capturing high-level image semantics, while offering precise control over message encoding and extraction. The ARL paradigm empowers the learnable watermarking module to pursue an optimal balance between robustness and fragility. This is achieved through interaction with a dynamic curriculum of benign and malicious image manipulations simulated by an adversarial attacker agent. Comprehensive evaluations on the CelebA and CelebA-HQ benchmarks reveal that our method consistently outperforms state-of-the-art approaches, achieving improvements of over 4.5% on CelebA and more than 5.3% on CelebA-HQ under challenging manipulation scenarios.

The post Daniel B. Shapiro testifies to House Foreign Affairs Committee on wartime prospects for regional integration appeared first on Atlantic Council.

The post The US Navy’s decision to reclassify 19 Virginia-class submarines is a breakthrough. Here’s why. appeared first on Atlantic Council.

arXiv:2607.23198v3 Announce Type: replace Abstract: We present Variance-Preserving Orthogonal Selection (VPOS), an unsupervised feature-selection method that performs sequential orthogonal deflation in the variance-weighted principal component analysis (PCA) loading space $\mathbf{V}_d\mathbf{\Lambda}_d^{1/2}$. After each feature is selected, its loading direction is projected out of all remaining candidates, so subsequent selections cover complementary directions of the rank-$d$ covariance approximation while returning original variables. We establish rank-reduction guarantees and a determinant-growth interpretation, and distinguish VPOS from greedy selection on raw data, unweighted eigenvector pivoting, Principal Feature Analysis (PFA), and Principal Variable Selection (PVS). Experiments enforce $k\leq d$, tune method-specific parameters on validation observations, and evaluate on unseen outer folds. Across seven labelled benchmarks, VPOS improves held-out normalised reconstruction error over matched PCA without deflation on every dataset, with reductions of 1--78%. It obtains the lowest mean reconstruction error on Wine, Breast Cancer, and MNIST and is within 1.7% of the lowest error on CIFAR-10 and HighDim. On CIFAR-10, VPOS is approximately 24$\times$ faster than the closely related PVS baseline while incurring a 1.7% reconstruction gap. These results establish VPOS as an efficient covariance-coverage method, particularly when correlated high-dimensional data must be represented by a small set of identifiable original variables.

The post Havrylov in Ukrinform: For the first time in Ukraine’s history, we are launching the licensed production of a defense system on Ukrainian soil appeared first on Atlantic Council.

