跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+197新增论文
+151其中 AI 相关
+16已有论文修订
+91新增 AI 导读

arXiv 最近完成扫描:2026-11-22 08:44 · 全站最近入库:2026-11-21 12:19各信源进度 →

统计口径与数据新鲜度

统计 2026-11-21 11:34 — 2026-11-22 11:34(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-22 08:44最近发现新增版本:2026-11-21 12:19库内论文最新更新时间:2026-11-20 02:59

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

29,158 已收录论文24,181 AI 相关8,992 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 23,345 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-22 08:42(北京时间)
版本更新 · v4arXiv:2606.10044

Business World Model

Cecil Pang, Hiroki Sayama

原文摘要

World model has emerged as a powerful paradigm in artificial intelligence, enabling agents to represent their environments, predict future states, and evaluate possible actions before acting. However, existing world models have largely been developed for visual or physical domains such as computer vision, robotics, gaming, and autonomous driving, which are governed by relatively stable dynamics. These formulations are not directly applicable to business practice, where the relevant environment is semantic, organizational, and market driven. Outcomes depend on context-sensitive factors such as customer behavior, pricing, competition, and regulation. This paper introduces the concept and architecture of a Business World Model, a world model specialized for business and organizational environments. We propose a business-semantics-centric formulation in which states, dynamics, and feasible actions are linked to business entities, their attributes, and their relationships. Within this framework, intelligent agents can simulate alternative action sequences, estimate their effects on future business outcomes, and evaluate trade-offs under uncertainty and business constraints. We present a reference design and open-source implementation in which LLM agents interpret natural-language business goals and explain results, while a deterministic, metadata-driven simulation engine integrating machine learning models, business rules, and action spaces performs all calculations. A churn case study of a digital subscription business demonstrates that the system recommends options meeting the churn-reduction target within the revenue constraint. This work establishes a conceptual foundation for autonomous business systems capable of moving from instruction-based execution toward goal-driven planning, optimization, and execution.

cs.AI

版本更新 · v2arXiv:2610.08983

Zero-Shot Brain MRI Inpainting with 2.5D Unconditional Flow Priors

Arnela Hadzic, Franz Thaler, Simon Johannes Joham, Martin Urschler

原文摘要

Generative inpainting of brain MRI volumes is essential for synthesizing healthy tissue in pathological regions, improving the accuracy and reliability of automated downstream brain analysis applications such as image registration, brain extraction, and segmentation. However, standard 3D approaches are computationally prohibitive, while efficient 2D slice-wise methods suffer from severe inter-slice discontinuities. Furthermore, traditional models rely on conditional training, requiring task-specific learning of masked inputs. We propose a zero-shot brain MRI inpainting framework utilizing 2.5D unconditional flow priors to capture spatial context along the superior-inferior axis without the overhead of full 3D convolutions. During training, our flow matching model learns the joint distribution of adjacent axial slice triplets, modeling the manifold of healthy brain anatomy while explicitly excluding pathological regions from the loss function. At inference, the model processes the input triplets autoregressively along the depth axis. We employ the Restora-Flow solver to constrain the unconditional prior using the input mask, achieving accurate zero-shot inpainting. Evaluations show our 2.5D strategy resolves the structural discontinuities of 2D baselines, synthesizing plausible healthy tissue while maintaining volumetric consistency across the axial, sagittal, and coronal planes. As a final step, we generate and average an ensemble of multiple stochastic reconstructions to form the final prediction. Quantitative results benchmarked on the official BraTS 2026 Inpainting Challenge validation set demonstrate the effectiveness of our proposed approach, yielding an SSIM of 0.816 $\pm$ 0.112, MSE of 0.007 $\pm$ 0.005, and PSNR of 22.923 $\pm$ 4.343. Code is available at https://github.com/imigraz/brats2026-inpainting.

cs.CV

新论文 · v1arXiv:2611.26096

From Sessions to Serving: Integrating LLM-Based Embeddings into Retrieval for Production Content Recommendation

Zhixiong Hu, Meng Li, Maggie Zhang, Deborshi Goswami, David Bailey, James Bradley

原文摘要

Personalized content recommendation on large marketplace websites must reconcile heterogeneous candidate content (e.g., customer stories, product pages, industry pages, and learning content) with sparse, fast-changing user behavior, placing increasing demands on the retrieval stage to represent user interest well, not just the downstream ranking stage. We present LSER (LLM Session Embedding Retrieval), a production content recommendation system for autodesk.com that derives session- and history-level embeddings from a pretrained large language model (LLM) and adds them as auxiliary user-side features in the retrieval-stage user tower alongside existing learned sequence encoders. To keep user-side embeddings fresh under real-time behavior, we build a nearline serving path that asynchronously refreshes sequential and LLM-derived user vectors for our main content recommendation server. We validate the incremental value of the LLM features through an ablation study across our full multi-content-type catalog. Our results indicate that integrating LLM-derived embeddings as additional retrieval signals, together with production-grade nearline serving, is technically feasible and can improve model performances.

cs.IR · cs.LG

新论文 · v1arXiv:2611.26095

Approximate Adjoint Models for Memory and Compute-Efficient Online Learning of Hybrid Models

