Search arXiv⌕ Search

SEARCH · Search arXiv

Search Search arXiv

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

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,603 records · Page 89Linked to original sources
arXiv2026

Matching Tasks to Agents: Mechanism Design for Multi-Agent Contracts

We study task assignment under adverse selection and moral hazard. In the single-task setting, agents privately know their types, which determine their success probabilities and effort costs. Based on the agents' reported types, the principal selects an agent to perform the task and offers a success-contingent payment. Although optimal Bayesian contract design is generally computationally intractable and Myerson's characterization does not directly extend to settings with hidden actions and multidimensional private information, we show that the single-task problem admits a simple Myerson-style characterization. Our key insight is that incentive constraints depend only on the ratio between effort cost and success probability. We characterize an optimal mechanism and show that it can be computed in polynomial time. We also establish a tight $Θ(1/(NK))$ worst-case bound on the ratio between the optimal profit and the full-information benchmark, where $N$ is the number of agents and $K$ is the number of types per agent. We then extend our analysis to multiple tasks, showing that an optimal mechanism may not exist. Nevertheless, for a constant number of agents, we provide a polynomial-time approximation scheme for the supremum profit.

cs.GT↗
arXiv2026

Agri-SCHOLAR: Simulation-Grounded In-Season Rolling Advisories for Agricultural Crops Using LLM-Bounded Optimization

Existing agricultural advisories to farmers suffer from several limitations: (a) remain largely expert-driven, (b) do not adequately account for in-season weather, and (c) do not scale to individual farms. In this paper, our objective is to close these gaps by developing Agri-SCHOLAR (Simulation-Grounded Crop Harvest Optimization with LLM-Guided In-Season Advisory Rollouts). Agri-SCHOLAR is a rolling, in-season advisory framework that issues, at intervals of every few days or every week, a personalized advisory using the following information: soil and crop properties, observed weather to date, short-range forecast of weather, and a season extrapolation of weather based on historical analogs. The novelty and depth of Agri-SCHOLAR lies in each advisory being produced by a closed-loop system which couples multiple LLM agents with a powerful crop simulator (such as APSIM in our case). The simulator crucially serves as an evaluation oracle. In Agri-SCHOLAR, the LLM agents propose agronomically sound and feasible bounds on actionable parameters and an optimizer searches within those bounds against a multi-objective reward spanning yield, resource utilization, and economic feasibility. We conduct a detailed 11-year analysis of the maize crop in two selected locations of South Asia and show that Agri-SCHOLAR consistently outperforms the region's standard packages of practice and satisfactorily addresses the limitations of existing agricultural advisories.

cs.AI↗
arXiv2026

Mitigating Saliency Collapse: Robust Saliency-Aware Long-Text Image-Text Alignment

Long-text image-text alignment is essential for fine-grained scene understanding, yet existing methods mainly focus on extending context length or improving local alignment. We identify an overlooked failure mode, saliency collapse, where models over-rely on text that features visual prominence while underutilizing non-salient context. Consequently, retrieval performance degrades sharply when salient content is reduced, despite that the non-salient context remains informative. To address this, we propose RoSA, a saliency-aware finetuning framework that decomposes image-text pairs into complementary salient and contextual views, enabling joint learning of prominent and contextual semantics. We propose a modality-independent decomposition strategy to combine object-level visual saliency estimation with LLM-based text decomposition to bridge vision and language. We employ auxiliary objectives to explicitly align these views, ensuring robust representation with both salient and contextual features. RoSA greatly improves long-text retrieval and demonstrates strong robustness to saliency collapse across benchmarks. Moreover, our auxiliary objectives are plug-and-play, bringing consistent gains to existing tuning methods. Code and checkpoints are available at: https://github.com/intelligolabs/RoSA

cs.AI↗
arXiv2026

Searching for substellar companion candidates with Gaia. V. M dwarfs within 15 pc studied with absolute astrometry, radial velocities, and direct imaging

