Convolutional neural networks (CNNs) trained on object recognition achieve high task performance but continue to exhibit vulnerability under a range of visual perturbations and out-of-domain images, when compared with biological vision. Prior work has demonstrated that coupling a standard CNN with a front-end (VOneBlock) that mimics the primate primary visual cortex (V1) can improve overall model robustness. Expanding on this, we introduce Early Vision Networks (EVNets), a new class of hybrid CNNs that combine the VOneBlock with a novel SubcorticalBlock, whose architecture draws from computational models in neuroscience and is parameterized to maximize alignment with subcortical responses reported across multiple experimental studies. Without being optimized to do so, the assembly of the SubcorticalBlock with the VOneBlock improved V1 alignment across most standard V1 benchmarks, and better modeled extra-classical receptive field phenomena. In addition, EVNets exhibit stronger emergent shape bias and outperform the base CNN architecture by 9.3% on an aggregate benchmark of robustness evaluations, including adversarial perturbations, common corruptions, and domain shifts. Finally, we show that EVNets can be further improved when paired with a state-of-the-art data augmentation technique, surpassing the performance of the isolated data augmentation approach by 6.2% on our robustness benchmark. This result reveals complementary benefits between changes in architecture to better mimic biology and training-based machine learning approaches. 1

Every emotional moment from the Gilgo Beach killer's sentencing: Rex Heuermann's shocking first words... and the chilling exchange that silenced the room Don Trump Jr. says Ted Cruz is'lying through his teeth' as GOP infighting over Iran deal continues to spiral LIZ JONES: The cracks in Harry and Meghan's perfect facade have started to show. It's so obvious he's tiring of her tone-deaf approach... and I predict there's serious trouble in store All my friends are suddenly getting divorced. Mid-life wives share taboo sex confessions about why they really leave... including common position that made one hate her husband: JANA HOCKING Trump says'fools who think I haven't been tough enough on Iran' are'jealous or stupid' after signing widely-criticised deal that includes giving Tehran $300billion Taylor Swift's bottomless thirst for attention, her greed and sheer tackiness are now truly unbearable... this latest stunt has shown her true colors: MAUREEN CALLAHAN The Ring star Daveigh Chase's friends searched for her on LA's Skid Row in months before her shock death at 35 Horrific new videos blow Texas woman's mystery death wide open: Her agonizing'final gasp'... unthinkably vile corpse claims... and sick past of man who saw her last'I can still see you': Princess of Wales shares adorable moment with shy little girl at Royal Ascot Brooklyn Beckham is savaged by fans for yet another'classless' swipe at his estranged family as new DoorDash ad is branded a'giant PR mess' Sensational REAL reason Jelly Roll is divorcing Bunnie XO: Insiders reveal'preacher's wife' bombshell that's the talk of Nashville... truth about legendary rocker cuckolding rumor... and G-string mishap Watch horrifying drone video that follows woman's plunge to death after bungee team threw her from bridge without rope Kylie Jenner and Timothee Chalamet put on VERY affectionate display during NYC bike ride... as it's revealed how the relationship has'changed' the actor Luxury fashion tycoon beloved by the stars hangs her head in shame as she's indicted for allegedly exploiting her workers and stealing $50k from their wages NBA star's fiancee breaks her silence after friend, 26, mysteriously dropped dead at her luxury bachelorette party in St Barts Tragedy as 8-year-old dies during World Cup watch party with cops blaming'accidental drowning' Instead, I lost a stone and dropped a dress size in one MONTH with a meal plan that's not even a diet. It's packed with carbs and so simple - anyone can do it in time for summer Gen Z are scared of ringing the DOORBELL: One in three youngsters now text or call when they arrive at someone's door because they think it's less awkward It's something most people do without thinking twice about it. But a new survey has revealed how ringing the doorbell is leaving many Gen Z petrified.

Characterizing interactions between brain areas is a fundamental goal of systems neuroscience. While such analyses are possible when areas are recorded simultaneously, it is rare to observe all combinations of areas of interest within a single animal or recording session. How can we leverage multi-animal datasets to better understand multi-area interactions? Building on recent progress in large-scale, multi-animal models, we introduce NeuroPaint, a masked autoencoding approach for inferring the dynamics of unrecorded brain areas. By training across animals with overlapping subsets of recorded areas, NeuroPaint learns to reconstruct activity in missing areas based on shared structure across individuals. We train and evaluate our approach on synthetic data and two multi-animal, multi-area Neuropixels datasets. Our results demonstrate that models trained across animals with partial observations can successfully in-paint the dynamics of unrecorded areas, enabling 39th Conference on Neural Information Processing Systems (NeurIPS 2025).

Action segmentation is a core challenge in high-level video understanding, aiming to partition untrimmed videos into segments and assign each a label from a predefined action set. Existing methods primarily address single-person activities with fixed action sequences, overlooking multi-person scenarios. In this work, we pioneer textual reference-guided human action segmentation in multi-person settings, where a textual description specifies the target person for segmentation. We introduce the first dataset for Referring Human Action Segmentation, i.e., RHAS133, built from 133 movies and annotated with 137 fine-grained actions with 33h video data, together with textual descriptions for this new task. Benchmarking existing action segmentation methods on RHAS133 using VLM-based feature extractors reveals limited performance and poor aggregation of visual cues for the target person. To address this, we propose a holistic-partial aware Fourierconditioned diffusion framework, i.e., HopaDIFF, leveraging a novel cross-input gate attentional xLSTM to enhance holistic-partial long-range reasoning and a novel Fourier condition to introduce more fine-grained control to improve the action segmentation generation. HopaDIFF achieves state-of-the-art results on RHAS133 in diverse evaluation settings.

