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 793 records · Page 44Linked to original sources

arXiv2026

The Neptunian Final Parsec Problem: How Inertial Waves Enable Neptunes and Gas Giants to Survive Tidal Migration

Observations indicate that some hot Neptunes and Jupiters underwent high eccentricity migration (HEM). A longstanding issue facing HEM theory remains, however: in order to shrink their orbits by ${\sim}1{-}2$ orders of magnitude, planets must dissipate ${\gtrsim}10{\times}$ more orbital energy than is contained in their structural binding energy. Previous work suggests that excitation of planets' fundamental modes (''$f$-modes'') at very high eccentricities, $e{\sim}$0.99, can shrink semi-major axes by ${\sim}$an order of magnitude (${\sim}5{\rightarrow}0.5$ au) while safely bleeding tidal energy to space. We examine how planets traverse the remaining order of magnitude in orbital shrinkage, $a{\sim}$0.5${\rightarrow}$0.05 au, which still poses a danger for Neptunian planets due to their lower binding energies (their ''final parsec'' problem). To assess whether tidally-migrating Neptunes balance their energy budgets, we perform coupled orbit/structure calculations tracking their frequency-dependent tidal response under inertial waves (IWs). We show that tidal radius inflation at moderate eccentricities sweeps Neptunes' fundamental modes back into resonance with tidal forcing; only a small amount of tidal heat relative to a Neptune's binding energy is needed to drive this $f$-mode excitation. This excitation of Neptunes' $f$-modes may offer a route to safe circularization, provided the $f$-mode dissipates without harming the planet. In contrast, eccentric gas giants ($e{\sim}0.9$) circularize without drama over ${\sim}$Gyr. Our theory predicts that warm Neptunes on a HEM track ($a{\gtrsim}0.1 \ \mathrm{au}, \ e{\gtrsim}0.8$) should be rare, in opposition to warm gas giants and lower eccentricity Neptunes. Observational confirmation of our predictions would constrain the interior structures of Neptunian planets via tidal theory.

astro-ph.EP↗

arXiv2026

CNN-based Position Reconstruction for the T0 Detector in the HIRFL-CSR External-target Experiment

The T0 detector provides the event start time for the HIRFL-CSR external-target experiment, and its multi-channel readout signals also contain 2D position information of incident particles. In the CEE experiment, the dedicated beam-profile detector has a rate capability of only ~1 kHz, far below the beam event rate of up to 1 MHz, so it cannot provide real-time beam-spot monitoring under high-rate data-taking or an event time reference. To fill this gap, this work studies a CNN-based position reconstruction method that enables the T0 detector to provide both start time and beam-spot information. The over-threshold times and time-over-threshold (TOT) values from 32 SiPM channels are reorganized into a geometry-preserving input tensor. Time features are corrected by event-wise mean-time subtraction, and TOT features are normalized by an event-wise trimmed mean before feature standardization. A customized CNN directly regresses the incident coordinates. Training and validation samples are built with Geant4 simulation and waveform digitization, and the method is evaluated using a pulsed-laser scanning data set with known position labels. On independent test sets, the residual standard deviations for uniformly distributed planar simulation data are about 1.11 and 1.04 mm in x and y. For the discrete laser-scanning geometry, the resolutions are about 2.36 and 1.34 mm in simulation and 2.42 and 1.41 mm in experiment. The trained model is applied to CEE zero-field beam data collected at the Institute of Modern Physics, Lanzhou, in January 2026, and a 2D beam-spot distribution is reconstructed. These results show the T0 detector's potential as a high-rate beam-spot monitor while retaining its role as an event start-time detector.

physics.ins-det↗

arXiv2026

Minions: A Dynamic and Distributed Virtual GPU Cloud Built on Idle Apple Silicon Devices

The increasing demand for GPU-accelerated machine learning inference has exposed limitations in centralized GPU infrastructures, including high cost, limited elasticity, and long wait times under resource constraints. At the same time, modern Apple Silicon-based macOS devices are equipped with powerful GPUs that remain largely idle during normal usage. This paper presents Minions, a dynamic and distributed virtual GPU cloud that aggregates idle Apple Silicon GPUs across macOS devices to execute large-scale machine learning inference workloads. Minions enables teams to leverage existing hardware without requiring additional GPU procurement. The system provides flexible GPU inference task execution, fault tolerance, and efficient resource utilization. Our experience shows that idle macOS devices can be transformed into a practical distributed GPU inference platform in environments where centralized GPU resources are constrained.

