A hyperclique in an $r$-uniform hypergraph is a set of vertices in which every $r$-subset is an edge. A simple exhaustive search algorithm detects a hyperclique on $\ell$ vertices in an $r$-uniform hypergraph on $m$ vertices by checking all $\binom{m}{\ell}$ sets of $\ell$ vertices. This algorithm can be expressed as a depth-two circuit with at most $m^\ell$ gates and wires. For hypercliques of *constant* size, Amano (2010) showed that this circuit is almost optimal: for large $r$, constant-depth circuits need $m^{\ell(1-O(\log r/r))}$ wires. We prove that this circuit remains hard to beat even for hypercliques of *polynomial* size. More precisely, circuits of depth $h$ need $m^{Ω(\ell)}$ gates for $\ell\leq m^{1/h-\varepsilon}$, and $m^{\ell(1-O(\log r/r))}$ gates for $\ell=m^{o(1)}$ that tends to infinity. Our hyperclique bound and the method behind it also lead to lower bounds for two central problems of fine-grained complexity. First, we extend the bound $n^{k/4}$ of Rossman (2008) for $k$-Clique from cliques of *constant* size to cliques of *polynomial* size: circuits of depth $h$ detecting a clique on $k$ vertices in an $n$-vertex graph need $n^{Ω(k)}$ gates for $k\leq n^{1/(h-1)-\varepsilon}$. Second, a reduction of Choudhury, Limaye, Sreenivasaiah, and Srinivasan (2026) turns our hyperclique bound into a lower bound of $n^{k-\varepsilon}$ gates for $k$-Orthogonal Vectors ($k$-OV) at every fixed depth for vectors of dimension $\log^{O(1)} n$; this proves the $k$-OV Conjecture for constant-depth circuits in polylogarithmic dimension. Finally, for $2$-OV in logarithmic dimension, we give a simple proof that depth-three circuits need $Ω(n^2)$ wires, which answers a question of Kane and Williams (2019).
High-content multi-channel imaging, from Cell Painting assays to remote sensing, is central to modern scientific pipelines. Because these channels are highly heterogeneous and experimental configurations frequently evolve, a practical vision backbone must be fundamentally channel-agnostic. Current channel-adaptive vision transformers rely on global self-attention with learnable channel embeddings, which can be computed only for previously known channels. Attempts to overcome this limitation of new channels either discard the embeddings, losing channel identity, or process each channel independently, thereby eliminating critical cross-channel interactions. To address this, we introduce ChanSFormer, an efficient Vision Transformer that replaces rigid embeddings with disentangled spatial-channel attention and individual channel CLS tokens. This architecture natively preserves both distinct channel identities and cross-channel information flow. Furthermore, the individual channel class token design unlocks more informative feature representations for self-supervised learning and enables representation-based channel sampling while remaining channel-agnostic. We evaluate ChanSFormer on two biology multi-channel datasets, CHAMMI, JUMP-CP and a satellite dataset, So2Sat. Experimental results show that ChanSFormer outperforms state-of-the-art methods by 3.93%-4.27% on JUMP-CP in classification accuracy. It also demonstrates exceptional cross-dataset transferability under a self-supervised setting, outperforming previous methods by up to 2.72% in accuracy in the same classification tasks. Furthermore, the disentangled attention reduces the quadratic complexity of channel-by-spatial sequence length of the global attention baseline, thus improving throughput by 55%-313%. Code is available at https://github.com/cvlab-stonybrook/ChanSFormer.
Santa Claus and Weighted Nash Social Welfare with Subadditive Valuations
Yuhang Chen, Yuda Feng, Weijiang Hu, Shi Li, Ruilong Zhang
AI 速览：针对次可加估值下的最大最小公平分配（Santa Claus 问题）与加权纳什社会福利，作者报告了仅依赖代理人数 n 的 O(√n) 近似算法，并给出配置 LP 的紧致积分间隙分析。
We study two fairness objectives for allocating $m$ indivisible items among $n$ agents with subadditive valuations: the max-min objective, also known as the Santa Claus problem, which maximizes the smallest value of an agent, and the weighted Nash social welfare, which maximizes the weighted geometric mean of the agents' values. For the max-min allocation problem, we give a randomized polynomial-time $O(\sqrt n)$-approximation algorithm for subadditive valuations with demand queries, improving the $2n-1$ bound of Khot and Ponnuswami. For additive valuations and for submodular valuations with value queries, we give simpler $O(\sqrt n)$-approximation algorithms. For additive valuations, this improves the $O(\sqrt{n\log n})$ bound of Saha and Srinivasan. For these two classes, $O(m^\varepsilon)$-approximation algorithms are also known for every constant $\varepsilon>0$. Our ratios depend only on $n$ and match, up to constants, the $Ω(\sqrt n)$ integrality gap of the configuration LP. For the weighted Nash social welfare problem with subadditive valuations and demand queries, we give an $O(Γ(w))$-approximation algorithm, where $Γ(w)=\exp\bigl((\ln n-H(w))/2\bigr)$ and $H(w)$ is the entropy of the weight vector $w$. The factor $Γ(w)$ equals $1$ for equal weights and is at most $\sqrt n$ for all weights. We show that this ratio is tight for algorithms analyzed against the configuration LP: for every weight vector, the configuration LP has an integrality gap of $Θ(Γ(w))$, even for XOS valuations, and this gap is $Ω(\sqrt n)$ when one agent has almost all the weight. For submodular valuations, the configuration LP has a constant integrality gap for all weights.