Emilio González-Zamora, Pierre Tandeo, Laurent Debreu, Said Ouala

原文摘要

Differentiable programming enables online calibration of numerical models by minimizing errors accumulated along simulated trajectories. This includes estimating physical parameters and training neural sub-models in hybrid configurations. Gradients are typically computed using adjoint methods, or backpropagation through time (BPTT), whose computational and memory costs grow with the length of the training window. This limits the duration and resolution of hybrid-model training, a major bottleneck for realistic geophysical applications. Approaches like truncated BPTT reduces memory use by discarding sensitivities across truncation boundaries, introducing gradient bias and neglecting long-range dependencies. We propose a class of gradient approximation methods that preserve the adjoint recurrence over the full trajectory while approximating only the pullback operators within it. Exact pullbacks are computed by automatic differentiation at sparse anchor states only. Between anchors, sensitivities propagate either through the identity (Zero-order) or through a reused, frozen pullback (Pullback Reuse). This approximation makes the parameter pullback constant within each chunk, allowing the gradient sum to factorize so that a single evaluation per chunk replaces one per time step. We evaluate both on three systems of increasing complexity, an incomplete Lorenz-63 model, a large-eddy closure for two-dimensional quasi-geostrophic turbulence, and bias correction in a multilayer quasi-geostrophic ocean model. Peak memory falls by up to an order of magnitude and runtime up to fourfold, with Pullback Reuse staying closest to the exact gradient direction. In all three cases, approximate gradients support hybrid-model training at comparable accuracy, showing that effective online learning does not require exact gradients.

cs.LG · math.OC · physics.ao-ph

新论文 · v1arXiv:2611.26094

Phantom of the Masquerade: Graph Neural Network-Based Detection of Masquerade Attacks in Controller Area Networks

Nehal Ameen, Pablo Moriano, Irfan Ahmed

原文摘要

The lack of built-in security in controller area networks (CANs) exposes modern vehicles and industrial systems to masquerade attacks that can compromise safety and reliability. These attacks inject malicious messages that mimic legitimate communication patterns, making them difficult to detect using existing intrusion detection systems (IDS). Prior graph-based IDS approaches often rely on shallow embeddings that decouple representation learning from classification, which may limit their ability to fully exploit structural and signal-level information. In this paper, we propose PhantoMasq, a graph neural network (GNN)-based framework for detecting masquerade attacks in CAN systems. PhantoMasq models CAN traffic as message sequence graphs enriched with lightweight statistical features derived from signal behavior, enabling joint learning of structural dependencies and temporal characteristics. The detection task is formulated as a graph-level classification problem and solved using graph convolutional networks (GCNs). We evaluate PhantoMasq on the ROAD dataset under a protocol that prevents overlapping windows from spanning training and test data. Across five masquerade attack types, PhantoMasq achieves mean AUC-ROC between 0.831 and 0.999, with performance depending on the transmission rate of the targeted identifier and the magnitude of the induced signal deviation. Against a state-of-the-art graph-embedding baseline evaluated on the same sliding-window configurations, PhantoMasq achieves higher AUC-ROC on all five attack types, by 0.13 to 0.36, and is less sensitive to window configuration. Window-processing latency ranges from 2.33 to 5.94 ms per window on CPU, corresponding to a throughput of 168 to 429 windows per second. We make our implementation and experimental artifacts available to support reproducibility.

cs.CR · cs.LG

版本更新 · v4arXiv:2606.08918

GeoMetric: Injecting Metric Geographic Structure into Worldwide Image Geo-Localization

Junchao Cui, Wenqi Shi, Xiangyang Luo, Biru Zhu, Xuanzi Ma, Nan Wu

原文摘要

Worldwide image geo-localization estimates where a photograph was taken on Earth. Existing methods infer location from visual appearance alone, an unreliable cue for geography: near-identical scenes occur on different continents. We present GeoMetric, a three-phase retrieval framework that encodes geographic distance directly in the representation. Phase I trains image, text, and location encoders into a unit hypersphere. The location encoder combines ScaleFusion and GeoAnchorAttention. ScaleFusion projects a coordinate onto an equal-area chart and encodes it with three Fourier bands, fusing them by attention into one multi-scale embedding. GeoAnchorAttention adds a global prior from fixed spherical anchors without mixing samples, so each coordinate embedding is batch-independent. A trimodal contrastive objective aligns images, reverse-geocoded texts, and coordinates, masking semantic and geographic false negatives. Phase II retrieves candidates, each carrying a GPS coordinate and granularity-aware geographic text. Phase III lets a frozen open-source 8B-parameter large multimodal model (LMM) choose the coordinate from those candidates. On four benchmarks, GeoMetric is best in 15 of 20 threshold cells, and ablations show that the components act at different scales: the text encoder carries coarse geographic context, while the location encoder resolves fine-grained distinctions.

