arXiv subjects

Search papers

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

At least 955 records · Page 53Linked to original sources

arXiv2026

Diviner: Agentic RCA with Human-Governed Adaptation for Large-Scale Infrastructure

Root-cause analysis (RCA) in large-scale distributed systems is challenging because a single symptom can cross service, data, configuration, compute, and organizational boundaries. Responders must assemble incomplete evidence and coordinate experts who hold different parts of the causal path. Generic language-model agents do not solve this problem through tool access alone: they can use stale or mistimed context, interpret failed queries as absent symptoms, report unsupported system state, and overload observability services designed for human-paced use. We present Diviner (Domain-Informed Verifiable Incident Navigation through Evidence and Review), an agentic RCA system operating at Meta and organized around a four-stage reliability lifecycle. Encode combines a large corpus of reviewed incidents with service-owner expertise to produce scoped, versioned skills, workflows, typed tools, and an incident knowledge graph. Execute uses an adaptive Lead and context-isolated workers to apply relevant knowledge to point-in-time evidence and maintain evidence-linked hypotheses. Evaluate compares saved investigation trajectories with responder-reviewed outcomes to measure exact cause, useful direction, and Groundedness. Adapt uses recurring gaps and on-call feedback to derive focused operational- knowledge updates, admitting them only after replay, quality checks, and owner approval.

cs.AI↗

arXiv2026

LICOA: Learning Instruction-Conditioned Optical Actions for Mobile Spatial Augmented Reality

Spatial augmented reality (SAR) augments physical environments with projected content, but most systems assume that the projector pose, target region, or motion path has been fixed or specified. We instead make projected light an embodied action: given a language instruction and current visual observations, a system must decide how to move and where and at what scale to place the projection. We introduce LICOA to realize this capability. Its differential-drive chassis provides room-scale approach and repositioning, while a steerable mirror with two degrees of freedom (2DoF) and projection scale provide rapid local control of the beam. A shared interface maps synchronized external and onboard RGB cameras with robot control state to physical optical actions across pretrained action models. Because instruction-conditioned optical action lacks dedicated training data and a common evaluation protocol, we construct a matched simulation embodiment and a corpus with 150 tasks and 7,500 rendered episodes, then evaluate models trained only on this corpus in previously unseen real-world environments. Across 480 complete closed-loop real-world execution trials (rollouts) on 12 tasks, the best adapted action model achieves a 57.50% success rate and 40.64% success weighted by path length, compared with 8.33% and 5.15% for the same architecture trained from scratch. A further 150 real-world ablation rollouts support the need for both camera views and coordinated control of the chassis and mirror. These results establish a physical platform, shared perception and optical action interface, corpus, and real-world evaluation protocol for treating projected light as an embodied action.

cs.RO↗

arXiv2026

Mallea: Building, Steering, and Memorizing Worlds with Coding Agent

Code world models represent world dynamics through explicit state and executable rules. We investigate how to coordinate object representations, behavioral revisions, and visual memory within this formulation. We present Mallea, a system for building, steering, and memorizing worlds from text. A coding agent infers object interaction requirements from the prompt and jointly plans object structure, properties, and rules, favoring compact representations that support those requirements. The world evolves under the planned rules; optional user actions steer it, while text requests can edit properties or extend capabilities with unaffected state preserved. Object texture memory retains reusable appearance only for whitebox-defined objects and adapts it as those objects change. The program advances world state, while a generative renderer combines current structure with remembered appearance. By organizing construction, intervention, and memory around the same objects, Mallea conditions subsequent observations on the latest state and compatible appearance. Project page: https://lxycopper.github.io/Mallea.