Subscribe The Mask Supply Chain! Manufacturers of any medical device are the first ones to get the blame when a patient complains of suffering injury or harm from using a medical device. How does this relate to the Medical Device Supply Chain? When filing a lawsuit for medical device liability due to defects, the lawyer only has to prove that the defect was present at the time immediately after the medical device or tool was manufactured. Because of this manufacturers must prioritize clinical investigations of the medical devices they are producing. But today’s manufacturers outsource most of the inputs and parts of their final manufactured medical devices and products. Some manufacturers have appealed for higher transparency so the burden of liability would be shared across the supply chain. To improve the transparency of where they get the inputs and parts of their final products, companies are looking for preferences for outsourcing medical device manufacturing requirements and should also consider implementing mechanisms. Here are some of the things manufacturers can do to increase transparency in your medical device supply chain. 1. FDA Mandated Unique Device Identification (UDI) One of the ways by which medical device manufacturers and their suppliers can increase transparency in the medical device supply chain is by using the Unique Device Identification (UDI) system. To improve their standards of transparency, stakeholders and players can benefit from it in the medical device manufacturing, supply, and distribution industry. It would help improve the package labeling and tracking of the various medical devices and the parts used to make them. 2. Mandated By FDA The UDI is a system mandated by the U.S. Food and Drug Administration (FDA) as guidance for players in the medical device industry on the labeling and tracking of their products. It was issued in response to the growing importance of tracking people responsible for designing, manufacturing, distributing, and utilizing these various medical devices. It is significant because when a patient gets hurt or suffers injury resulting in the use of the medical device, liability sometimes is imposed on the manufacturer of the device. The UDI Rule was started by FDA in 2013 to develop regulatory mechanisms, track down medical device products, and enhance product safety for patients. It outlined specific requirements and standards on how medical devices should be labeled. The 2014 Rule was for class III devices. The 2015 Rule was for life-sustaining and life support products that can be embedded in patients. The 2016 Rule was for class II medical devices. The deadline for enforcement of acquiescing with the requirements for class III and class II devices has already lapsed. In compliance with the enforcement regulations of these products was prioritized because they’re considered to have the highest level of risk to patient safety. 3. Class I Extended To 2022 The deadline for complying with the enforcement rules for class I medical devices and unclassified devices was extended till the 22nd of September 2022. Class I medical devices were made last priority because they have been categorized as having the lowest level of risk to patient safety. Class I covers a broad scope of various medical device products. Among the products in this category are braces, bandages, and minor surgical tools and instruments. The FDA has a database containing a comprehensive list of the various kinds of medical devices, tools, and instruments. This database is called the Global Unique Device Identification Database (GUDID). According to their type and use, the medical devices contained in the list have been classified by the FDA. Medical device manufacturers can use the FDA database to search for the kind of product and its classification. They can type in the generic name of the medical device to find out the correct classification type of the medical device or instrument. It is very beneficial in increasing transparency in the industry because even manufacturers sometimes can’t agree on the classification of medical instruments and devices. The FDA UDI rule mandates that UDI should be placed on the labels and packages of all medical equipment, devices, tools, implants, and instruments. This UDI should have two parts, so the first part is the ‘device identifier’ or DI. The second part is the ‘production identifier’ or PI. The FDA has also required a standard format for the date which will appear on the label. The date should appear as (YYYY-MM-DD). Also, the FDA rule requires that the UDI should be marked in formats that can be read by humans as well as machines. There are medical devices, tools, and instruments used more than once such as surgical clamps, forceps, tweezers, and tongs. The UDI should be marked directly on these instruments and not on their packaging or boxes. 4. Manufacturer’s Product Information Management (PIM) Another way to meet the medical technology challenge of improving transparency in the supply chain is through the implementation of its Product Information Management (PIM) system and platform across the company’s sourcing, procurement, manufacturing, packaging, sales, and distribution workflows processes. A PIM would require all suppliers and providers to assign unique markers, bar codes, labels, RFID, QR codes, or other identification methods for every part or product they supply. The manufacturing company may come up with its PIM system with its identification, attribution, tracing, tracking, and monitoring methods. It would be of great help in enabling the manufacturer to instantly identify and recognize the manufacturer or supplier of each part of their medical devices, wearable therapeutics and products. They can require all their suppliers and providers to follow the manufacturer’s PIM system if they wish to continue getting contracts. The other way is for the manufacturer to permit their suppliers and providers to come up with their product identification system for all the products, parts, and supplies that the manufacturer sources from each of them. The supplier or provider’s PIM system may be encoded and integrated into the manufacturer’s ERP modules and applications suites for procurement, production, sales, distribution, and logistics. It would enable the manufacturer to track the use and location of each part provided by a third-party supplier or provider, even if the specific manufacturer’s product has already been sold to its eventual user. Manufacturers can even make use of their PIM to increase their engagement with their institutional clients, along with individual buyers. They can upload information about the various parts and products incorporated into their final manufactured medical devices and products. It would greatly increase the company’s transparency about the input materials and parts to their products. It would also help customers who have inquiries or complaints about the quality or safety of the manufacturer’s medical device products. Conclusion The penalties for actual and punitive damages imposed on manufacturers of medical devices for product defects can sometimes amount to tens or hundreds of millions of dollars. It can be unfair at times if the product defect can be traced to a defect in a part that was outsourced to a third-party supplier or vendor. It has led some manufacturers to appeal for greater transparency in the medical device supply chain. They’ve asked for distribution of the risks and liability for product defects that cause injuries to the general public. Medical Device Supply Chain article and permission to publish here provided by Claire Glassman. Originally written for Supply Chain Game Changer and published on February 3, 2022. Cover photo by sandro porfirio on Unsplash. The post How to Increase Transparency in the Medical Device Supply Chain! appeared first on Supply Chain Game Changer™.