cs.CV

新论文 · v1arXiv:2611.26092

Control of Complex Social Systems via Committed Minorities and Opinion Broadcast

Roberta Raineri, Mengbin Ye, Lorenzo Zino

原文摘要

Complex social systems are often characterized by a deep interconnection between individuals' opinion formation and their decision-making processes. In this chapter, we show how these two dynamics can be incorporated within a unifying game-theoretic coevolutionary model for actions and opinions. We present the main theoretical results concerning the convergence of the coevolutionary model, and its recent extensions to incorporate further features of real-world social systems. Then, we focus on the problem of controlling such a coevolutionary model. In particular, we present two distinct approaches: i) introducing a committed minority in the population, and ii) broadcasting information to some target individuals. For both of these methods, we discuss their implementation and provide theoretical guarantees on their performance. Finally, we demonstrate the effectiveness of the proposed control methods on different case studies, illustrating their implementation on simple academic examples, as well as exploring their performance in realistic scenarios using network structures inferred from real-world data. In summary, this chapter provides an overview of the mathematical approaches developed to model a population of individuals with coevolving actions and opinions, providing effective tools to study and, ultimately, control this complex socio-technical system.

eess.SY · cs.SI · math.DS · math.OC

新论文 · v1arXiv:2611.26091

The Density of Interpolants in Modal Logic: First Results

Roman Kontchakov, Louwe Kuijer, Frank Wolter

原文摘要

We initiate a quantitative study of interpolation in propositional modal logic. Instead of asking only whether a logic has Craig or uniform interpolation, we measure the density of formulas, or pairs of formulas, admitting interpolants. For quasi-normal modal logics, formulas of bounded modal depth over finite signatures are counted modulo logical equivalence. We show that interpolation failure is local to cells determined by the retained signature, yielding product decompositions of both uniform and Craig interpolation densities. These decompositions allow us to analyse their asymptotic behaviour as either the retained or the forgotten signature grows. When the retained signature grows, we obtain general results for all quasi-normal modal logics; when the forgotten signature grows, the behaviour is subtler, particularly when formulas are counted after existential quantification over forgotten variables. For locally tabular logics, we introduce cell incompatibility graphs and use them for a fine-grained analysis covering tabular logics, the difference logic Diff, and bounded-depth extensions of Grz.

cs.LO

版本更新 · v2arXiv:2611.15073

When Does Causality Help in Multivariate Time-Series Anomaly Detection?

Long Zhang, Wei Huang, Aishui Rao, Jian Hu, Binxiang Wu, Jie Li, Chen Gong, Linxi Feng

原文摘要

Causality-aware scoring is common in multivariate time-series anomaly detection, yet the literature rarely states when it helps. Holding one causality-guided autoencoder (DeCauAE) fixed, we vary graph quality, anomaly type and protocol against ground truth. Graph quality fixes the payoff: on a twenty-variable DAG the gain over a graph-free autoencoder decays as the graph is corrupted. Anomaly type fixes the target: the structural factor lifts window F1 from 0.800 to 0.976 where anomalies break conditional structure, and its advantage vanishes below a break severity s 0.3. Protocol fixes the ranking: point-adjusted F1 lets a random score match the full model, while range-aware VUS separates the autoencoder family from the classical baselines. On real benchmarks the causal module pays only where density makes a system eligible and cross-subsample reproducibility selects entities inside it; enabling it everywhere costs 0.012 VUS-ROC. The structural factor also transfers to classical detectors .

cs.LG

新论文 · v1arXiv:2611.26090

Dynamic Human-Robot Role Allocation through Entropy-Based Monitoring and LLM Reasoning

Elena Merlo, Marta Lagomarsino, Hamidreza Raei, Arash Ajoudani

原文摘要

Role allocation strategies in human-robot collaboration typically assign tasks to both the human and the robot, limiting the user's ability to take initiative. Furthermore, allocation decisions may conflict with the user's intentions, and the rationale behind such allocations may remain unclear, potentially reducing both acceptance and collaboration fluency. The novelty of the proposed framework is that the human is left free to contribute to the task, while the robot is dynamically allocated activities through a cost-driven Behavior Tree architecture according to user-expressed preferences and behavior. Explicit preferences are extracted from natural language instructions using a large language model, whereas human activity is monitored online through a Shannon entropy-based analysis of hand-object interactions. Furthermore, the framework can recognize anomalies in human execution and autonomously trigger corrective robot actions. The approach was validated through a pilot study and a multi-subject experiment involving 14 participants. Results demonstrate the potential of such an allocation strategy for enabling intuitive and engaging collaboration without increasing the perceived workload.

cs.RO

版本更新 · v2arXiv:2607.24542

From transcription to semantic corpus analysis: unsupervised learning of sentence representations for ancient languages