In this paper, we examine the convergence landscape of multi-agent learning under uncertainty. Specifically, we analyze two stochastic models of regularized learning in continuous games--one in continuous and one in discrete time--with the aim of characterizing the long-run behavior of the induced sequence of play. In stark contrast to deterministic, full-information models of learning (or models with a vanishing learning rate), we show that the resulting dynamics do not converge in general. In lieu of this, we ask instead which actions are played more often in the long run, and by how much. We show that, in strongly monotone games, the dynamics of regularized learning may wander away from equilibrium infinitely often, but they always return to its vicinity in finite time (which we estimate), and their long-run distribution is sharply concentrated around a neighborhood thereof. We quantify the degree of this concentration, and we show that these favorable properties may all break down if the underlying game is not strongly monotone--underscoring in this way the limits of regularized learning in the presence of persistent randomness and uncertainty.

Verifiers can improve language model (LM) capabilities by providing feedback or selecting the best response from a pool of generated candidates. Currently, high-quality verifiers are either unscalable (e.g., humans) or limited in utility (e.g., tools like Lean for formal proofs). While LM judges and reward models have become broadly useful as general-purpose verifiers, a significant performance gap remains between them and oracle verifiers. To help close this gap, we introduce WEAVER, a framework for designing a strong verifier by combining multiple weak, imperfect verifiers. First we find that weighted ensembles of verifiers, which typically require learning from labeled data, significantly outperform unweighted combinations due to differences in the verifiers. To reduce the dependency on labeled data, WEAVER leverages weak supervision to estimate each verifier's accuracy and combines their outputs into a unified score that better reflects true response quality.

The 2026 World Cup promises a global celebration. Many Arab fans may find themselves excluded. For the first time in World Cup history, eight Arab nations have qualified for this year's tournament, including Morocco, Tunisia, Egypt, Algeria, Saudi Arabia, Qatar, Iraq, and Jordan--double the number of teams that qualified for Qatar in 2022. Yet, the tournament is taking place at an unprecedented moment of heightened geopolitical tension. The US-Israel war with Iran, which began in February of this year, has caused ripple effects across Gulf states and neighboring countries in the Levant, including Lebanon, Palestine, and Jordan, reshaping the security around travel and mobility for fans and players hailing from the region. The US State Department has fully suspended visa issuance for nationals from countries with teams that qualified, including Iran and Haiti--despite it being the first time Haiti has qualified for a World Cup since 1974.

Things to Do in L.A. Tap to enable a layout that focuses on the article. Inside the FBI's battle to keep drones out of World Cup sites amid heightened terrorism fears Drones were seized at SoFi Stadium and the L.A. Coliseum. This is read by an automated voice. Please report any issues or inconsistencies here . As World Cup soccer fans cheer for their teams in Los Angeles and stadiums across the nation, FBI agents are working in command centers, watching out for unauthorized flying objects.

We study how training data contributes to the emergence of toxic behaviors in large language models. Most prior work on reducing model toxicity adopts reactive approaches, such as fine-tuning pre-trained (and potentially toxic) models to align them with human values. In contrast, we propose a proactive approach-- IF-GUIDE--that leverages influence functions to identify and suppress harmful tokens in the training data. To this end, we first show that standard influence functions are ineffective at discovering harmful training records. We then present a novel adaptation that measures token-level attributions from training data to model toxicity, along with techniques for selecting toxic training documents and a learning objective that can be integrated into both pre-training and fine-tuning. Moreover, IF-GUIDE does not rely on human-preference data, which is typically required by existing alignment methods. In our evaluation, we demonstrate that IF-GUIDE substantially reduces both explicit and implicit toxicity--by up to 10 compared to uncensored models, and up to 3 compared to baseline alignment methods such as DPO and RAD--across both pre-training and fine-tuning scenarios. IF-GUIDE is computationally efficient: a billion-parameter model is not necessary for computing influence scores; a million-parameter model--with 7.5 fewer parameters--can effectively serve as a proxy for identifying harmful data.

Modern reinforcement learning (RL) algorithms have found success by using probabilistic models, such as transformers, energy-based models, and diffusion/flowbased models. To this end, researchers often choose to pay the price of accommodating these models into their algorithms - diffusion models are expressive, but are computationally intensive due to their reliance on solving differential equations, while autoregressive transformer models are scalable but typically require learning discrete representations. Normalizing flows (NFs), by contrast, seem to provide an appealing alternative, as they enable likelihoods and sampling without solving differential equations or autoregressive architectures. However, their potential in RL has received limited attention, partly due to the prevailing belief that normalizing flows lack sufficient expressivity. We show that this is not the case. Building on recent work in NFs, we propose a single NF architecture which integrates seamlessly into RL algorithms, serving as a policy, Q-function, and occupancy measure. Our approach leads to much simpler algorithms, and achieves higher performance in imitation learning, offline, goal conditioned RL and unsupervised RL.1