This paper presents a formal framework for sequencing instructions in AI agents, inspired by the Indian philosophical system of Mimamsa. The framework formalizes sequencing mechanisms through action object pairs in three distinct ways: direct assertion (Srutikrama) for temporal precedence, purpose driven sequencing (Arthakrama) for functional dependencies, and iterative procedures (Pravrittikrama) for distinguishing between parallel and sequential execution in repetitive tasks. It introduces the syntax and semantics of an action object imperative logic, extending the MIRA formalism (Srinivasan and Parthasarathi, 2021) with explicit deduction rules for sequencing. The correctness of instruction sequencing is established through a validated theorem, which is based on object dependencies across successive instructions. This is further supported by proofs of soundness and completeness. This formal verification enables reliable instruction sequencing, impacting AI applications across areas like task planning and robotics by addressing temporal reasoning and dependency modeling.
The standard method of exponential moments for proving concentration bounds can often be replaced by an argument based on elementary symmetric polynomials. We introduce an additional element of randomness into this framework, which reduces the problem to bounding product moments over a uniformly sampled set of indices. We show that this approach gives useful bounds in three settings. For read-$Δ$ families under limited independence, we obtain bounds governed by the degrees of randomly induced dependency subgraphs, improving the dependence on worst-case degrees. For random linear hashing over arbitrary finite fields with (semi-)random inputs, we derive fixed-bin and maximum-load bounds by controlling the rank defect of random tuples of input keys. Finally, for stochastic processes, we show how decay of product moments yields concentration bounds, recovering the spectral and mixing-time scales for finite-state Markov chains.
Cafer Avcı, Alireza Yazdiani, Roopak Srinivasan, Megh Khaire, Mohammad Tayarani, H. Oliver Gao
Transportation is a major source of energy consumption, greenhouse gas emissions, and criteria air pollutants, making efficient emissions estimation essential for environmental planning and policy analyses. The U.S. Environmental Protection Agency's Motor Vehicle Emission Simulator (MOVES) supports detailed emissions modeling, but its computational demands constrain large-scale scenario analyses. This study introduces uTECH Turbo-MOVES, a software architecture that expedites MOVES computation without modifying its Calculator modules or input/output structure. By combining RunSpec partitioning, container isolation, in-memory intermediate storage, and concurrent workload orchestration, it improves execution efficiency. We evaluate six benchmark cases spanning County, Default, and Project scales and both Emission Inventory and Emission Rates modes. Results demonstrate a measured 5.5x and an estimated 15.3x per-run speedup for Default-scale workloads, and an 84.8x campaign speedup for an eight-county Emission Rates analysis. Although gains vary by workload and hardware configuration, uTECH Turbo-MOVES enables more extensive environmental assessments and transportation policy evaluation.
Yidan Sun, Viktor Schlegel, Srinivasan Nandakumar, Iqra Zahid, Yuping Wu, Yulong Wu, Hao Li, Jie Zhang 等 12 位作者
Synthetic text generation with Differential Privacy (DP) guarantees emerges as a principled approach that can enable the sharing of sensitive datasets across institutional and regulatory boundaries, while bounding the risks of re-identification and membership inference. LLM-based methods deliver promising results; however, comparisons are exacerbated by differing evaluation setups and "private" datasets, potential pre-training contamination is not considered and guarantees are not verified with DP audits. To advance this field, we introduce a unified evaluation framework with standardised utility and fidelity metrics and privacy audits, encompassing nine curated datasets that capture domain-specific complexities such as technical jargon, long-context dependencies, and specialised document structures. In a large-scale empirical study, we benchmark LLM-based state-of-the-art DP text generators of varying sizes (between 1--8B). Our results indicate that DP synthetic text generation remains an unsolved challenge, with quality deteriorating more as the private datasets deviate further from the generators' pre-training corpora. Our analyses suggest that synthetic data quality may be overestimated when LLMs have been pre-trained---without DP---on portions of the "private" data to be generated, and our membership inference (MIA) auditing setup for synthetic text reveals measurable membership signals even under tight privacy budgets. Finally, we note that under the "public pre-training, private generation" paradigm, the nominal $(\varepsilon,δ)$ guarantee does not account for exposure during pre-training, motivating contamination-aware evaluation of DP synthetic text generation. Code and data splits: https://github.com/ImperialGlobalSingapore/synbench