Th{é}otime de la Selle

原文摘要

Automatic Text Recognition (ATR) now supplies digital humanities with large volumes of unstruc- tured, heterogeneous, and often noisy text in ancient languages. Downstream semantic analyses- text reuse identification, alignment, and semantic search-rely on sentence embeddings, yet ex- isting methods transfer poorly to ancient languages: generic multilingual encoders underperform, specialized language models yield anisotropic representation spaces, and labeled similarity data is scarce. We study two fully unsupervised strategies-TSDAE and unsupervised SimCSE (hence- forth CSE)-that adapt a specialized token-level language model into a corpus-specific sentence encoder using only raw sentences. We validate them on biblical reuse in patristic literature in Latin and Ancient Greek: 2,936 sentences carrying expert-verified reuses in works of Augustine, Jerome, and Athanasius, matched against the Vulgate, the Septuagint, and the Nestle--Aland New Testament. We decompose reuse identification into binary detection and correspondence retrieval, evaluate both tasks separately, and benchmark the adapted encoders against multilingual, specialized, distilled, and supervised fine-tuned baselines, on clean text and on text perturbed by simulated HTR errors and scribal abbreviations. On these test sets, the adapted encoders lead all baselines on both tasks, with complementary profiles: TSDAE leads detection given a large in-domain corpus, while CSE leads retrieval, reaches its optimum with as few as 4--8k raw in-domain sentences-a few tens of sec- onds of training on a laptop GPU-and transfers across works and authors. Under simulated noise, retraining CSE on the perturbed text itself outperforms a clean-trained encoder. Measurements of alignment, uniformity, and anisotropy relate the geometric effect of each strategy to the measured gains. The full pipeline-segmentation, fine-tuning, cross-corpus semantic search-is available to non-specialists through the online tool Paraphrasis.

cs.CL · cs.DL · cs.IR

版本更新 · v4arXiv:2601.22978

Triosecuris: Formally Verified Protection Against Speculative Control-Flow Hijacking

Jonathan Baumann, Yonghyun Kim, Yan Farba, Catalin Hritcu, Julay Leatherman-Brooks

原文摘要

This paper introduces Triosecuris, a formally verified defense against Spectre BTB, RSB, and PHT that combines CET-style hardware-assisted control-flow integrity with compiler-inserted speculative load hardening (SLH). Triosecuris is based on the novel observation that in the presence of CET-style protection, we can precisely detect BTB misspeculation for indirect calls and RSB misspeculation for returns and set the SLH misspeculation flag. We formalize Triosecuris as a transformation in Rocq and provide a machine-checked proof that it achieves relative security: any transformed program running with speculation leaks no more than what the source program leaks without speculation. This strong security guarantee applies to arbitrary programs, even those not following the cryptographic constant-time programming discipline.

cs.CR · cs.PL

新论文 · v1arXiv:2611.26086

ScaleWire: An End-to-End Scale-Up Ethernet Simulation Framework for LLM Systems

Kai Ma, Quanfeng Lv, Bowei Dai, Jingguo Ge, Kefan Ruan, Ruoqi Li

原文摘要

LLM training and inference rely on scale-up networks for high-bandwidth, low-latency memory-semantic communication among accelerators (XPUs). Ethernet is emerging as an open interconnect for scale-up systems. In Ethernet systems, network-interface (NI) mechanisms across transport, network, and link layers interact, jointly shaping communication performance and on-chip resource requirements. We present ScaleWire, an end-to-end scale-up Ethernet simulation framework for LLM systems. It couples packet-level simulation of cross-layer NI mechanisms with a resource exploration engine that combines analytical estimates and simulation to size on-chip resources. We validate ScaleWire against hardware measurements and nine analytical checks, and evaluate six NI designs across four LLM communication workloads. For DeepSeek-R1 dispatch, NI design choices lead to a 4.5$\times$ difference in per-XPU on-chip storage budget. In grid-comparison experiments, the exploration engine uses 6-19% of the simulations required by exhaustive search. Using vLLM prefill traces with kernel times measured on GH200 GPUs, ScaleWire further guides resource provisioning under a completion-time constraint for inference execution. ScaleWire thus helps designers select NI mechanisms and provision on-chip resources to balance communication performance and resource cost.

cs.NI

新论文 · v1arXiv:2611.26083

Reference-Twist-Based Velocity-Domain Spatiotemporal Calibration for mmWave Radar, Event Camera, and IMU

Zhenghao Zou, Yang Lyu, Xiaohu Guo, Chunhui Zhao, Quan Pan

原文摘要

Multi-sensor navigation requires a common spatial and temporal basis, but geometric calibration relies on targets, features, point clouds, or pose associations that are difficult to establish across mmWave radars, event cameras, and inertial measurement units (IMUs). These sensors nevertheless share rigid-body motion. This paper formulates reference-conditioned velocity-domain spatiotemporal calibration for a rigid suite of two mmWave radars, one event camera, and two IMUs. Given a trusted reference body twist, radar Doppler, event normal flow, and inertial measurements are regressed against the same motion to estimate relative rotations, lever arms, time offsets, and sensor-specific bias and scale without artificial targets, structure from motion, or cross-modal point-cloud registration. The reference is treated as a conditioning input rather than a sensor to be calibrated. We compare joint and reference- conditioned estimation, derive conditional identifiability and CRLB-level covariance and bias bounds, and state the operating conditions under which conditioning is informative. Indoor and forest recordings with axis-separated motion evaluate initial- ization, temporal alignment, motion degeneracy, reference-input sensitivity, and residual consistency. Because the reference does not add physical observability, the estimates remain conditional on reference accuracy, synchronization, motion excitation, and projection diversity.