cs.DC↗

arXiv2026

PredictIO: Decoupling Workload Prediction from OS-Controlled Storage Prefetching

Storage stalls remain a significant source of application latency even on fast storage devices. Prefetching can hide part of this latency by fetching data before it is demanded, but general-purpose operating-system mechanisms typically infer future accesses from recent file offsets. This approach is ineffective when accesses are determined by workload structure, such as recurring sparse accesses during application launch, logical traversal in database scans, or model-layout metadata in LLM runtimes. Existing workload-aware prefetchers usually embed prediction logic inside individual applications and couple execution to application-level I/O paths, making policies difficult to reuse across workloads and systems. We present PredictIO, a programmable interface that separates workload prediction from prefetch execution. PredictIO delivers kernel-observed accesses and application hints to user-space policies, which express predictions as uniform page-range requests. A kernel-space executor validates, bounds, and filters these requests against cached and in-flight pages before issuing asynchronous prefetch I/O. We implement PredictIO for application launch, SQLite range scans, and LLM model loading. Across these workloads, PredictIO reduces latency by 5.3-26.3%, achieves up to 2.99x speedup for SQLite scans, and reduces LLM loading latency by 8.8-47.9%. We further show that a common prediction policy can be reused across SQLite and LMDB through lightweight adapters, that an application-integrated RocksDB readahead policy can be migrated while preserving its workload selectivity, and that the execution backend determines how much benefit migration provides for different access patterns and storage devices.

cs.OS↗

arXiv2026

Adaptive Whole-Body MPC with Optimal Posture Planning for Humanoid Robots under Unknown Payloads

Humanoid robots performing lifting and material-handling tasks must maintain balance despite unknown and varying payloads. However, payload uncertainty significantly affects whole-body dynamics and may destabilize conventional controllers. We propose an adaptive whole-body control architecture that integrates three components: (i) a centroidal model predictive controller (MPC) that generates dynamically feasible reference motions, (ii) an adaptive fuzzy balance controller that compensates for payload-induced model mismatch by generating corrective torques distributed across the supporting-leg joints, and (iii) a waypoint-based optimal posture planner built on a triple-integrator model that produces smooth sit-to-stand and stand-to-sit trajectories satisfying center-of-mass (CoM) stability constraints, joint limits, and whole-body feasibility requirements. Together, these components enable stable lowering, grasping, lifting, carrying, and placing behaviors under significant payload uncertainty. Extensive simulation and hardware experiments demonstrate that the proposed framework improves balance robustness and reduces tracking degradation during posture transitions and lifting tasks compared to non-adaptive baselines.

cs.RO↗

arXiv2026

An Ultraviolet Instrument on the Habitable Worlds Observatory: Considerations and Capability Trades on the Multi-Object, Integral-field, and High-resolution modes of a potential UVI

The Habitable Worlds Observatory (HWO) is NASA's next-generation astrophysics platform, delivering transformative capability in the search for life and study of the origins of the universe. Amongst the instruments under consideration for HWO is a multi-purpose ultraviolet spectrograph and imager that provides a deep toolbox for some of the most pressing science investigations of the next several decades. In this paper, we present a basic draft model of a potential UVI, including a multi-object spectrograph, far ultraviolet imager, high resolution spectrograph, and integral-field spectrograph. The draft designs for these channels of the UV instrument serve as a basis for a general discussion of the instrument architecture and design challenges, while also identifying potential trades, enhancements, and other considerations that the HWO project will need to address as it moves into the formulation phase. We conclude with a presentation of several scientific simulations that have helped guide the capabilities of the draft design. This paper is intended to serve as a summary of UV instrument modes, technology, and potential design ahead of planned instrument studies in the late 2020's, and then Phase A in the early 2030's.

astro-ph.IM↗

arXiv2026

SmartSalesBench: A Resistance-Aware Multi-Turn Benchmark for Evaluating LLM Sales Agents