Jingtao Li, Sadiq Jaffer, Anil Madhavapeddy, Pedro Sousa, Anthony Ortiz, Kyle Gao, Zhengpeng Feng, Yanfei Zhong 等 10 位作者
Repeated satellite imagery permits understanding of how our planet changes over time, letting us predict its future and plan mitigating actions. Most existing studies that rely on a time series of satellite observations forecast the spectral values at each pixel for a future target date. However, clouds, shadows, and sensor noise often corrupt these observations, so that spectral forecasters waste much of their capacity on low-level features or noise rather than real change. With the advent of Earth foundation models, we argue that forecasting in embedding space is a feasible and more effective alternative. To this end, we design a token space that handles irregular histories from different sensors and foundation models and supports forecasts at any query date. Tokens from this space are inputs to our proposed forecaster, E$^2$Forecast, which is a transformer with two branches: a Gaussian branch that directly predicts a mean point forecast and per-dimension standard deviations, and a rectified-flow branch that generates forecast ensembles. On Sentinel-1 and Sentinel-2 benchmarks, forecasting embeddings instead of spectral values halves the cosine distance (from 0.20 to 0.10) and greatly reduces overconfidence, lowering the z-score standard deviation from 2.63 to 1.22. By comparing new observations with its forecasts, E$^2$Forecast can also detect anomalous changes without labels or retraining, raising F$_1$ from 41.58 to 48.70 on a global benchmark. This consistent superiority is validated on nine remote sensing foundation models. Beyond change detection, we believe our approach advances the field by permitting effective reuse of diverse pretrained task heads. Our code is publicly available at https://github.com/ucam-eo/e2forecast.
Voice conversion (VC) models can be used to anonymize speech data while preserving naturalness. Existing models, however, are mainly optimized for English and other high-resource languages, and they fail to generalize well to Arabic due to its distinct phonology. In this work, we comprehensively evaluate existing zero-shot VC systems for Arabic speech, identifying limitations in pronunciation preservation. We then train an Arabic-centric model based on EZ-VC on diverse Arabic speech, and demonstrate strong performance in intelligibility, naturalness, and pronunciation preservation, while maintaining high speaker dissimilarity. Through subjective evaluation, we further demonstrate its robustness to code-switching and cross-dialectal conversion, as well as its generalization to Arabic dialects with limited representation in the training data. Our findings highlight the limitations of zero-shot VC and the importance of target-language adaptation for robust voice anonymization.
LucidLyra: Feed-Forward Generative 3D Scene Reconstruction and Joint Understanding via Multi-Task Distillation
Oliver Hahn, Rodrigo Marcuzzi, Sangeetha Grama Srinivasan, Alessandro Burzio, Sherwin Bahmani, Jiahui Huang, Stefan Roth, Shengyu Huang 等 10 位作者
AI 速览：LucidLyra 从视频潜变量出发，以前馈方式生成可查询语义与实例的三维高斯场景，无需逐场景优化。
We present LucidLyra, a framework that makes generatively reconstructed 3D scenes directly queryable for semantics and instances in a feed-forward manner from video latents. While generative 3D reconstruction can synthesize virtual environments from a single image, downstream applications in robotics and physical AI require not only visual realism but also access to scene semantics and object identities. Existing 3D understanding approaches, however, require captured multi-view scenes, have limited generalization capabilities, or rely on expensive per-scene processing and optimization. To bridge this gap, our method distills complementary knowledge from a camera-controlled video diffusion model and frozen 2D vision foundation models into a unified model to jointly infer 3D Gaussians encoding geometry, appearance, open-vocabulary semantics, and class-agnostic instance cues. The resulting scenes support novel-view rendering, open-vocabulary segmentation, and object grouping without per-scene optimization, while training requires neither manual 3D annotations nor captured multi-view data. Our method outperforms baselines combining state-of-the-art 3DGS segmentation approaches with generative 3D scene reconstruction on curated synthetic and real-image-conditioned benchmarks, while preserving reconstruction quality. Furthermore, our latent formulation naturally extends to monocular video, reconstructing structured dynamic 3D scenes across time. By turning renderable but unstructured outputs into queryable 3D worlds, LucidLyra enables downstream applications including 3D scene editing, 3D asset mining, and robotic simulation.