In this work, we survey M-type stars within 15 pc observed by Gaia based on their colour, magnitude, and quality of astrometric data. We aim to detect new substellar companions using primarily Gaia astrometry and, when available, proper-motion anomaly (PMa) along with radial velocity (RV) and direct imaging (DI) data. We used the GaiaPMEX tool to search for bound companions around 582 M dwarfs. We identified all previously known stellar and planetary companions that were consistent with the Gaia excess signal. Then for all newly identified companions, we computed detection limits in the (mass, semi-major axis) parameter space using archival RV and DI data, to possibly rule out the presence of a stellar companion. The joint analysis of the parameter space allowed by astrometry and constrained by RV and DI provides tight constraints on the nature of the companion. We found 156 M dwarfs to have a bound companion within the Gaia sensitivity domain. The astrometric signals of 37 and seven of them are due to known stellar and substellar companions, respectively. We report 35 new substellar companions among which at least ten giant planets orbiting between 0.2 au and 2 au, with masses close to 1 M_Jup. We also report 77 substellar candidates, and provide detection limits using astrometry+RV+DI for 50 stars with known exoplanets and no significant Gaia excess noise. This work reveals a significant population of new substellar companions to M-type stars, challenging core accretion models. These results will be crucial inputs for estimating occurrence rates and studying the architecture of planetary systems around M dwarfs. Some new companions presented here, located beyond the water ice line, are prime targets for DI or coronagraphy observations with JWST or ELT.

astro-ph.EP↗
arXiv2026

Noise redistribution by stabilizer codes: an organizing view of concatenated three-qubit codes

We present an organizing framework for concatenated quantum error correction based on the observation that the syndrome projection of a stabilizer code reshapes the noise channel in a code-determined manner. For any stabilizer code, one level of QEC maps the physical noise to a block-diagonal logical channel whose diagonal Pauli weights are fixed by the code geometry. Coherences between different syndrome classes vanish, and no physical Pauli twirl is required to reach this block-diagonal form. Concatenation can then be viewed as the iteration of the same redistribution step: the dominant logical error after one level selects the code used at the next. We instantiate the framework with the two smallest stabilizer codes, the bit-flip and phase-flip codes, and apply it to amplitude damping. The redistribution rules give the concatenation protocol $C_2(\hatσ_y) C_1(\hatσ_x)$, whose ideal-recovery fidelity scales as $f_l=1-\mathcal O(γ^2)$. Solving the crossover condition $F_l>F_{l-1}$ in the ideal-gate limit yields a critical damping parameter $γ_c\approx0.30$; in terms of the initial channel fidelity, $f=1-1.72755γ^2$. Under the same ideal-recovery assumption, we compare the overall gate accuracy rate with that of the five-qubit code. The scope of the framework, in particular for coherent and correlated noise, is stated explicitly.

quant-ph↗
arXiv2026

Beyond Global Importance: A Heterogeneity Index for Sensitivity Analysis

Global sensitivity analysis quantifies how strongly inputs influence a model output, however, input importance often varies across the model space. We introduce a heterogeneity index H to quantify this variation. The index compares normalized regional sensitivity profiles using mean pairwise total-variation distance and its values range from zero to one. We test H on nine benchmark functions and examine its robustness to the number of regions, sample size, and sampling strategy. We further demonstrate it using models of wooden-pallet carbon footprint, flood risk, and steel-structure reliability. Across these cases, H reveals changes in sensitivity that global indices obscure, including inputs whose importance changes substantially across the model space and differences between rare behavioral regimes. The index complements global sensitivity analysis by indicating when a single global sensitivity profile is insufficient. An open-source implementation is provided in the simdec package.

stat.ME↗
arXiv2026

Revisiting Distribution Matching for State-to-Vision Policy Distillation in Laparoscopic Robotic Manipulation

End-to-end control policies trained via Reinforcement Learning (RL) have emerged as an alternative to conventional surgical robotic control approaches, with recent studies exploring state-based policies trained on privileged state observations. However, such approaches rely on precise physical states that are difficult to obtain reliably in surgical environments. Vision-based RL provides an alternative by learning manipulation policies directly from RGB observations, avoiding explicit state estimation and manually engineered perception pipelines while enabling policies to acquire task-relevant visual representations. Nevertheless, obtaining vision-based policies via pure RL methods for surgical manipulation remains challenging. Thus, imitation learning (IL) becomes another solution, yet it requires a large number of expert demonstrations, which are difficult to obtain in minimally invasive surgery. In this work, we propose a framework for state-to-vision policy distillation in laparoscopic robotic manipulation. Specifically, a state-based policy with privileged observations is employed as the teacher, and a vision-based policy learns from RGB observations and robot proprioceptive information, which requires less expert demonstrations than IL and introduces superior performance than RL methods. We conduct extensive experiments on various surgical manipulation tasks in the simulated environment. Experimental results show that, in the context of laparoscopic robotic manipulation, training vision-based policies with general-purpose RL methods is challenging, whereas our proposed state-to-vision policy distillation leads to more consistent and stable performance across tasks.