We present SmartSalesBench, the first multi-turn conversational benchmark explicitly designed to evaluate large language model (LLM) sales agents against psychologically grounded, resistant buyer personas. Existing task-oriented dialogue benchmarks assume cooperative users and measure task completion; SmartSalesBench inverts this assumption, placing agents against personas programmed to resist and challenge across conversations of up to twenty turns. The benchmark includes five agent profiles, twelve buyer personas, fifteen smartphone products across six price tiers, and four controlled experiments with 3 iterations totaling 13,680 sessions. Both sales-agent and buyer persona generations are grounded through retrieval-augmented generation (RAG). We introduce nine evaluation metrics grouped into three tiers: primary outcome metrics (CR, BRS, RIR), secondary quality metrics (SMS, PFS, ETI), and exploratory metrics planned for future reporting (ITS, ARI, SPI). Our evaluation combines automated LLM-as-judge scoring with internal human assessment (n = 50, \k{appa} = 0.71, Pearson r = 0.81 with BRS).

cs.AI↗

arXiv2026

Diffusion Precipitation Generation with Medium-range Weather Forecasting

Accurate daily precipitation forecasts are important for weather-sensitive decisions. Although recent Weather Foundation Models (WFMs) have achieved significant progress, they are trained on reanalysis data, which exhibit regional and intensity-dependent biases. Since retraining WFMs on observation-based targets is costly, we study Dynamics-Precipitation Decoupling, a plug-and-play framework that uses a separately trained generator to map non-precipitation forecasts from a frozen weather model to observation-aligned daily precipitation. Its key challenges are to capture heterogeneous rainfall distributions and align daily accumulations with atmospheric signals that evolve across space and time. To address them, we train a diffusion generator conditional on atmospheric states and satellite-gauge estimates of precipitation. A multi-scale recurrent encoder summarizes atmospheric evolution, while cross-attention aligns atmospheric and precipitation features during denoising. Extensive evaluation of various backbones, including Pangu-Weather, GraphCast, and HRES, on global and regional forecasting shows improved daily precipitation accuracy and event detection, with better representation of localized accumulation patterns. Global ensemble evaluation shows lower Brier scores than ENS and training-from-scratch GenCast at all reported lead times, with lower CRPS in most comparisons. Remarkably, a heavy-rainfall case study using 2,288 stations in China further supports improved daily accumulation patterns and event detection against ground observations.

cs.LG↗

arXiv2026

Beyond Semantic IDs: Finite-Beam Reachability and Continuous-Query Generation in Generative Recommendation

Generative recommendation (GR) of social media platforms commonly retrieves items by autoregressively decoding discrete Semantic IDs (SIDs) with finite-width beam search. This work examines how this output interface constrains whether held-out targets enter finite returned lists, while distinguishing catalog-wide scoreability from empirical top-K support. We introduce two continuous-query paradigms: CQG-Single generates one retrieval query in a single pass, and CQG-AR autoregressively generates continuous residual updates, retaining step-wise decoding while removing the SID item-token vocabulary, code lookup, and discrete beam expansion. These designs thereby eliminate the SID-specific reachability constraint induced by finite-beam decoding and avoid the technical challenge of aligning a discrete SID vocabulary with the pretrained backbone's vector space. Both methods make every encoded item directly scoreable through dense or approximate-nearest-neighbor retrieval. Under a unified protocol with valid-prefix constrained decoding, CQG-Single exceeds the TIGER-style baseline in all seven runs while using 55.3% as many trainable parameters. CQG-AR remains comparable on ML-1M and is numerically stronger on Beauty and Sports datasets. A controlled depth ablation further shows that additional continuous refinement does not monotonically improve shallow accuracy. These findings motivate decoupling output representation, autoregressive modeling, and retrieval budget when evaluating generative recommendation systems.