Computer-use agents plan on a screenshot for several seconds and then click at a position on the screen; if the page has changed in the meantime, the click lands on an unintended element. Prior work treats the planning interval as the source of this risk. Measuring planned clicks on 31 public web pages, we find that the risk lies in page loading: 0.4% of clicks planned on fully loaded pages became stale within 30 seconds, against 17% of clicks planned in the first second after a page opened. On these pages, waiting about one second for network activity to stop removes most of this risk, yet the wait rules of eight of ten widely used open-source agents leave more than 1% of clicks stale. Commercial pages keep loading for longer: in the HTTP Archive's public crawl, 15% of 65 desktop booking and buying home pages went 500 ms without network requests within 3 seconds of their HTML being parsed, against 63% of 3,028 .gov pages, and 52% shifted their layout within 6.5 seconds of a network-idle wait capped at 3 seconds, against 8% (61% against 26% among pages recorded in full). Where a check before the click is needed, as before irreversible actions and on pages that are still loading, a pixel comparison verifies the target's position as accurately as a frontier vision-language model at one eighth of the latency, and a 4B model matches the frontier model in verifying whether the previous action succeeded. We release a benchmark generator, a break-even formula for when a check pays, and three rules for agent developers: https://github.com/multimodal-ai-engineering/preflight-bench
Vikas Dwivedi, Balaji Srinivasan, Monica Sigovan, Bruno Sixou
Neural-network-based PDE solvers require architectural choices, particularly the number of layers and neurons per layer. Fixing a single Gaussian radial basis function (RBF) feature layer and taking the number of features to the continuum limit removes the need to prescribe a finite feature count or select individual feature centers, thereby reducing representation arbitrariness. Motivated by this principle, we introduce Continuum-Feature Kernel Solver (CFKS), a deterministic framework for PDEs obtained by analytically marginalizing a bounded continuum of Gaussian RBF features. CFKS works directly with the resulting continuum-feature mean kernel and its operator-applied counterparts, parameterized by prescribed Gaussian widths and bounded center distributions. For linear PDEs, CFKS leads to symmetric operator-kernel collocation systems with null-space enforcement of physical constraints. Nonlinear PDEs are addressed through linearized initialization and nonlinear least-squares residual minimization using analytical kernel derivatives and Jacobians. Residual-guided collocation enrichment and sequential space--time-slab marching provide additional resolution strategies, while prescribed multiscale kernel refinement helps resolve localized nonlinear structures. Finite-feature experiments confirm convergence of empirical value and operator Gram matrices toward their continuum counterparts. Numerical studies include oscillatory and multiscale ODEs, irregular-domain Poisson and biharmonic problems, singular perturbations, transient transport equations, and viscous Burgers' equation. The results establish CFKS as a direct computational connection between finite Gaussian-RBF representations and continuum-feature kernel PDE solvers. Codes: https://github.com/vikas-dwivedi-2022/CFKS_Codes
LOL-v1 and LOLv2-Real are the standard benchmarks for supervised low-light image enhancement, and several leading methods ship hand-set test-time adjustments with their evaluation scripts. We audit both benchmarks with 18 public checkpoints of eight methods under one metric implementation that reproduces every published number to within 0.09 dB, and find that their test labels differ from their training labels. (i) The 15 LOL-v1 test labels are rendered differently: the same captures appear in LOLv2 with labels of identical luminance but 1.48x lower saturation, and LOL-v1's own training labels match the LOLv2 rendering; we confirm this on the official release. (ii) The LOLv2-Real test set is the contiguous block of images 690-789, whose labels are 14% darker than the training labels (Cliff's delta = -0.45, p < 1e-12) with no comparable shift in the input statistics we examined; methods trained without any labels score up to 5.5 dB higher on test than on train. (iii) All 15 LOL-v1 test inputs reappear in LOLv2-Real, 12 in its training set. The released test-time settings are consistent with compensating for these shifts. Separating what can be calibrated on training data from what only test-set tuning finds, the test-only residual has median 0.00 dB on LOLv2-Syn, which has no shift, but 0.37 dB on LOL-v1 and 0.92 dB on LOLv2-Real (up to 3.9 dB). Under the shift, test-set checkpoint selection inflates a from-scratch model by 0.9-2.0 dB, and removing the released settings changes which method ranks first. We give concrete reporting recommendations.
We introduce Nemotron 3.5 Super and Nemotron 3.5 Lightning, open and efficient Mixture-of-Experts Hybrid Mamba-Transformer models designed for reasoning, long-context understanding, and agentic workloads. Super is a 120B-parameter, 12B-active multimodal model, while Lightning is a smaller 30B-parameter, 3B-active text model; both designed for efficient deployment. Both models combine large-scale distillation, long-context training, and native speculative decoding, while Super further incorporates Indic continued pretraining and agentic software-engineering mid-training. We develop a comprehensive post-training pipeline spanning supervised fine-tuning and reinforcement learning (RLVR and RLHF), complemented in Super by specialized teacher training and multi- teacher on-policy distillation (MOPD). This transforms Super from a text-only base model into a multimodal agent capable of reasoning over text, images, video, and interactive tool environments. We further explore test-time scaling through parallel agent execution and self-verification. Across training and deployment, we jointly optimize capability and efficiency. Indic continued pretraining improves Indian-language benchmarks by up to 9.82 percentage points, while agentic mid-training strengthens software-engineering capabilities and robustness across agent harnesses. The post-trained Super model achieves 51.78% on SWE-bench Pro, surpassing Nemotron 3 Ultra despite using approximately 4.6x fewer active parameters. Post-training MTP boosting improves speculative- decoding acceptance length by up to 21.8% under probabilistic draft sampling, while NVFP4 quantization achieves 99.62% median accuracy recovery relative to BF16. Inference optimizations enable up to 4,856 output tokens/s/GPU on GB200 while sustaining at least 50 tokens/s/user.