cs.AI↗

arXiv2026

Thermodynamics of Viscously Heated Circumplanetary Disks

Circumplanetary disks (CPDs) play an important role in the formation of gas giants and their satellites, yet remain poorly understood owing to the lack of observables and comprehensive theoretical models. Incorporating $α$-viscosity, a free-fall prescription of mass infall and boundary layer accretion, we develop an axisymmetric viscous radiation hydrodynamic (RHD) model. We evolve this model for different physical parameters using the RHD code {\tt Guangqi}, which solves RHD equations under the flux-limited diffusion approximation. In eight out of nine simulations, we obtain quasi-steady-state solutions and verify self-consistent energy conversion and transport. Overall, disk thickness increases with accretion rate and decreases with planet mass, consistent with the expected trend in conventional $α$-disk models. On the other hand, we highlight that the boundary layer plays a significant role in setting angular momentum transport and thermodynamics of the inner CPD, yielding ``hot'' ($s_{\rm real}>11\,k_{\rm{B}}\,m_{\rm{H}}^{-1}$) accreted gas, and potentially leading to dust sublimation. We identify a supersonic accretion flow, amounting to $2\%$--$5\%$ of the total accretion rate, located at high elevations above the CPDs with low angular momentum in our simulations. If the planet is massive ($\gtrsim3\,M_{\rm J}$), this low angular momentum flow impacting the planetary surface could generate observable H$α$ emission when the shock accretion rate is $\gtrsim5.5\times10^{-8}\, M_{\rm J}\,\rm{yr}^{-1}$.

astro-ph.EP↗

arXiv2026

A General Framework for Matching with Delays

We present a simple deterministic $O(\log n)$-competitive algorithm for the online minimum-cost perfect matching with delays (MPMD) problem on arbitrary, unknown metric spaces, where $n$ is the (unknown) number of requests. This improves upon the recent $O(\log^5 n)$-competitive algorithm of Dufay and Wattenhofer (SODA 2026) and nearly matches the $Ω(\log n / \log \log n)$ randomized lower bound of Ashlagi et al. (APPROX 2017). Our algorithmic framework rests on two ingredients: an online moat-growing process that grows active dual variables at a deliberately slowed constant rate $α< 1$ to control temporal asymmetry, and a delayed eligibility rule that matches requests in the same dual component only after their waiting time is proportional to their time-expanded distance. The framework extends naturally to arbitrary concave delay functions, maintaining an $O(\log n)$ competitive ratio, and to $k$-way matching with delays, yielding an $O(k \log n)$-competitive deterministic algorithm.

cs.DS↗

arXiv2026

The Density of Thought: Predicting and Steering Reasoning in Large Language Models

We investigate whether variation in information density across chain-of-thought (CoT) reasoning steps predicts answer correctness and can serve as a controllable inference-time target. We prompt frozen pretrained models to solve problems via CoT and measure the density of each reasoning step with two independent metrics: LP variance (how unevenly information is distributed across steps, computed as the variance in mean per-token negative log-likelihood) and the Caveman compression ratio. We then test whether density correlates with, and can be steered to improve, answer correctness. Experiments across six model scales, two training paradigms, and 60 GSM8K problems yield four main findings. First, lower LP variance predicts correctness in supervised fine-tuned (SFT) models, but this relationship weakens with scale and reverses in reinforcement-learning (RL) distilled models. Training paradigm therefore appears more influential than model size in determining density behaviour. Second, synthetic experiments reveal an intrinsic complexity preference: models trained on uniformly distributed data learn complexity-favouring distributions 2.0 to 2.6 times more accurately than simplicity-favouring ones. Third, direct inference-time density steering reduces accuracy from 60.0% to 8.3%, indicating that natural density patterns are functionally important to reasoning. Fourth, pre-generation nudging combining hint injection with contrastive KV-cache steering improves accuracy by 13.3 percentage points, consistent across both metrics. Finally, we introduce Entropy-Gated Density Routing (EGDR), a self-supervised method using token-level Shannon entropy to gate proportional KVcache adjustments during generation, improving accuracy by 10.0 percentage points without any external guidance.

cs.AI↗

arXiv2026

ReState: Noise-Level-Aware State Correction for Training-Free Object Removal in Few-Step Diffusion Models