cs.AI↗

arXiv2026

Learning Representations to Guide Pixel Diffusion Transformers

Visual representations are largely developed for understanding, yet what makes them useful for generation remains unclear. We study this question in pixel diffusion transformers and uncover two key insights. First, better representation quality does not necessarily translate into greater generative utility, as externally aligned semantics can be progressively lost during pixel synthesis. Second, architectural separation with specialized pixel decoders mainly redistributes semantic and spatial modeling, without necessarily yielding representations better suited to generation. These insights suggest that a useful generative representation must be both properly formed and actively exploited. Motivated by these insights, we introduce Representation Guidance (RG), which directly forms and exploits internal representations while retaining a plain Transformer architecture. A low-dimensional bottleneck forces generation-relevant information into a compact representation, which is then explicitly used to guide decoder generation by modifying the flow-matching target according to its induced prediction offset. RG learns clearer semantic and spatial structure while scaling efficiently in both parameters and computation. On ImageNet 256x256, RG-L/16 with only 459M parameters outperforms substantially larger pixel diffusion transformers, while RG-XL/16 achieves an FID of 1.55. RG further reaches an FID of 1.74 at 512x512 and extends to text-to-image generation, achieving 0.89 on GenEval and 87.1 on DPG-Bench.

cs.CV↗

arXiv2026

Hecke Eigenform Product Identities for Full-Level Hilbert Modular Forms

We study full-level Hilbert Hecke eigenform product identities over totally real fields of degree at least two, allowing arbitrary compatible narrow class characters. Combining our results with previous work, we classify products of two Eisenstein eigenforms of positive parallel integral weights and products of an Eisenstein eigenform of weight at least two with a cuspidal eigenform of positive parallel integral weight. The unique identity in each of these two classes is given by $E_{4}=60 E_{2}^{2}$ and $h_{8}=120 E_{2} \cdot h_{6}$, respectively, both over $\mathbb{Q}(\sqrt{5})$. The proofs use constant-term comparisons, indecomposable Fourier exponents, integrality of reciprocal Eisenstein constant terms, and a Rankin-Selberg formula for Petersson inner products. In the remaining case of a weight-one Eisenstein factor and a cuspidal factor, we also establish $h_5=12E_1(1,\varepsilon)\cdot h_4$ over $\mathbb Q(\sqrt3)$, where $\varepsilon$ is the nontrivial narrow class character. All eigenforms in these identities have unit-ideal coefficient one.

math.NT↗

arXiv2026

Quantum Information-Theoretic Analysis of Optical Fiber Sensing via Stimulated Brillouin Scattering

Stimulated Brillouin scattering (SBS) is a powerful mechanism for distributed sensing in optical fibers, as the Brillouin response is highly sensitive to local material parameters. Assuming the input probe is a quantum state, we develop a quantum-information-theoretic framework for SBS-based fiber sensing and intrusion detection. Starting from a linearized opto-acoustic Hamiltonian, we derive an effective input-output relation for a short fiber segment. We then model each fiber segment as a single-mode bosonic Gaussian channel and the entire fiber as a cascade of quantum channels. We then compute the quantum Fisher information for general SBS-dependent parameters and identify optimal measurement settings to achieve the quantum limit, comparing them with receiver-constrained strategies such as heterodyne detection and photon counting. In addition, we propose an adaptive approach that achieves first-order asymptotic optimality and provides a novel second-order correction to address preliminary miscalibration. As an application, we apply the framework to the fiber-tapping detection problem.

quant-ph↗

arXiv2026

The Evidence Anchoring Boundary: Where Visual Interventions Matter and When They Help in Multimodal LLMs

Visual interventions provide a training-free way to reduce hallucinations in multimodal large language models. However, an intervention may have little leverage at some layers and may damage a correct answer even when it changes the output. We investigate where visual interventions retain leverage and when they improve predictions. First, we apply the same visual-residual edit at successive decoder layers, matching its RMS strength across depth. Across six checkpoints from three model families, we observe a persistent drop in intervention effect and define its operational onset for this edit as the \textbf{Evidence Anchoring Boundary (EAB)}. With boundaries fixed on calibration queries, held-out effects on disjoint evaluation queries before EAB are $13.78\times$--$26.10\times$ larger than those after it. Comparison with intermediate readouts shows that the edit can lose leverage before answer preferences stabilize. Second, we compare counterfactual subtraction across presence, attribute, and relation questions. Subtraction can correct a misleading preference, leave a decision unchanged, or remove a preference supporting the correct answer. Based on these observations, we develop calibrated selective routing, which preserves the original prediction unless calibration supports an intervention for the question type. It improves full-set attribute accuracy by $1.56$--$1.81$ percentage points across three checkpoints, while universal subtraction is neutral or harmful.