William Chen, Prem Seetharaman, Ke Chen, Oriol Nieto, Kevin Duarte, Siddharth Srinivasan Iyer, Mamshad Nayeem Rizve, Zhiwen Cao 等 13 位作者
Multimedia editing requires generative models to interpret complex, compositional instructions and modify only the desired elements across modalities, a capability largely beyond existing systems. Current approaches are typically limited to simple, single-attribute edits within one modality, since diverse, high-quality editing pairs are hard to obtain, especially for cross-modal tasks where aligned audiovisual (AV) data is scarce. We present CrossEdit, a unified omni-modal editing model for images, audio, and video that performs AV movie scene edits zero-shot through cross-modal transfer. Our key observation is that complex instruction-following, once learned in any modality, generalizes to others. We exploit the additive nature of acoustic signals to procedurally generate large-scale audio editing pairs with compositional instructions, and show that fine-tuning on this synthetic data and self-supervised AV masked reconstruction, alongside a targeted set of curated cross-modal tasks, induces instruction-following that transfers zero-shot to unseen modality and instruction combinations. To evaluate this capability, we release CrossEditBench, a human-annotated benchmark of AV edits on movie scenes, and propose AV-FES, a metric that jointly scores instruction following and consistency. We show that our proposed techniques improve performance on zero-shot AV movie scene editing, while maintaining or improving performance on lip-synced speech editing and standard image, video, and audio editing benchmarks. Demos are available at https://wanchichen.github.io/crossedit/.
Interactive world models hold immense potential for high-stakes applications in robotics, game engines, and immersive digital twins, yet existing benchmarks primarily focus on perceptual similarity rather than the physical-causal coherence essential for real-world utility. We introduce WMEval, a comprehensive benchmark that evaluates interactive world models against Unity-based physical simulations. By leveraging Unity as a high-fidelity ground-truth oracle, we shift evaluation beyond perceptual visual assessment toward the objective state-based verification required for downstream applications. Spanning 9 diverse scenes, WMEval assesses action controllability, scene and character consistency, prompt adherence, kinematics, fluid mechanics, physical causality, and lighting. By executing matched action sequences in both the simulator and the world model, WMEval enables evaluation at both the perceptual level and the physical-causal level, using simulator-derived states as reference. Experiments on state-of-the-art models reveal significant room for improvement in dimensions that remain hidden by standard perceptual metrics but are vital for real-world deployments. Our code is available at: https://github.com/hao-ai-lab/WMEval.
Jeremy A. Collins, Seungjae Lee, Rachanon Wachakorn, Krishnan Srinivasan, Animesh Garg, Vitor Campagnolo Guizilini, Paarth Shah
Pretrained video generation models encode rich priors about physical interactions, but leveraging them for robot control typically requires either resource-intensive video generation at test time or action heads trained from scratch that do not fully utilize the pretrained backbone. We introduce Tokenizing Actions as Video Patches in a Shared Transformer Policy (TAPESTRY), a world action model that repurposes a pretrained video diffusion transformer into a robotic policy without requiring video generation at inference. Our key idea is to treat robot actions as single-patch tokens in the video model's latent space. A lightweight linear projection maps each action timestep to the patch dimensions of the pretrained VAE, and the resulting tokens are processed by the same patchifying layers and transformer backbone used for video. The shared backbone is optimized through three diffusion objectives sampled per minibatch element: Video diffusion uses VAE latents from robot and human videos to establish dynamics priors. Joint video+action diffusion trains on robot demonstrations using a coupled diffusion timestep and an asymmetric attention mask. Action-only diffusion denoises actions conditioned on clean image latents. At test time, the model generates a compact action token sequence in a single DDIM step from pure noise, producing smooth action chunks at 84 ms per action chunk on consumer hardware and avoiding the cost of full video generation. Across simulation benchmarks and real-robot experiments, TAPESTRY achieves strong performance with substantially lower inference cost than comparable joint video--action policies. More information can be found at https://tapestry-wam.github.io.
Jiaming Hu, Shi Bo, Yan Zheng, Tian Wang, Florian Dubost, Alejandro Mottini, Junze Liu, Arvind Srinivasan 等 14 位作者
Joint-embedding predictive architectures (JEPAs) enable world models to plan from images by predicting future embeddings. However, their visual prediction objectives do not explicitly incorporate task-relevant state information, so the physical state that determines task progress may not be reliably captured in the latent space. Supervising world models with simulator state can make planning more informative, but it requires deciding which state features matter. Features are typically selected by hand or weighted uniformly, and a uniform weighting gives every standardized state feature equal influence, so redundant coordinates can overcount the same variation, while irrelevant ones can obscure task-relevant differences. To address this limitation, we propose SWIM (State Weighting with Inverse-Dynamics Model), which learns state weights from offline trajectories by predicting recorded actions from current states and observed future displacements. The learned weights quantify each feature's contribution to action prediction and define a weighted state metric for state-supervised training, which we demonstrate by aligning latent and state distances. The inverse-dynamics model is used only during training, so the resulting world model plans from images with no additional planning-time computation. Across five simulated robotic manipulation tasks, SWIM consistently improves planning success over uniform state weighting.