Training-free object removal commonly redirects self-attention away from the target region, but on few-step distilled diffusion models this often leaves object-shaped residuals. We identify a temporal mismatch: a structured prediction forms inside the mask at the first forward, and the short trajectory offers few later chances to revise it, while attention rerouting acts mainly in later steps. Across four distilled backbones, magnitude-matched single-forward interventions show that background-directed state correction affects the final result more than attention rerouting. We therefore propose ReState, which corrects the masked clean prediction toward background evidence retrieved from the input using a noise-level-aware weight that vanishes at the final forward. ReState retains object-excluding attention as a complementary constraint and adds no denoiser evaluations. Across two paired benchmarks and two few-step backbones, ReState achieves the best masked PSNR, SSIM, and LPIPS among training-free methods, improving PSNR by 3.17 dB on average over the strongest baseline. On OBER-Test, it also outperforms the 50-step attention-forcing reference on all three metrics at 6.7x lower latency. The same configuration transfers to two additional distilled backbones without retuning. Project page: https://tdrlee.github.io/ReState

cs.CV↗

arXiv2026

Dynamical phase transition for a logarithm observable in chaotic systems

We consider the full distribution $P_N(A)$ of the dynamical observable $A = - \sum_{i=1}^N \ln x_i$ where $x_1, \dots, x_N$ is a sequence generated by a chaotic map $f(x)$ via $x_{n+1} = f(x_n)$. We study the particular cases where $f(x)$ is the doubling, tent or logistic map, and focus on the limit $N \gg 1$. Typical fluctuations of $A$ are Gaussian with mean $\left\langle A\right\rangle$ and variance which are both of order $\sim N$. We find that the left tail of the distribution ($A < \left\langle A\right\rangle$) scales as $P_{N}(A) \sim e^{-NI(A/N)}$, and calculate the rate function $I(a)$ which depends on the map $f(x)$. Interestingly, we find that the right tail of the distribution ($A > \left\langle A\right\rangle$) follows an anomalous scaling: $P_{N}(A) \sim e^{-N^{1/3} F((A-\left\langle A\right\rangle) / N^{2/3})}$. The large-deviation function $F(y)$ has a universal form for the three maps, differing only in the scaling constants. At a critical value $y=y_c$, it exhibits a corner singularity which has the character of a condensation transition. The transition separates between a subcritical phase $0< y y_c$ in which a small fraction of the $x_i$'s give a macroscopic contribution to the fluctuation. Despite this anomalous behavior of the right tail, we find that the cumulants of the distribution $P_N(A)$ grow linearly with $N$ at large $N$, and calculate the associated coefficients.

cond-mat.stat-mech↗

arXiv2026

A universal framework for Hopf algebras in combinatorics

A universal framework for Hopf algebras in combinatorics provides a unified understanding of combinatorial objects. To establish such a framework, we dualize the theory of combinatorial Hopf algebras developed by Aguiar, Bergeron, and Sottile, obtaining the category of rooted Hopf algebras. An object in this category is a graded connected Hopf algebra equipped with a group-like element, called a root, in its completion. These roots dualize the character theory of combinatorial Hopf algebras and provide a natural way to extend Bernstein and raising operators to other rooted Hopf algebras. The resulting constructions extend Schur, Hall--Littlewood, and Macdonald functions to rooted Hopf algebras. As an application, we develop these constructions for symmetric functions in noncommuting variables.

math.CO↗

arXiv2026

Bias Control and Injection for Multi-Agent Aggregation

Multi-agent reasoning with large language models (LLMs) raises a central question for collective decision-making: when agents disagree, whose decision should be trusted? Resolving such disagreements is difficult because agents' decisions are susceptible to non-semantic biases such as majority pressure and response-style preferences, which can obscure the quality of the underlying reasoning. Our analysis shows how bias can be both mitigated and exploited: bias control improves initial answer selection, while bias injection elicits decision instability that serves as a reliability signal. Building on these findings, we introduce a test-time aggregation framework that requires neither validation data nor prior knowledge of relative model strength. The framework implements bias control by presenting one style-normalized reasoning trace per distinct answer, and bias injection by placing each aggregating agent's initially selected answer in the minority. Overall, our framework achieves higher average accuracy and better token efficiency than competing non-oracle aggregation methods across three mathematical reasoning benchmarks and four heterogeneous LLM combinations. It also improves final-answer accuracy when applied to multi-agent debate and Mixture-of-Agents, without modifying their interaction or synthesis processes.

cs.AI↗

arXiv2026

Modularity criteria for multiple Eisenstein series

We give two criteria for a linear combination of multiple Eisenstein series of fixed weight to be a modular form, one in terms of its Fourier coefficients and one in terms of its $L$-function. The first criterion follows from an exact transformation formula under $SL_2(\mathbb{Z})$, which describes the modular defect of multiple Eisenstein series explicitly. The second criterion is a Hecke converse theorem for multiple Eisenstein series, which shows that the functional equation of the $L$-function alone already implies modularity.