cs.RO

新论文 · v1arXiv:2611.26081

Learning to Compare Neural Representations from Parameterized Computation Graphs

Ya-Wei Eileen Lin, Daniel Herbst, Stefanie Jegelka

原文摘要

Comparing internal representations across trained models is important for identifying corresponding layers and selecting components for model reuse. Yet, direct comparisons at inference require evaluating candidate models on known reference inputs and collecting their internal activations. In this paper, we study whether internal representation similarity measured across independently trained models can instead be learned to predict in unseen models from their parameterized computations. We introduce a framework that applies a graph metanetwork for mapping each internal site to an embedding derived from its upstream computations while accounting for hidden-coordinate permutation and internal attention-basis transformations. We analyze the structural properties of our estimators and derive conditional guarantees for layer retrieval and stitching candidate selection. Empirically, we find that the learned predictors support relation estimation and low-regret layer retrieval. They also select useful stitching connections while not being trained on stitching outcomes. This opens up a new way to study representation similarity directly from trained neural networks: representational relations learned across a collection of trained models can generalize to new computations, and that the predicted relations can guide candidate selection before direct measurement or adapter fitting for functional evaluations.

cs.LG

版本更新 · v3arXiv:2506.16557

Scaling GR(1) Synthesis via a Compositional Framework for LTL Discrete Event Control

Hernan Gagliardi, Victor Braberman, Sebastian Uchitel

原文摘要

We present a compositional approach to controller synthesis of discrete event system controllers with linear temporal logic (LTL) goals. We exploit the modular structure of the plant to be controlled, given as a set of labelled transition systems (LTS), to mitigate state explosion that monolithic approaches to synthesis are prone to. Maximally permissive safe controllers are iteratively built for subsets of the plant LTSs by solving weaker control problems. Observational synthesis equivalence is used to reduce the size of the controlled subset of the plant by abstracting away local events. The result of synthesis is also compositional, a set of controllers that when run in parallel ensure the LTL goal. We implement synthesis in the MTSA tool for an expressive subset of LTL, GR(1), and show it computes solutions to that can be up to 1000 times larger than those that the monolithic approach can solve.

cs.SE

新论文 · v1arXiv:2611.26080

Running-Time Lower Bounds under FPT $\neq$ W[1]

Hendrik Molter

原文摘要

We revisit running-time lower bounds historically based on the Exponential Time Hypothesis. Our case study asks how much of their quantitative content can be recovered under the assumption FPT $\neq$ W[1]. The reductions encode the choice of a clique in few bits and control the target parameter in terms of that encoding. This makes a sufficiently fast algorithm for the target problem yield a fixed-parameter algorithm for deciding whether a graph has a clique of the requested size. We obtain twelve lower bounds for deterministic algorithms across graph, string, and integer feasibility problems, including bounds that hold under structural restrictions. In several cases the same function of the parameter appears in the exponent as in the historical bound, with the additional requirement that the cutoff for any prescribed saving in the exponent be computable.

cs.DS · cs.CC

版本更新 · v2arXiv:2608.30429

EvoSkill Injection: Red-Teaming Autonomous Skill Generation and Evolution in Self-Evolving Agents

Doyun Kim, Chanwoo Kim, Sugyeong Eo, Yeo-Chan Yoon, Chanjun Park

原文摘要

LLM-based agent systems increasingly adopt skill-based architectures to reduce repetitive reasoning costs and improve stable, efficient task execution. Recent studies propose self-evolving agents that autonomously generate, refine, and reuse skills from past experiences to enable continuous capability evolution. However, autonomous skill evolution introduces a new attack surface in which malicious capabilities are generated, stored, and reused as legitimate skills. In this paper, we define EvoSkill Injection as a threat model targeting the autonomous skill generation and evolution pipeline of self-evolving agents. We further propose SARGE (Red-teaming Autonomous Skill Generation and Evolution in self-evolving agents), a red-teaming framework for evaluating this threat model through iterative generation, escalation, and reinforcement interactions. To support our framework, we construct EvoSkillBench, a benchmark dataset of malicious interaction trajectories for inducing malicious skill formation in self-evolving agents, and introduce EvoSkillSafetyBench, a post-attack benchmark for evaluating whether injected malicious skills are subsequently retrieved and activated as harmful behaviors. Our evaluation shows that SARGE induces malicious skill formation and that injected skills are persistently stored and repeatedly activated, highlighting the risk of persistent capability corruption.

cs.AI · cs.CL

新论文 · v1arXiv:2611.26077

Same Judgement, Different Scores: Numerical Scoring Topology in LLM-as-a-Judge Evaluation