Feed-forward 3D reconstruction models such as VGGT regress cameras and geometry in a single forward pass, but they report no uncertainty for the camera poses they predict: an incorrect relative pose looks no different from a correct one. We ask whether the frozen model nonetheless knows when its pose is wrong. On HardMatch, a benchmark of 1,000 deliberately difficult image pairs on which VGGT fails on 35% of pairs, a linear probe on VGGT's frozen features predicts pose failure on held-out categories with ROC-AUC 93.5%, compared with 79.7% for VGGT's own point-map confidence and 77.7% for a classifier trained on all of VGGT's outputs. The probe catches failures VGGT is confident about, lets a downstream system abstain on the pairs it distrusts, and transfers without retraining to rotation failures on PhotoTourism and ScanNet-1500, beyond what co-visibility explains. This knowledge is absent from the per-image encoder: it is computed partway through the layers that fuse the two views, and the same holds for a second model, $π^3$. In essence, we find that VGGT knows more about its own failures than it reports.
Large organizations consume large language models (LLMs) from several vendors at once, each with its own API dialect, credential model, streaming protocol and pricing. Integrated directly, this heterogeneity yields credential sprawl, format lock-in, unbounded spend and inconsistent governance. We present the design and production operation of the Cosmos LLM Gateway, the governed entry point for enterprise LLM traffic at PayPal. Built on Envoy, it splits work by latency profile: I/O-bound control decisions (authentication, credential brokering, policy and budget evaluation) run in an out-of-process external-processing service, while wire-format translation, token metering and prompt-cache shaping run in-proxy as a Rust WebAssembly filter, so that, by default, streamed response body chunks never wait on an out-of-process call. Models, providers and prices are largely validated data rather than code. Governance is declarative policy evaluated by Open Policy Agent and managed through a control plane with role-scoped authoring, approvals, pilots, shadow-mode rollout, usage notifications and versioned policy bundles. For use cases that require it, a data-loss-prevention layer can block, redact or pseudonymize detected sensitive content before it reaches a provider. Spend scopes are views over per-model usage counters, so a new budget applies to spend already recorded in its window without migration. As of September 2026, the gateway serves more than 12,000 developers and hundreds of business use cases, handling roughly 200 billion tokens per day across more than 100 models at 99.999% availability to business. In a benchmark of the core request path against a synthetic upstream, it adds a median of 4 ms and a p99 of 11 ms of overhead at more than 2,200 requests per second on two 4-vCPU instances. We report the architecture, policy model, evaluation and lessons learned.
The growing computational power of modern GPU data centers has let a few industrial giants train and deploy large language models (LLMs) on massive amounts of data. Yet the sheer scale of these models and their training data makes centralized development increasingly hard to sustain, and at the same time shuts out smaller institutions that hold relevant private data but lack comparable infrastructure. In this paper, we organize our positions on distributed LLM training around the pretraining and fine-tuning paradigms, covering federated learning, split learning, and decentralized approaches, and we assess their readiness for geo-distributed settings with heterogeneous hardware and unreliable connectivity. We then present a {\em unified} system architecture built around a cloud-hosted parameter server that coordinates heterogeneous clients over outbound gRPC connections, reusing one communication, compression, and fault-tolerance stack for both pretraining and a family of fine-tuning methods. We validate the design with real geo-distributed experiments on GPT-2 Medium and Llama3-1B pretraining across multiple university sites, showing stable convergence under realistic conditions. Finally, we chart the road ahead, identifying open problems that span frontier model support, elastic fault tolerance, adaptive compression, federated alignment, privacy and trust, incentive design, and system--algorithm co-design, all of which must be addressed to make democratized LLM training a reality.
Srinath Srinivasan, Lunxiao Li, Tim Menzies, Wesley Assunção
Early stopping promises to reduce compute costs for autonomous coding agents by terminating unpromising runs early. Deciding when to abandon a run requires separating baseline issue difficulty from the trajectory of an individual attempt. We study the distinction in 147,002 execution traces from two public corpora spanning four model-framework settings. Comparing pre-execution issue context against early execution prefixes, we evaluate linear count models, initial-turn text embeddings, task-conditional rankers, and a fine-tuned Qwen3-0.6B monitor. Although early prefixes modestly improve cross-task outcome ranking, within-task discrimination remains near chance across early landmarks (5k-20k tokens): initial-turn embeddings and 20k-token task-conditional models yield AUCs around 0.50, and three-seed neural evaluations across 32,406 attempts at 20k tokens achieve a mean within-task AUC of 0.50 [0.48, 0.51] across tasks of varying solve rates. To evaluate how these predictions translate into stopping decisions, we replay score-guided and matched random restarts under shared token budgets alongside uninterrupted continuation. Under a 75k-token cap, learned restarting improves resolution over matched random restarting by up to 1.98 percentage points. An exact algebraic decomposition resolves this paradox: the gain over random restarting stems almost entirely from between-task stopping allocation (avoiding stops on easy issues), whereas attempt-level selection within a task contributes near zero (-0.11 to +0.24 percentage points). Across all four settings at the 75k cap (and 216 of 224 tested budget-threshold configurations overall), uninterrupted continuation achieves higher solve rates at lower mean token cost than both restart policies. Practical evaluations of early stopping must therefore establish attempt-level discrimination and demonstrate net compute savings relative to continuation.