cs.RO↗
arXiv2026

Evaluating Artificial Wisdom (AW): A Trajectory-Based Framework for Context-Sensitive Decision Regulation

Capability benchmarks measure what large language models can do, but not whether their recommendations remain appropriate as decision conditions change. This paper proposes a trajectory-based measurement framework and prospective validation protocol for artificial wisdom, defined behaviorally as context-sensitive regulation of judgment. Versioned reference specifications determine admissible responses and transition materiality before evaluation. Seven stage-level criteria assess judgment under current conditions, while two transition components assess material-update adequacy and commitment-preserving stability. Stability requires prior adequacy; switching within registered commitment classes is permitted, and persistent error earns no stability credit. A nine-component profile is primary, with a secondary bottleneck score and separate critical-failure reporting. Three simplified simulations examine precision, bottleneck estimation, and sensitivity to model-family count. A single-criterion simulation met its precision criteria at 24 scenarios per domain. Near a bottleneck tie, the estimated minimum had a downward bias of 0.55 points at that size. A simplified incremental-prediction simulation detected a Brier improvement of 0.027 with probabilities of .508 and .742 using eight and twelve families, respectively; it does not establish power for the operational protocol. An exploratory predecessor-rubric study comprising 71 evaluation records illustrates aggregation sensitivity but does not validate the proposed instrument. The twelve-family target remains provisional pending feasibility assessment and calibration of the complete analysis. Whether the profile predicts independent decision outcomes beyond capability remains an open empirical question.

cs.AI↗
arXiv2026

Multi-Tool Selection for Robotic Systems: Comparing Semantic Retrieval and LLM

Natural-language interfaces allow users to command robotic systems without interacting directly with mission planners or robotics middleware. Their reliable use depends on robot tool routing, defined here as the selection of the smallest complete subset of predefined executable capabilities required to satisfy a request. This task is particularly difficult for compound instructions, where incomplete selection may omit necessary sensing or monitoring capabilities, while excessive selection may expose irrelevant tools. We compare four routing strategies: three progressively refined ChromaDB-based methods - fixed-threshold retrieval, relative-distance retrieval, and clause-aware aggregated retrieval - and a cloud-hosted 120-billion-parameter language-model baseline. The methods are evaluated on 164 manually annotated prompts organised into three complexity levels, from direct single-tool requests to compound requests requiring at least three tools. Using exact-match accuracy, completeness, precision, a completeness-weighted composite score, and routing latency, CDB-v3 achieves the best macro-average score among the retrieval-only methods, with 0.7822, compared with 0.8869 for the large-model baseline. Although the LLM provides higher routing quality, CDB-v3 exhibits substantially lower latency and can be deployed locally without requiring cloud connectivity. These results identify a quality-latency trade-off between direct large-model selection and clause-aware semantic retrieval for privacy-sensitive, connectivity-constrained, and resource-constrained robotic systems.

cs.RO↗
arXiv2026

CARL: Chat Assisted Reinforcement Learning

Failure of applied Reinforcement Learning (RL) to solve a given problem predominantly stems from unaddressed specification gaps, where the intent of human experts in the problem's domain is inaccurately or incompletely translated by AI into mathematical objectives, constraints, and reward functions. Failure can also stem from unaddressed performance gaps, where human domain experts are unable to determine the validity of the RL based solution or where parameters (i.e., hyperparameters or weights on the reward terms comprising the reward) are poorly chosen. Chat Assisted Reinforcement Learning (CARL) bridges these gaps through a human-AI co-design interface architecture that enforces interactive specification prior to code generation and then enables subsequent verification and parameter alterations via iteration. More specifically, CARL leads domain experts through a structured ten-step workflow for each iteration using step-specific skill files, domain-aware suggestions via an assistant persona (called Trillian), and compact decision summaries that maintain context across iterations without context overload. Once the domain expert confirms the Markov decision process (MDP) specified by this procedure, CARL's autonomous tool-use loop generates, tests, executes, and debugs an end-to-end RL training pipeline, followed by an interactive post-training diagnosis and revisit step that allows for verification and parameter alteration before running a new iteration if desired.