Doyun Kim, Chanwoo Kim, Jiwoo Kum, Minsoo Song, Chanjun Park

原文摘要

LLMs are increasingly used as automated evaluators to reduce the cost and time required for human annotation. However, the role of scoring-scheme design in the reliability of LLM-as-a-Judge evaluation remains underexplored. To address this problem, we introduce numerical scoring topology, which characterizes the structural design of a scoring scheme in terms of its scale width, granularity, offset, symmetry, and boundary placement. Building on this definition, we conduct the first systematic study of how numerical scoring topology influences zero-shot LLM-as-a-Judge evaluation across diverse benchmarks and judge models. Our results show that numerical scoring topology systematically affects judge scores, calibration, and robustness, particularly in judgment-intensive evaluations. Shifted and narrow ranges tend to induce some of the largest score deviations across many settings. We further find that stronger and larger judge models generally achieve higher discriminative validity and greater robustness to scoring-topology changes. These findings suggest that numerical scoring topology should be treated as an integral component of LLM-as-a-Judge evaluation and highlight the need for more principled and standardized scoring protocols.

cs.CL

新论文 · v1arXiv:2611.26074

Controlling the Humanoid Robot H1-2 With a Simulated Biologically Plausible Network

Fatima Atwi, Constant Roux, Christian Bergaud, Olivier Stasse

原文摘要

Artificial Spiking Neural Networks have gained remarkable attention in the last few years due to their proven ability in simulating biological neural cultures and producing similar activity and trends. Nevertheless, whether they can be trained for humanoid locomotion while retaining biological plausibility and robust performance at once remains an open question. We study this problem by proposing a biologically inspired spiking network, trained by different learning schemes and tested on the H1-2 Unitree Humanoid. These schemes differ in how recurrent weights are trained. Ordered from most to least biologically faithful, they are: a fixed reservoir mimicking an in-vitro neural culture, where recurrent weights are kept frozen and only a linear readout is trained, the same network but with its synapses trained by reward-modulated spike timing dependent plasticity and eligibility propagation, and finally backpropagation through time as a non biological baseline for comparison. All schemes relied on distillation of a reinforcement learning expert policy with dataset aggregation. We show that e-prop reached 100% survival rate matching backpropagation through time, whereas both the fixed reservoir and the reward-modulated STDP, though the closest to biological cultures, plateaued at a much lower performance (~43.7% and ~11% survival). We explore in further detail which trends a real culture could and could not produce until now. Finally we demonstrate brief bipedal stepping on a full size humanoid H1-2.

cs.AI · cs.RO

新论文 · v1arXiv:2611.26073

AltRA-Test: An Alternative Annotator Test for LLM Judges Without Shared Annotators

Nagham Omar, Rotem Dror, Almog Gueta, Avigdor Gal, Roi Reichart, Nitay Calderon

原文摘要

Large language models (LLMs) are increasingly used as judges and graders to replace human annotators, raising the question of when such replacement is statistically justified. To support such decisions, the Alternative Annotator Test (Alt-Test) was introduced for dense settings where all items are annotated by the same annotators. However, in many large-scale annotation settings, items are annotated by different subsets of annotators. We therefore propose the Alternative Annotator Test with Resampled Annotations (AltRA-Test), a statistical test that does not require that all items are annotated by the same annotators. AltRA-Test uses a bootstrap that resamples items, draws k human annotations for each item, and substitutes one of them with an LLM annotation. This produces paired all-human and mixed human-LLM panels. A replacement is supported when the inter-annotator agreement (IAA) of the mixed panel is statistically non-inferior to the IAA of the all-human panel, i.e., when the LLM preserves the signal that the human annotators share. We establish that AltRA-Test is a consistent test of asymptotic level alpha. We then evaluate it through simulations and real-world experiments in which we test 18 LLMs and 5 prompting techniques on 20 datasets across text, image, audio, and video. Across these evaluations, about 100 items often suffice for the test. In simulations, where we know whether the LLM is a better or a worse annotator than the humans, AltRA-Test makes more correct replacement decisions than existing baselines. In real-world data experiments, LLMs pass the test in about half of the setups, depending mainly on the task. Since replacement is not always justified, we encourage researchers to test it for their task.

cs.CL · cs.AI

新论文 · v1arXiv:2611.26072

DfD-GS: Defocus Image Prior for Metric-Scale Single-View Depth Estimation and View Synthesis

Shiori Ueda, Rinto Yagawa, Hideo Saito, Shohei Mori

原文摘要

Recovering metric-scale geometry from observations at a single viewpoint remains fundamentally ill-posed, limiting the reliability of novel view synthesis. Recent methods rely on learned monocular depth priors, which provide strong structural estimates but can suffer from scale ambiguity and dataset-dependent biases. We propose DfD-GS, a feed-forward framework that predicts a metric-scale 3D Gaussian scene representation from a single-view all-in-focus--defocused image pair. The key idea is to exploit defocus blur as a physical metric depth cue. DfD-GS first estimates a defocus-based metric disparity map from the all-in-focus--defocused image pair and the corresponding defocus strength. This estimate is used to align and refine a monocular depth prior to metric scale. Visible and hidden 3D Gaussians are then predicted for occlusion-aware novel view synthesis. Experiments on synthetic-defocus datasets constructed from Booster and Middlebury show that DfD-GS outperforms single-image baselines in metric depth estimation and novel view synthesis. On WildRGB-D, DfD-GS achieves lower AbsRel and LPIPS but higher depth RMSE than the single-image baselines. Additional analyses show that DfD-GS maintains a more stable metric scale under field-of-view changes. Qualitative results on a real dataset show recovery of the main scene structure under optical defocus.