Gauri Kambhatla University of Texas at Austin, Sanjana Gautam, Angela Zhang, Alexander Liu University of Texas at Austin, Ravi Srinivasan University of Texas at Austin, Junyi Jessy Li University of Texas at Austin, Matthew Lease Amazon and University of Texas at Austin
The ability to accurately align LLMs with diverse population groups on subjective questions would have great value. In this work, we show that adding simple supervision can more consistently improve the alignment of LLM-generated distributions with diverse population groups, as measured across three datasets spanning public health, public opinion, and values and beliefs. Beyond evaluating average alignment, we also report how alignment varies across specific groups. Our broad findings provide insights into the distributional alignment of LLM generations with diverse populations. By conducting evaluation over many LLMs and prompting strategies, we provide a benchmark to stimulate future research.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Keyu He, Tejas Srinivasan University of Southern California, Brihi Joshi, Xiang Ren University of Southern California, Jesse Thomason University of Southern California and Amazon, Swabha Swayamdipta University of Southern California
When people query Vision-Language Models (VLMs) but cannot see the accompanying visual context (e.g. for blind and low-vision users), augmenting VLM predictions with natural language explanations can signal which model predictions are reliable. However, prior work has found that explanations can easily convince users that inaccurate VLM predictions are correct. To remedy undesirable overreliance on VLM predictions, we propose evaluating two complementary qualities of VLM-generated explanations via two quality scoring functions. We propose Visual Fidelity, which captures how faithful an explanation is to the visual context, and Contrastiveness, which captures how well the explanation identifies visual details that distinguish the model’s prediction from plausible alternatives. On the A-OKVQA, VizWiz, and MMMU-Pro tasks, these quality scoring functions are better calibrated with model correctness than existing explanation qualities. We conduct a user study in which participants have to decide whether a VLM prediction is accurate without viewing its visual context. We observe that showing our quality scores alongside VLM explanations improves participants’ accuracy at predicting VLM correctness by 11.1%, including a 15.4% reduction in the rate of falsely believing incorrect predictions. These findings highlight the utility of explanation quality scores in fostering appropriate reliance on VLM predictions.
尚无 AI 解读，展示原文摘要
cs.CL · cs.CV
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
AI 速览：作者报告，经过恰当提示并配备合适工具后，现代LLM能为几乎任意提示稳健生成高质量定制界面，人类偏好显著超过标准markdown输出。
AI models excel at creating content, but typically render it with static, predefined interfaces. Specifically, the output of LLMs is often a markdown "wall of text". Generative UI is a long standing promise, where the model generates not just the content, but the interface itself. Until now, Generative UI was not possible in a robust fashion. We demonstrate that when properly prompted and equipped with the right set of tools, a modern LLM can robustly produce high quality custom UIs for virtually any prompt. When ignoring generation speed, results generated by our implementation are overwhelmingly preferred by humans over the standard LLM markdown output. In fact, while the results generated by our implementation are worse than those crafted by human experts, they are at least comparable in 50% of cases. We show that this ability for robust Generative UI is emergent, with substantial improvements from previous models. We also create and release PAGEN, a novel dataset of expert-crafted results to aid in evaluating Generative UI implementations, as well as the results of our system for future comparisons. Interactive examples can be seen at https://generativeui.github.io
In this work, we explore an untapped signal in diffusion model inference. While all previous methods generate images independently at inference, we instead ask if samples can be generated collaboratively. We propose Group Diffusion, unlocking the attention mechanism to be shared across images, rather than limited to just the patches within an image. This enables images to be jointly denoised at inference time, learning both intra and inter-image correspondence. We observe a clear scaling effect -- larger group sizes yield stronger cross-sample attention and better generation quality. Furthermore, we introduce a qualitative measure to capture this behavior and show that its strength closely correlates with FID. Built on standard diffusion transformers, our GroupDiff achieves up to 32.2% FID improvement on ImageNet-256x256. Our work reveals cross-sample inference as an effective, previously unexplored mechanism for generative modeling.