math.NT↗

arXiv2026

MR-MPCC: Model Predictive Contouring Control for Robotic Manipulator Tracking under Moving Reference Frames

Maintaining small path-tracking errors under moving reference frames is essential for robotic manipulation tasks with tight execution tolerances, such as machining and surgical automation. Conventional time-indexed controllers, including tracking Model Predictive Control (MPC), rely on a fixed temporal correspondence with the reference trajectory and may therefore accumulate path-relative error under abrupt reference-frame motion or timing mismatch. This paper presents MR-MPCC, a Model Predictive Contouring Control (MPCC) framework for velocity-controlled robotic manipulators with 6-DoF end-effector pose regulation under Moving-Reference disturbances. A hierarchical control architecture is adopted, in which an upper-level MPCC optimizes path progress while penalizing contouring and orientation errors, and a constrained resolved motion rate control (RMRC) layer maps the commanded Cartesian twist to feasible joint velocities under joint-level and safety constraints. MR-MPCC is implemented in a ROS2--MuJoCo pipeline and validated on a UR5e platform under moving-reference disturbances, where the path frame is perturbed and tracked by electromagnetic sensors. Compared with PI control and time-indexed tracking MPC under matched disturbance, MR-MPCC achieves lower geometric path-following error and smaller deviations in the moving path frame, demonstrating improved path-following performance under reference-frame motion.

cs.RO↗

arXiv2026

Audio-Based Condition Monitoring for Agricultural Machinery: Stakeholder Needs, Adoption Barriers, and MVP Requirements

Failures of agricultural machinery are costly when breakdowns occur during narrow work windows and are discovered too late for planned service. Audio-based condition monitoring offers a low-cost retrofit sensing path because operators already treat abnormal sound as an informal diagnostic cue. Whether such systems are adopted, however, depends on how well they fit farm workflows, repair constraints, and stakeholder decision-making. This paper reports a stakeholder discovery study of audio-based condition monitoring for agricultural machinery. We synthesize semi-structured interviews with farm operators, dairy operators, dealership and service personnel, manufacturers, integrators, and domain experts, together with a short survey. Across sources, downtime avoidance emerged as the dominant value driver, and failures were most commonly recognized through unusual noise and loss of functionality. Survey responses further indicate a gap in maintenance visibility and low penetration of predictive tools. These findings narrow the opportunity from generic ``AI diagnosis'' to low-friction early warning and evidence generation for maintenance planning. Modern tractors appear to be a weaker initial wedge because many are already instrumented and supported by mature service programs. Implements, attachments, hidden rotating components, small-to-medium operations, and continuous-use settings such as dairy present clearer need because failures in these contexts are harder to perceive, sensing coverage is lower, and downtime consequences are higher. We translate these findings into requirements for a minimum viable product (MVP) that prioritizes maintenance actionability over full automation: commodity audio capture, contextual annotation, asset-specific baselines, persistent-deviation detection, temporally stable alerts, and evidence artifacts that support operator inspection and dealer triage.

cs.ET↗

arXiv2026

Can Capability Enhancement be Localized?

Transformer-based language models have shown strong capabilities across many tasks and domains. However, prior work shows that, for a given task, not all attention heads matter equally. Specifically, they suggest that many \textit{redundant} heads can be safely removed with little-to-no effect, while a small set of \textit{important} heads, when ablated, severely damages the model's performance on the given task. In this work, we investigate a complementary aspect, called \textit{Capability Enhancement Localization} (CEL): \textit{given a model and a task, does there exist a small set of heads whose removal improves the task performance?} We first investigate this question \textit{empirically} across models from different families. Using a simple two-stage routine that screens candidate heads and evaluates their joint ablations, we identify sparse head combinations whose removal improves task-specific performance without further fine-tuning. Motivated by these observations, we ask \textit{theoretically whether such a phenomenon can naturally arise from training dynamics, and under what conditions}. To this end, we construct a simplified multi-head softmax attention model and two tasks with conflicting features, and show that gradient descent from small random initialization provably concentrates the conflicting features into particular heads. We further show that ablating such heads improves performance on one task while harming the other, even though training converges to a global optimum of the joint objective. This establishes a possible training-induced theoretical explanation for CEL.