cs.CV

新论文 · v1arXiv:2611.26069

SheetAlign: Grounded Interaction and Stage-Isolated Learning for Spreadsheet Agents

Yuxiang Wang, Junhao Gan, Jianzhong Qi

原文摘要

Spreadsheets are widely used for data analysis and supporting decision-making. Automating spreadsheet analysis while protecting sensitive data and controlling deployment costs motivates locally deployed open-weight large language models (LLMs). Meanwhile, complex spreadsheet tasks require both task reasoning and workbook understanding, while predefined operation interfaces impose additional tool-selection and invocation demands. We introduce SheetAlign, a spreadsheet manipulation framework with stage-isolated policy learning, organized around where to operate, what to change, and how to implement transformations. Task-guided inspection and API grounding anchor code generation in workbook observations and available operations, while stage-isolated learning optimizes program generation and review using edit-progress rewards. At inference, execution consistency selects programs through agreement among workbook states. On SpreadsheetBench, our grounded task interface improves Qwen3-4B's accuracy from 13.5% to 29.2% on average without training. With learning and execution consistency, SheetAlign achieves 32.8% accuracy on average on SpreadsheetBench and 30.1% on Domain-Spreadsheet, outperforming comparable baselines. The code will be made publicly available upon publication.

cs.CL · cs.SE

新论文 · v1arXiv:2611.26068

Physics-Informed Deep Learning for Channel Prediction in Satellite-Mounted Metasurface Reflection Communication Networks

Aolin Sun, Ruihong Jiang, Jincong Mo, Yangbo Guo, Zhuangzhuang Cui

原文摘要

Satellite-mounted metasurfaces (MS) provide a low-power programmable paradigm for 6G satellite-terrestrial networks, effectively compensating for severe free-space path loss under stringent payload constraints. However, accurate channel prediction for such systems remains fundamentally challenging due to the coupling of dynamic orbital topology, stochastic user geometry, and hardware-imposed phase quantization. This paper proposes a physics-informed deep learning (PIDL) framework that embeds the physics of wireless communication, including monotonicity constraints and asymptotic decay properties, into a deep neural network for outage probability (OP)-aware channel prediction. We construct a realistic system model incorporating spherical user distribution, Spherical Poisson Point Process-based satellite constellation, and practical metasurface quantization. Simulation results demonstrate that the proposed framework achieves over 0.99 prediction accuracy with physically consistent outage trends under varying geometric configurations, constellation densities, and quantization resolutions.

cs.IT

版本更新 · v4arXiv:2611.15592

Small Heads, Strong Decisions: A Controlled Study of Typed-Decision Models

Ritvik Rastogi

原文摘要

In September 2026, TypeSafe announced Jev as the first model in its System One category: unstructured state goes in, and typed, calibrated decisions that software can use directly come out. Jev's architecture and training data are unpublished, but many open typed-decision models have followed. They answer structured questions about a piece of text, such as a yes/no check, a choice among labelled options or a graded score, by reading the text with a language model and scoring the options with a small readout. Each pairs a different readout with a different backbone, training mix and recipe, so it is unclear which design choices matter. We re-implement five readouts and train all of them under one protocol: the same backbone, the same 15 public English training sources, the same recipe with a per-head learning-rate sweep, and the same seed. Our experiments demonstrate that: The complexity of the head does not affect accuracy: on Qwen3-0.6B every readout reaches 81.3-82.0% in-domain accuracy, and a 0.5M-parameter cosine head, which scores each option by its cosine similarity to the question, matches the 27M-parameter joint schema head adapted from Clef in-domain at half its inference time. Reading all of a text's questions from one prompt also matches one prompt per question in accuracy, with 1.6 times faster inference and 2.1-2.5 times faster training on one H100. The backbone's size matters more than either choice: moving from Qwen3-0.6B to Qwen3-4B adds 4 points in-domain and 11 points on held-out sources, while the base, instruct and thinking checkpoints of the same model tie. Checkpoints trained further for embedding or reranking bring no gain. We recommend spending effort on the backbone and keeping the head minimal.

cs.AI

版本更新 · v2arXiv:2408.12100

A Unified Plug-and-Play Algorithm with Projected Landweber Operator for Split Convex Feasibility Problems

Shuchang Zhang, Hongxia Wang

原文摘要

In recent years Plug-and-Play (PnP) methods have achieved state-of-the-art performance in inverse imaging problems by replacing proximal operators with denoisers. Based on the proximal gradient method, some theoretical results of PnP have appeared, where appropriate step size is crucial for convergence analysis. However, in practical applications, applying PnP methods with theoretically guaranteed step sizes is difficult, and these algorithms are limited to Gaussian noise. In this paper,from a perspective of split convex feasibility problems (SCFP), an adaptive PnP algorithm with Projected Landweber Operator (PnP-PLO) is proposed to address these issues. Numerical experiments on image deblurring, super-resolution, and compressed sensing MRI experiments illustrate that PnP-PLO with theoretical guarantees outperforms state-of-the-art methods such as RED and RED-PRO.