arXiv:2608.03228v2 Announce Type: replace Abstract: Existing low rank KV cache methods preserve either model weights or key variance, neither of which directly reflects the attention scores used during inference. We derive the expected attention score distortion caused by rank r key compression and show that it yields a covariance weighted low rank objective. Under a margin condition, controlling this distortion also improves top k recall. The optimal rank r solution has a closed form asymmetric factorization obtained from the SVD of the covariance weighted query key operator. This motivates SAKI, a training free KV cache index that directly preserves attention scores rather than key reconstruction quality. Across LLaMA 3.1 8B, Qwen 2.5 7B, Mistral 7B v0.1, and Llama 3.2 3B, SAKI outperforms key PCA at every tested rank. At rank 32, it removes 13 to 30 percent of PCA's remaining top 64 recall error, including improvements from 0.748 to 0.799 on LLaMA 3.1 8B and from 0.786 to 0.850 on Qwen 2.5 7B. It improves 68 to 89 percent of attention heads per model, with the largest gains in deeper layers. Predicted score MSE reductions closely match empirical measurements, with a Pearson correlation of 0.997, while ablation studies confirm that the gains arise from optimizing the attention score objective rather than covariance weighting alone. Analysis of the scoring operator further explains why weight only, invariant subspace, and key reconstruction methods can be suboptimal. SAKI uses random-matrix theory to separate genuine covariance signal from autocorrelated sampling noise, matching PCA with only 512 calibration tokens and adding value exactly where PCA sees no reliable signal.

Bleach Thousand Year Blood War anime season 1 episode 42, titled ‘Son of Darkness’, continued...

Election 2026 News

arXiv:2602.08290v2 Announce Type: replace-cross Abstract: In federated learning (FL), decentralized model training allows multi-ple participants to collaboratively improve a shared machine learning model without exchanging raw data. However, ensuring the integrity and reliability of the system is challenging due to the presence of potentially malicious or faulty nodes that can degrade the model's performance. This paper proposes a novel trust-based incentive mechanism designed to evaluate and reward the quality of contributions in FL systems. By dynamically assessing trust scores based on fac-tors such as data quality, model accuracy, consistency, and contribution fre-quency, the system encourages honest participation and penalizes unreliable or malicious behavior. These trust scores form the basis of an incentive mechanism that rewards high-trust nodes with greater participation opportunities and penal-ties for low-trust participants. We further explore the integration of blockchain technology and smart contracts to automate the trust evaluation and incentive distribution processes, ensuring transparency and decentralization. Our proposed theoretical framework aims to create a more robust, fair, and transparent FL eco-system, reducing the risks posed by untrustworthy participants.

Welcome to the steamy side of ANN's Spring Manga Guide.

arXiv:2510.05589v3 Announce Type: replace-cross Abstract: Effective time series forecasting enables various real-world applications, benefiting from the proliferation of mobile devices. However, the volume of time series data may vary significantly across domains due to high data acquisition costs and data regulations. To maximally create value from sparse data, this study focuses on a new problem of source-free time series forecasting, aiming to adapt a pretrained model from sufficient source time series to the sparse target time series without access to the source data, enabling data protection. To achieve this, we propose TimeID, a novel source-free time series forecasting framework with a large language model (LLM) centric proxy denoising inspired by the powerful generalization capabilities of LLMs. Specifically, TimeID consists of three key components: (1) dual-branch invariant disentangled feature learning that enforces representation- and gradient-wise invariance by means of season-trend decomposition; (2) lightweight, parameter-free proxy denoising that dynamically calibrates systematic biases of LLMs; and (3) knowledge distillation that bidirectionally aligns the denoised prediction and the original target prediction. Extensive experiments on real-world datasets demonstrate that TimeID outperforms state-of-the-art baselines, improving MSE and MAE by 10.7% and 9.3% on average. The code is available at https://github.com/decisionintelligence/TimeID.

arXiv:2607.27594v3 Announce Type: replace Abstract: Post-training alignment in large reasoning models (LRMs) has significantly improved their adaptability to diverse safety compliance settings. However, as LRMs personalization for downstream users takes center stage, the demand for varying levels of policy compliance grows as different user-specific LRMs must adhere to distinct subsets of safety policies. Training a separate LRM for each policy subset introduces severe combinatorial overhead. While in context learning methods overcome this combinatorial overhead, they introduce additional computational challenges associated with long context generation. To address this challenge, we propose \ours, a unified adaptive hypernetwork-based framework for multi-policy compliance. In our framework, safety policies serve as customizable inputs to a LoRA adapter generator, which learns to produce policy compliant LoRA weights for downstream LRM. When added to the LRM these weights enable the generation of responses compliant with the specified policy subsets. In this work, we demonstrate that training such a hypernetwork enables on-demand policy adjustments on a single LRM without sacrificing task performance across reasoning models of different sized and different evaluation datasets. This highlights the effectiveness and practicality of adaptive hypernetwork based alignment in LRMs.