尚无 AI 解读，展示原文摘要
cs.CV
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition · 2026 · 正式发表
Satellite Earth-observation (EO) time series in the optical and microwave ranges are often irregular due to orbital patterns and cloud obstruction, and while compositing addresses these issues, it loses critical phenological information. To overcome this, we present TESSERA, a pixel-wise foundation model for multi-modal (Sentinel-1/2) EO time series that learns robust, label-efficient embeddings. During training, TESSERA uses Barlow Twins and sparse random temporal sampling to enforce invariance to the selection of valid observations, aided by two key regularizers: global shuffling to decorrelate spatial neighborhoods and mix-based regulation for invariance under extreme sparsity. We find that for diverse classification, segmentation, and regression tasks, TESSERA embeddings deliver state-of-the-art accuracy with high label efficiency, often requiring only a small task head and minimal computation. To democratize access, adhere to FAIR principles, and simplify use, we release global, annual, 10m, pixel-wise int8 embeddings together with open weights/code and lightweight adaptation heads, providing practical tooling for large-scale retrieval and inference at planetary scale. All code and data are available at https://github.com/ucam-eo/tessera.
尚无 AI 解读，展示原文摘要
cs.CV
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition · 2026 · 正式发表
Venna, Venkata Kesav, Gunda, Sai Madhusudan, Jinka, Jyothi Swaroopa, Rachakonda, Hrithik Sagar, Srinivasan, Anirudh, Sarvadevabhatla, Ravi Kiran
**Document QA** requires not only accurate answers but also identifying where each answer is grounded on the page. Most models treat the task as text-only generation, while existing answer grounding methods generate coarse bounding boxes that fail to capture curved text. We introduce **M3Grounder, a hybrid vision-language and segmentation architecture that formulates document grounding as pixel-level segmentation. It produces fine-grained evidence masks** refined by a bleed-suppression loss to prevent spillover. M3Grounder autoregressively generates answer text interleaved with [GROUND] tokens that link individual answer spans to their corresponding evidence regions. Also, **M3Grounder grounds evidence hierarchically across phrase, line, and block levels** using an enclosure loss that enforces spatial containment. We release **GroundingDocQA dataset (200K documents, 2M multi-span and multi-granular QA pairs with pixel-level grounding masks)**, built through a data engine that handles complex layouts, curved-text, and graphics-rich documents. We also release **GroundingDocQA-Bench, a diverse and challenging human-verified benchmark**. M3Grounder sets **a new state of the art in grounded DocVQA, advancing from coarse boxes to hierarchical, fine-grained and contextually grounded evidence.**
尚无 AI 解读，展示原文摘要
cs.CV
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition · 2026 · 正式发表
Hao Cui, Zahra Shamsi, Gowoon Cheon, Xuejian Ma, Shutong Li, Maria Tikhanovskaya, Peter Norgaard, Nayantara Mudur 等 34 位作者
Scientific problem-solving involves synthesizing information while applying expert knowledge. We introduce CURIE, a scientific long-Context Understanding,Reasoning and Information Extraction benchmark to measure the potential of Large Language Models (LLMs) in scientific problem-solving and assisting scientists in realistic workflows. This benchmark introduces ten challenging tasks with a total of 580 problems and solution pairs curated by experts in six disciplines - materials science, condensed matter physics, quantum computing, geospatial analysis, biodiversity, and proteins - covering both experimental and theoretical work-flows in science. We evaluate a range of closed and open LLMs on tasks in CURIE which requires domain expertise, comprehension of long in-context information,and multi-step reasoning. While Gemini Flash 2.0 and Claude-3 show consistent high comprehension across domains, the popular GPT-4o and command-R+ fail dramatically on protein sequencing tasks. With the best performance at 32% there is much room for improvement for all models. We hope that insights gained from CURIE can guide the future development of LLMs in sciences. Evaluation code and data are in https://github.com/google/curie
Gemini Team, Petko Georgiev, Ving Ian Lei, Ryan Burnell, Libin Bai, Anmol Gulati, Garrett Tanzer, Damien Vincent 等 1137 位作者
In this report, we introduce the Gemini 1.5 family of models, representing the next generation of highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio. The family includes two new models: (1) an updated Gemini 1.5 Pro, which exceeds the February version on the great majority of capabilities and benchmarks; (2) Gemini 1.5 Flash, a more lightweight variant designed for efficiency with minimal regression in quality. Gemini 1.5 models achieve near-perfect recall on long-context retrieval tasks across modalities, improve the state-of-the-art in long-document QA, long-video QA and long-context ASR, and match or surpass Gemini 1.0 Ultra's state-of-the-art performance across a broad set of benchmarks. Studying the limits of Gemini 1.5's long-context ability, we find continued improvement in next-token prediction and near-perfect retrieval (>99%) up to at least 10M tokens, a generational leap over existing models such as Claude 3.0 (200k) and GPT-4 Turbo (128k). Finally, we highlight real-world use cases, such as Gemini 1.5 collaborating with professionals on completing their tasks achieving 26 to 75% time savings across 10 different job categories, as well as surprising new capabilities of large language models at the frontier; when given a grammar manual for Kalamang, a language with fewer than 200 speakers worldwide, the model learns to translate English to Kalamang at a similar level to a person who learned from the same content.