In reinforcement learning with sparse rewards, demonstrations can accelerate learning, but determining when to imitate them remains challenging. We propose Smooth Policy Regularisation from Demonstrations (SPReD), a framework that addresses the fundamental question: when should an agent imitate a demonstration versus follow its own policy? SPReD uses ensemble methods to explicitly model Q-value distributions for both demonstration and policy actions, quantifying uncertainty for comparisons. We develop two complementary uncertainty-aware methods: a probabilistic approach estimating the likelihood of demonstration superiority, and an advantage-based approach scaling imitation by statistical significance.

Large language models (LMs) often struggle to generate diverse, human-like creative content, raising concerns about the long-term homogenization of human thought through repeated exposure to similar outputs. Yet scalable methods for evaluating LM output diversity remain limited, especially beyond narrow tasks such as random number or name generation, or beyond repeated sampling from a single model. To address this gap, we introduce INFINITY-CHAT, a largescale dataset of 26K diverse, real-world, open-ended user queries that admit a wide range of plausible answers with no single ground truth. We introduce the first comprehensive taxonomy for characterizing the full spectrum of open-ended prompts posed to LMs, comprising 6 top-level categories (e.g., creative content generation, brainstorm & ideation) that further breaks down to 17 subcategories.

In enterprise settings, organizational data is segregated, siloed and carefully protected by elaborate access control frameworks. These access control structures can completely break down if an LLM fine-tuned on the siloed data serves requests, for downstream tasks, from individuals with disparate access privileges. We propose Permissioned LLMs (PermLLM), a new class of LLMs that superimpose the organizational data access control structures on query responses they generate. We formalize abstractions underpinning the means to determine whether access control enforcement happens correctly over LLM query responses. Our formalism introduces the notion of a relevant response that can be used to prove whether a PermLLM mechanism has been implemented correctly. We also introduce a novel metric, called access advantage, to empirically evaluate the efficacy of a PermLLM mechanism. We introduce three novel PermLLM mechanisms that build on Parameter Efficient Fine-Tuning to achieve the desired access control. We furthermore present two instantiations of access advantage-(i) Domain Distinguishability Index (DDI) based on Membership Inference Attacks, and (ii) Utility Gap Index (UGI) based on LLM utility evaluation. We demonstrate the efficacy of our PermLLM mechanisms through extensive experiments on five public datasets (GPQA, RCV1, SimpleQA, WMDP, and PubMedQA), in addition to evaluating the validity of DDI and UGI metrics themselves for quantifying access control in LLMs.

Can we scale 4D pretraining to learn general space-time representations that reconstruct an object from a few views at some times to any view at any time? We provide an affirmative answer with 4D-LRM, the first large-scale 4D reconstruction model that takes input from unconstrained views and timestamps and renders arbitrary novel view-time combinations. Unlike prior 4D approaches, e.g., optimizationbased, geometry-based, or generative, that struggle with efficiency, generalization, or faithfulness, 4D-LRM learns a unified space-time representation and directly predicts per-pixel 4DGaussian primitives from posed image tokens across time, enabling fast, high-quality rendering at, in principle, infinite frame rate. Our results demonstrate that scaling spatiotemporal pretraining enables accurate and efficient 4D reconstruction. We show that 4D-LRM generalizes to novel objects, interpolates across time, and handles diverse camera setups. It reconstructs 24-frame sequences in one forward pass with less than 1.5 seconds on a single A100 GPU.

Over the last decade, ransomware detection has become a central topic in cybersecurity research. Due to ransomware's direct interaction with storage devices, analyzing I/O streams has become an effective detection method and represents a vital area of focus for research. A major challenge in this field is the lack of publicly accessible data featuring individual command labeling. To address this problem, we introduce the Command LEvel Annotated Ransomware (CLEAR) dataset, a large-scale collection of storage devices' stream data. The dataset comprises 1,045 TiB of I/O traffic data, featuring malicious traffic from 137 ransomware variants.