cs.LG↗

arXiv2026

Personalized Projected Ascent for Federated Stackelberg Optimization

Federated minimax problems, such as distributionally robust optimization and adversarial training, include an inner maximization variable built from each client's local data. These variables differ in meaning and dimension from one client to the next, so averaging them across clients, as standard federated minimax methods do, is ill-posed. We therefore study a federated minimax problem where the minimization variable $\mathbf{x}$ is shared and aggregated, while each client retains the maximization variable $\mathbf{y}_i$ (personalized). Because these applications constrain $\mathbf{y}_i$ to a compact convex set, we analyze projected ascent directly. We show that, in the nonconvex-concave regime, this projected personalized scheme attains the standard sublinear rate. In the nonconvex-PL regime, however, it hits a stepsize-invariant noise floor due to stochastic $\mathbf{y}_i$ updates. To mitigate this, we propose VR-SGDA, incorporating a variance-reduced ascent step where each client maintains a private variable snapshot. By evaluating current and snapshot gradients on the same sample, the constant variance is replaced by a vanishing drift under a two-timescale regime. VR-SGDA eliminates the noise floor, attaining sublinear leading rates while preserving personalization and projection. Experiments on subpopulation-robustness and NLP adversarial training confirm both the worst-group robustness benefits and the two-timescale tracking mechanism.

cs.LG↗

arXiv2026

On the fairness of single round-robin tournaments: Evidence from European football

Single round-robin tournaments are widely regarded as unfair because of their home-away imbalance: some competitors play more home games than the others. Therefore, most sports leagues are played in a double round-robin format. However, some European football leagues contain a single round-robin (sub)tournament. To test whether empirical data can validate their unfairness, we construct a unique database and develop a strength of schedule metric based on Elo ratings with favourable theoretical properties. In contrast to the robust impact of relative team strength, the strength of schedule is found to have no significant effect on the number of points scored under the assumption of a league-specific home advantage. Consequently, the theoretical unfairness is not a serious issue in practice, or the standard approach of modelling home advantage in Elo ratings might need to be reconsidered.

econ.EM↗

arXiv2026

GeoPano: Towards Geometrically Accurate 3D Reconstruction from a Single Panorama

Panoramic 3D reconstruction enables the holistic recovery of complete 360 $^\circ$ scene geometry, offering significant benefits for robotics and AR/VR applications. However, directly estimating geometry from ERP images is inherently challenged by spherical distortions. To reduce spherical projection distortions, a common strategy is to decompose panoramas into distortion-free perspective views, but independently processing these views often introduces cross-view geometric inconsistencies. To address these challenges, we introduce GeoPano, a feed-forward framework for single-view panoramic 3D reconstruction that recovers a geometrically accurate and consistent point cloud from a single panorama. Our approach is driven by two key innovations: First, we propose a panoramic-aware attention aggregator, featuring a streamlined attention architecture equipped with a novel Pano-Attention mechanism, which explicitly exploits the intrinsic spherical geometry of the panorama to guide cross-view interactions, thereby suppressing geometric inconsistencies. Furthermore, we formulate a hybrid supervision strategy to jointly enforce fine-grained local details in perspective space and global structural coherence in spherical space. Experiments demonstrate that our method significantly outperforms previous SOTA approaches in reconstruction accuracy, geometric consistency, and visual fidelity.

cs.CV↗

arXiv2026

On the dimension conjecture for global Galois deformation rings in characteristic two

Let $K$ be a number field, and let $S$ be a finite set of primes of $K$ containing all primes above $2$ and all archimedean primes. Let $G_{K,S}$ be the Galois group of the maximal algebraic extension of $K$ unramified outside $S$. We prove two refined lower bounds for the Krull dimension of unrestricted global Galois deformation rings of continuous absolutely irreducible representations of $G_{K,S}$ over finite fields of characteristic $2$. As applications, we construct the first counterexamples to the dimension conjecture in every dimension $n\geq 2$ over every number field with a real prime. Over every totally real number field, these constructions also yield the first examples of such universal deformation rings that fail to be complete intersections. Furthermore, we propose a modified dimension conjecture for $K=\mathbb{Q}$ and $n=2$.

math.NT↗