We present variational inference with sequential sample-average approximations (VISA), a method for approximate inference in computationally intensive models, such as those based on numerical simulations. VISA extends importance-weighted forward-KL variational inference by employing a sequence of sample-average approximations, which are considered valid inside a trust region. This makes it possible to reuse model evaluations across multiple gradient steps, thereby reducing computational cost. We perform experiments on high-dimensional Gaussians, Lotka-Volterra dynamics, and a Pickover attractor, which demonstrate that VISA can achieve comparable approximation accuracy to standard importance-weighted forward-KL variational inference with computational savings of a factor two or more for conservatively chosen learning rates.

Vision-Language Models (VLMs) remain limited in spatial reasoning tasks that require multi-view understanding and embodied perspective shifts. Recent approaches such as MindJourney attempt to mitigate this gap through test-time scaling where a world model imagines action-conditioned trajectories and a heuristic verifier selects helpful views from such trajectories. In this work, we systematically examine how such test-time verifiers behave across benchmarks, uncovering both their promise and their pitfalls. Our uncertainty-based analyses show that MindJourney's verifier provides little meaningful calibration, and that random scoring often reduces answer entropy equally well, thus exposing systematic action biases and unreliable reward signals. To mitigate these, we introduce a Verification through Spatial Assertions (ViSA) framework that grounds the test-time reward in verifiable, frame-anchored micro-claims. This principled verifier consistently improves spatial reasoning on the SAT-Real benchmark and corrects trajectory-selection biases through more balanced exploratory behavior. However, on the challenging MMSI-Bench, none of the verifiers, including ours, achieve consistent scaling, suggesting that the current world models form an information bottleneck where imagined views fail to enrich fine-grained reasoning. Together, these findings chart the bad, good, and ugly aspects of test-time verification for world-model-based reasoning. Our code is available at https://github.com/chandar-lab/visa-for-mindjourney.

We perform experiments on high-dimensional Gaussians, Lotka-V olterra dynamics, and a Pickover attractor.

Some migrants coming to the UK will need to speak English to an A-level standard under tougher new rules set to be introduced by the government. Applicants will be tested in person on their speaking, listening, reading and writing at Home Office-approved providers, with their results checked as part of the visa process. The changes, which come into force from 8 January 2026, form part of wider plans to cut levels of immigration to the UK outlined in a white paper in May. Home Secretary Shabana Mahmood said: If you come to this country, you must learn our language and play your part. Those applying for skilled worker, scale-up and high potential individual (HPI) visas will be required to reach B2 level - a step up from the current B1 standard which is equivalent to GCSE.

At least 60 people have been killed in a drone strike at a displacement shelter in el-Fasher, a besieged Sudanese city on the brink of collapse. The resistance committee for el-Fasher, made up of local citizens and activists, said the paramilitary Rapid Support Forces (RSF) hit Dar al-Arqam camp, located within a university, with two drone strikes and eight artillery shells. Children, women and the elderly were killed in cold blood, and many were completely burned, a statement from the group said. Eyewitnesses described scenes of panic as rescuers pulled bodies from the rubble. Hospitals already struggling under months of siege have been overwhelmed, with doctors treating the wounded on floors and in corridors.

We perform experiments on high-dimensional Gaussians, Lotka-V olterra dynamics, and a Pickover attractor.

Zero-shot anomaly detection (ZSAD) aims to detect anomalies without any target domain training samples, relying solely on external auxiliary data. Existing CLIP-based methods attempt to activate the model's ZSAD potential via handcrafted or static learnable prompts. The former incur high engineering costs and limited semantic coverage, whereas the latter apply identical descriptions across diverse anomaly types, thus fail to adapt to complex variations. Furthermore, since CLIP is originally pretrained on large-scale classification tasks, its anomaly segmentation quality is highly sensitive to the exact wording of class names, severely constraining prompting strategies that depend on class labels. To address these challenges, we introduce ViP$^{2}$-CLIP. The key insight of ViP$^{2}$-CLIP is a Visual-Perception Prompting (ViP-Prompt) mechanism, which fuses global and multi-scale local visual context to adaptively generate fine-grained textual prompts, eliminating manual templates and class-name priors. This design enables our model to focus on precise abnormal regions, making it particularly valuable when category labels are ambiguous or privacy-constrained. Extensive experiments on 15 industrial and medical benchmarks demonstrate that ViP$^{2}$-CLIP achieves state-of-the-art performance and robust cross-domain generalization.

Few-shot anomaly detection streamlines and simplifies industrial safety inspection. However, limited samples make accurate differentiation between normal and abnormal features challenging, and even more so under category-agnostic conditions. Large-scale pre-training of foundation visual encoders has advanced many fields, as the enormous quantity of data helps to learn the general distribution of normal images. We observe that the anomaly amount in an image directly correlates with the difference in the learnt embeddings and utilize this to design a few-shot anomaly detector termed FoundAD. This is done by learning a nonlinear projection operator onto the natural image manifold. The simple operator acts as an effective tool for anomaly detection to characterize and identify out-of-distribution regions in an image. Extensive experiments show that our approach supports multi-class detection and achieves competitive performance while using substantially fewer parameters than prior methods. Backed up by evaluations with multiple foundation encoders, including fresh DINOv3, we believe this idea broadens the perspective on foundation features and advances the field of few-shot anomaly detection.

The Chinese government unveiled a program to woo foreign talent just as the US cracked down on H-1Bs with a $100,000 fee. The move immediately provoked xenophobic backlash. While President Donald Trump makes it harder to hire skilled foreign workers in the US, Chinese leader Xi Jinping is trying to lure them in. On Wednesday, China officially launched a new visa program designed to make it easier for young professionals and people with degrees in science and technology from top universities to study and do business in the country. While many of the details of the K visa program have yet to be announced, Chinese authorities have said that applicants won't be required to obtain an invitation letter from a specific company, meaning the visa isn't tied to individual employers.

President Trump sent shockwaves through the tech industry over the weekend by announcing a $100,000 payment for new employer-filed H-1B visa applications submitted after September 21, 2025. Since 1990, hundreds of thousands of foreigners have come to work for U.S. tech companies via the visa system. But in a proclamation, Trump wrote that the system had been "deliberately exploited to replace, rather than supplement, American workers with lower-paid, lower-skilled labor." Many experts agree that the H-1B system is flawed and needs amending. But TIME spoke with three professors in economics or business who believe that Trump's new fee system could be counterproductive: that it might push talent overseas; render universities and nonprofits unable to recruit foreign experts; and harm American tech innovation, including in the rapidly emerging field of AI.