'Ringleader' of alleged UFC drone attack to kill Trump is unmasked as illegal migrant who was granted DACA stay under Obama Watch horrifying drone video that follows woman's plunge to death after bungee team threw her from bridge without rope 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 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 Spy world panic as Tulsi Gabbard prepares to unleash bombshell file dumps on secret CIA'mind control' project and Dr. Fauci Beloved mattress company backed by Travis Kelce files for bankruptcy... but makes key promise to customers Olivia Wilde, 42, complains about being on Maxim's Hot 100 List calling it the'most f***** up thing in the world' 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 Kanye West's wife Bianca Censori raises eyebrows in plunging white lace lingerie as she photographs a nude model at Art Basel in Switzerland Knicks set to come face to face with Trump after president was'thunderously booed' at NBA Finals game 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 Has Taylor Swift already revealed her wedding dress designer? Teen tourist thrown to death by Central Park horse was trying to save mom who flew out of carriage during family's first visit to Big Apple Father keeps his cool as shouting man calls cops on him for taking his two young daughters into women's restroom Trump privately frets Bibi Netanyahu's zeal to'bomb everyone' could turn him into another disgraced president'Moscow will burn', Zelensky vows as Russia's capital is blanketed in toxic smoke following huge Ukraine drone attack He drove a Rolls-Royce and lived the American dream. But behind the Gucci was the ATF's most unlikely secret weapon. What do aliens EAT? Scientist reveals the foods extraterrestrials would go for on Earth - and why E.T.'s favourite Reese's Pieces are off the cards In the 1982 blockbuster, E.T. the Extra-Terrestrial, E.T. is lured out of hiding using a trail of Reese's Pieces.

Visual navigation is essential for robotics and embodied AI. However, existing foundation models, particularly those with transformer decoders, suffer from high computational overhead and lack interpretability, limiting their deployment in resource-tight scenarios. To address this, we propose DynaNav, a Dynamic Visual Navigation framework that adapts feature and layer selection based on scene complexity. It employs a trainable hard feature selector for sparse operations, enhancing efficiency and interpretability. Additionally, we integrate feature selection into an early-exit mechanism, with Bayesian Optimization determining optimal exit thresholds to reduce computational cost. Extensive experiments in real-world-based datasets and simulated environments demonstrate the effectiveness of DynaNav. Compared to ViNT, DynaNav achieves a 2.26 reduction in FLOPs, 42.3% lower inference time, and 32.8% lower memory usage, while improving navigation performance across four public datasets.

Recent advances in Large Language Models (LLMs) have highlighted the critical importance of extending context length, yet the quadratic complexity of attention mechanisms poses significant challenges for efficient long-context modeling. KV cache compression has emerged as a key approach to address this challenge. Through extensive empirical analysis, we reveal a fundamental yet previously overlooked asymmetry in KV caches: while adjacent keys receive similar attention weights (local homogeneity), adjacent values demonstrate distinct heterogeneous distributions. This key-value asymmetry reveals a critical limitation in existing compression methods that treat keys and values uniformly. To address the limitation, we propose a training-free compression framework (AsymKV) that combines homogeneity-based key merging with a mathematically proven lossless value compression. Extensive experiments demonstrate that AsymKV consistently outperforms existing long-context methods across various tasks and base models.

We investigate optimal transport (OT) for measures on graph metric spaces with different total masses. To mitigate the limitations of traditional Lp geometry, Orlicz-Wasserstein (OW) and generalized Sobolev transport (GST) employ Orlicz geometric structure, leveraging convex functions to capture nuanced geometric relationships and remarkably contribute to advance certain machine learning approaches. However, both OW and GST are restricted to measures with equal total mass, limiting their applicability to real-world scenarios where mass variation is common, and input measures may have noisy supports, or outliers. To address unbalanced measures, OW can either incorporate mass constraints or marginal discrepancy penalization, but this leads to a more complex two-level optimization problem. Additionally, GST provides a scalable yet rigid framework, which poses significant challenges to extend GST to accommodate nonnegative measures.

End-to-end autonomous driving has substantially progressed by directly predicting future trajectories from raw perception inputs, which bypasses traditional modular pipelines. However, mainstream methods trained via imitation learning suffer from critical safety limitations, as they fail to distinguish between trajectories that appear human-like but are potentially unsafe. Some recent approaches attempt to address this by regressing multiple rule-driven scores but decoupling supervision from policy optimization, resulting in suboptimal performance. To tackle these challenges, we propose DriveDPO, a Safety Direct Preference Optimization Policy Learning framework.