cs.HC↗
arXiv2026

XALIGN: A Human-centric Visual Analytics Approach to Understanding AI Explanations

Explainable AI (XAI) methods are usually evaluated for technical properties such as faithfulness and robustness. In high-stakes application domains, experts also need to check model predictions against established knowledge. Different explanation methods often disagree, and this disagreement is hard to see when each method is inspected separately. We propose XALIGN, a visual analytics workflow and interface that shows rule-based explanations and several feature attribution methods (TreeSHAP, KernelSHAP, and LIME) side by side for the same Random Forest model. Agreements and contradictions between the methods become visible signals that users can explore and compare with domain knowledge. We apply XALIGN in two case studies: the German Credit dataset and a meteorological dataset, in which events derived from ICON model output and Cloudnet observations are classified. In both, the views reveal discrepancies between the rule-based explanation and the attribution methods that a single method would not show. In an expert evaluation with three specialists, participants found XALIGN useful for identifying influential features and for avoiding over-trust in predictions. Linking rules to attributions was harder for the participant without an XAI background.

cs.LG↗
arXiv2026

A general quantified Wiener-Ikehara theorem

Building on the technology developed in \cite{Deb-qik}, we establish a quantified Tauberian theorem under a flexible Wiener-Ikehara Tauberian condition and under several types of boundary behavior for the Laplace transform.

math.CA↗
arXiv2026

Mostly Reading: What Open-Book Scores Measure in Document-Level LoRA Adapters

Single-document LoRA adapters have been proposed as units of knowledge to be merged and retrieved; yet open-book scoring, with the document in the prompt, can credit a document's own adapter (its specialist) for reading that document as readily as for holding it. Training one adapter for each of 150 Wikipedia articles on Llama-3.1-8B-Instruct, we find that open-book scores mostly measure a reading behaviour every adapter shares: a random adapter trained on another document recovers 82 % of the specialist's open-book advantage over the base model but only 2 % of its advantage on a cloze probe that asks for the same document's facts without providing the source text, a share we cannot distinguish from zero. Content does reach the weights, since the specialists beat the base model in cloze recall by 0.094, and the geometry of their updates follows the documents they read; yet, in a test of 30 queries, adapters chosen by weight-space proximity or text embedding are not distinguishable from random ones on either probe, although that test cannot exclude an advantage of about four fifths of what a query's own adapter gains over a random one. At 16 documents or more, correctly scaled merges recover 87 % to 89 % of the specialists' open-book advantage but only 12 % to 16 % of their cloze advantage, with intervals that run from below zero to at most 32 %, against 94 % for joint training.We also report four silent errors, each of which produced a clean, wrong result.

cs.CL↗
arXiv2026

Exact Archimedean Flicker--Rallis Periods on Cohomological Minimal $K$-Types

For the symmetric pair $(\mathrm{GL}_n(\mathbb C),\mathrm{GL}_n(\mathbb R))$, we compute the exact Flicker--Rallis period of a distinguished vector in a cohomological minimal $U(n)$-type, normalized by the Gel'fand--Tsetlin pairing of Hara--Miyazaki--Namikawa. Our main tool is a weight-independent identity between the Whittaker period and a compact $O(n)$-average. We determine the proportionality constant using the archimedean newform formula of Humphries--Jo, and evaluate the compact functional on the normalized distinguished vector. This yields an explicit formula for the Flicker--Rallis period for every cohomological weight.

math.RT↗
arXiv2026

Chunk Recomputation under Real Document Revisions: Measuring Update Cost and Retrieval Quality of Fixed, Recursive, and Content-Defined Chunking

Chunking strategies for retrieval-augmented generation are almost always evaluated on a static snapshot, yet the documents underlying a knowledge base continue to change. Because chunks are embedded independently, a boundary that moves after an edit forces unchanged text to be re-embedded. We measure this update cost on 120 pairs of adjacent Chinese and English Wikipedia revisions for three stateless chunkers (non-overlapping fixed-size, recursive paragraph-first, and content-defined chunking) under a strict byte-identical reuse rule and a frozen dense encoder. On the 116-article common-success pool, with frozen, length-calibrated parameters, fixed-size chunking recomputes 57-68% of its full new-version embedding-input workload, recursive chunking 38-56%, and content-defined chunking 43-54% (it failed on 4 of 62 Chinese articles), while at these parameters mean evidence recall under a fixed 2048-token retrieval budget differs by less than two percentage points within either language. A cost-only parameter grid shows that the ranking of the two content-aware chunkers changes with chunk-length settings, whereas non-overlapping fixed-size chunking has the highest recomputation cost in every length-matched condition. Controlled edits and a verbatim-presence diagnostic support boundary drift as one contributor to recomputation of text already present in the previous version. Update cost is therefore a substantial, parameter-sensitive dimension that static retrieval evaluations do not expose and that should be reported alongside retrieval quality.