cs.CV

版本更新 · v6arXiv:2603.28584

ORSIFlow: Saliency-Guided Rectified Flow for Optical Remote Sensing Salient Object Detection

Haojing Chen, Zhihang Liu, Yutong Li, Tao Tan, Haoyu Bian, Qiuju Ma

原文摘要

Optical Remote Sensing Image Salient Object Detection (ORSI-SOD) remains challenging due to complex backgrounds, low contrast, irregular object shapes, and large variations in object scale. Existing discriminative methods directly regress saliency maps, while recent diffusion-based generative approaches suffer from stochastic sampling and high computational cost. In this paper, we propose ORSIFlow, a saliency-guided rectified flow framework that reformulates ORSI-SOD as a deterministic latent flow generation problem. ORSIFlow performs saliency mask generation in a compact latent space constructed by a frozen variational autoencoder, enabling efficient inference with only a few steps. To enhance saliency awareness, we design a Salient Feature Discriminator for global semantic discrimination and a Salient Feature Calibrator for precise boundary refinement. Extensive experiments on multiple public benchmarks show that ORSIFlow achieves state-of-the-art performance with significantly improved efficiency.

cs.CV

新论文 · v1arXiv:2611.26065

Domain-Aware Conformal Prediction for Trustworthy Spacecraft Pose Estimation under Sim-to-Real Shift

Nidhal Eddine Chenni, Arunkumar Rathinam, Djamila Aouada

原文摘要

Autonomous rendezvous and proximity operations require not only an accurate estimate of the pose of the target spacecraft, but also a reliable indication of when this estimate can be trusted. Conformal prediction provides such an indication with a formal guarantee, by producing pose uncertainty sets that contain the true pose with a specified probability. Such a guarantee, however, assumes that calibration and test data follow the same distribution, an assumption that is violated in orbital perception, where labelled calibration images are available only from synthetic renderings while operational images are real. In this work, we show on the SPEED+ dataset that conformal pose estimation calibrated on synthetic images loses its coverage guarantee on real images, and that many of the resulting failures are silent: the predicted pose set appears valid but excludes the true pose. To address this, we calibrate on synthetic images translated to the appearance of real images. Since translation alters appearance while preserving geometry, the synthetic pose labels remain valid and no labelled real images are required. Experiments with two translation models show that this calibration restores coverage above the specified level for miscoverage levels of 0.1 and above, and reduces silent failures from up to 43.5% to at most 3.3% of the test images, bringing conformal pose estimation a step closer to trustworthy deployment in space.

cs.CV

新论文 · v1arXiv:2611.26062

GenSPARQL: A Generative Extension of SPARQL for Incomplete Knowledge Graphs

Jingcheng Wu, Yi Wang, Hongkuan Zhou, Steffen Staab, Bo Xiong

原文摘要

Knowledge graphs on the Web are incomplete, and SPARQL answers a query only from the stored triples. Large language models (LLMs) can predict missing facts, but an LLM returns names instead of IRIs and requires one call per binding of a query variable. We present GenSPARQL, a generative extension of SPARQL 1.1 with an operator GENOP that binds values generated by an LLM to query variables during query evaluation. A similarity join maps generated names to entities of the graph, and the query planner invokes the LLM only for the bindings selected by the other graph patterns and filters. In a web workbench, attendees query two graphs built from Wikidata and inspect generated values, mapped entities and LLM calls. In our evaluation with three open-weight LLMs, the embedding similarity maps 83% of the entities named by the LLMs and yields only one incorrect match.

cs.DB · cs.AI

新论文 · v1arXiv:2611.26060

Iterative Scarcity-Guided Exploration: Bootstrapping Generative Auto-bidding from Narrow Support

Qingmao Yao, Guangzheng Hu, Yusen Huo, Zhilin Zhang, Chuan Yu, Jian Xu, Xu Chu, Bo Zheng 等 9 位作者

原文摘要

Auto-bidding is a core algorithmic component in online advertising auctions. In practical cold-start settings, iterative training is commonly used to compensate for low-quality, narrowly supported historical data. Unfortunately, in iterative training, representative diffusion-based AI-Generated Bidding (AIGB) methods fail to sustain extrapolation beyond the narrow data support and stagnate at a suboptimal level. In this paper, we theoretically attribute this stagnation to Signal-to-Noise Ratio (SNR) collapse: the weak radial return signal is overwhelmed by a curvature-induced penalty. To break this stagnation, we propose \textbf{Iterative Scarcity-Guided Exploration (ISGE)}, which introduces a guidance handoff: as the return signal collapses, scarcity subsequently takes over as an exploratory signal to elevate the SNR above a critical threshold. Specifically, ISGE iteratively alternates between a \emph{Judger} that assigns scarcity scores to trajectories and an \emph{Explorer} that performs guided generation of high-scarcity and high-return trajectories, thereby bootstrapping from the narrow support with such self-generated trajectories. Extensive experiments on the industrial AuctionNet benchmark demonstrate that ISGE, starting from cold-start datasets, effectively surpasses the performance of the Full-Dataset baseline within three iterations. Code is available at https://github.com/alimama-tech/ISGE

cs.LG