cs.CL↗
arXiv2026

Demonstration-Based Robotic Feeding Using Dynamic Movement Primitives

This paper presents a demonstration-based robotic feeding framework that integrates human demonstration, Dynamic Movement Primitive-based motion planning, vision-based scene perception, and robot control. A human operator first demonstrates the scooping and feeding motions while the position and Euler-angle orientation of the spoon are recorded using a motion-capture system. The recorded trajectory is segmented into scooping and feeding motions, which are encoded using separate Dynamic Movement Primitives. Before task execution, an RGB-D camera estimates the participant's mouth position, while two AprilTags are used for robot-frame registration and tray localization. The detected tray pose is used to position the scooping motion, whereas the detected mouth position defines the positional goal of the feeding motion. The approach, scooping, and feeding trajectories are then executed sequentially by a Kinova Gen3 manipulator through a ROS~2-based framework. A real-world experiment demonstrates the feasibility of the integrated pipeline and shows that the learned spoon motions can be reproduced within the detected experimental scene.

cs.RO↗
arXiv2026

On Spurious Mesh-Scale Oscillations Associated with Resonance-Cone Regimes in Wave-Plasma Simulations

Spurious mesh-scale oscillations can compromise frequency-domain finite-element simulations of electromagnetic wave propagation in magnetized cold plasmas. We investigate their origin and mitigation through a verification framework combining a mixed Nédélec-Lagrange solver, a waveport accommodating non-orthogonal modes in anisotropic media, and an analytical generalized Fresnel benchmark at a vacuum-plasma interface. Comparisons with this benchmark show that the observed oscillations occur in resonance-cone regimes and are absent from the analytical fields, confirming their numerical origin. Systematic tests assess the effects of approximation order, variational formulation, mesh geometry, and numerical collisionality. Plane-wave residual and Fourier spectral analyses suggest that local mesh variations induce spectral coupling to additional wavevector components, predominantly populating the dispersion branches in collisionless resonance-cone regimes where the reduced scalar equation loses ellipticity. Uniform rectangular grids recover the analytical wave structure, further supporting the role of mesh geometry and symmetry. However, strict symmetry requirements limit the applicability of this approach, motivating numerical regularization. In resonance-cone regimes, introducing sufficient numerical collisionality suppresses the mesh-scale oscillations, yielding smooth electric-field solutions and reducing the error relative to the analytical benchmark. For one resonance-cone benchmark configuration, the error-minimizing damping decreases with mesh refinement according to an empirical $Z_{\mathrm n}\propto h^{1.48}$ scaling, sustaining error reduction consistent with second-order convergence and avoiding the accuracy limit imposed by fixed collisionality. ...

physics.plasm-ph↗
arXiv2026

Effective-action formulation of $\mathbf{k}$-odd bianisotropy and unidirectional polarization rotation in nonlocal temporal metamaterials

We derive the effective action of temporal metamaterials up to first order in spatial derivatives through a multiscale analysis of the electromagnetic Lagrangian density. We show that the first-order spatial correction gives rise to an effective bianisotropic coupling tensor that is odd in the wavevector $\mathbf{k}$. Remarkably, chiral, pseudochiral, and omega-type bianisotropic responses emerge as distinct directional manifestations of a single nonlocal coupling tensor, even when the underlying electric and magnetic response tensors remain symmetric. Furthermore, we establish a framework for analyzing Lorentz reciprocity based on the Hessian of the effective action. We demonstrate that Lorentz reciprocity is preserved in the presence of omega-type coupling but is broken when chiral or pseudochiral coupling emerges. This nonreciprocity manifests as the unidirectional rotations of normalized Stokes vectors incurred by Floquet-engineered chirality and pseudochirality.

physics.optics↗