cs.CV↗

arXiv2026

Silent breakdown of the perturbative polarization expansion in the terahertz regime

Nonlinear optics expands the polarization in powers of the drive and retains only the lowest-order terms. This truncation is safe at optical frequencies, where the convergence field lies far above the damage threshold. In the terahertz regime, phonon polaritons enhance the nonlinear susceptibilities by orders of magnitude, lowering the convergence field into the sub-damage range and opening a window in which the expansion fails while the crystal remains recoverable. We term this window silent breakdown. At resonance, the nearest singularities of the driven lattice mode are a complex-conjugate pair of branch points below the multistability threshold, so the response remains smooth and single-valued even though the $χ^{\left( n \right)}$ description fails. The convergence field scales as $Q^{-3/2}$, so sharpening a resonance enhances the nonlinearity but narrows the range over which the expansion applies. It follows from the mode shift $S$ alone, $E_{\mathrm{conv}}=E_{\mathrm{ref}}(3\sqrt{3}QS)^{-1/2}$, giving 2$\sim$9 MV/cm for $\mathrm{LiNbO}_{3}$. Since resonant enhancement is central to terahertz nonlinear optics, this boundary defines the field range over which susceptibility-based descriptions are quantitatively reliable.

physics.optics↗

arXiv2026

Fourier--Laplace Injectivity for Oblique and Polygonal Cylinders in Passive Inverse Problems

We study Fourier--Laplace transforms on null phases arising in passive inverse problems. Densities with codimension-one dependence in oblique cylinders are first considered. Integration in the distinguished direction reduces the data to an entire transform on a real-analytic maximally totally real submanifold. Holomorphic uniqueness then gives injectivity. A fixed transverse polytope swept along a polygonal curve, forming a polygonal cylinder, is next considered. The density is assumed to be polynomial in the material transverse variables on each swept piece. A weighted Brion--Barvinok expansion separates the vertex terms and the successive pieces. The same representation gives a finite-dimensional Lipschitz estimate when the polynomial degree is fixed. We apply these transform results to passive wave and heat equations in arbitrary dimension. The initial state is zero outside the observation domain. Equality of the Laplace-domain boundary data determines the quantity of the initial data divided by the speed squared in the acoustic model and the quantity of the initial data divided by the coefficient of thermal conductivity in the parabolic model. The question of what remains true when only weighted generalized boundary data on a finite interval $[0,T]$ are available is then studied. These data determine the harmonic moments of the quotient up to the weighted terminal residual. This identity gives a Hölder estimate in an oblique cylinder and a Lipschitz estimate in a polygonal cylinder. Finally, a continuous generating curve is approximated by a polygonal curve and a continuous material density by piecewise transverse polynomials. The resulting reconstruction estimate separates measurement noise, density approximation error, geometric error, and the condition number of the finite sampling matrix.

math.AP↗

arXiv2026

Work2Skill: Capturing Human GUI Experience for Reusable Agent Skill Induction

Computer-use agents can perform increasingly complex GUI tasks, yet unfamiliar software and workflows requiring human domain knowledge remain challenging. Human demonstrations convey procedural knowledge through the same GUI interaction space, but reuse requires interpreting actions in context and separating workflow rules from task-specific details. We thus introduce Work2Skill, a framework that captures human GUI experience for reusable agent skill induction. WorkRecorder provides the framework's infrastructure for collecting this experience, aligning screen recordings, mouse and keyboard events, and available accessibility-tree information. Using WorkRecorder, we construct WorkDemo-350, the first synchronized multimodal dataset of human desktop GUI demonstrations for reusable agent skill induction and downstream evaluation. Its 350 demonstrations are screened for completeness and task success. We then propose an agentic skill-induction method that inspects evidence on demand to produce reusable skills capturing workflow rules, step dependencies, and completion checks. When the same four execution models are equipped with skills induced by each method under shared execution settings, Work2Skill achieves the highest scores and success rates. Mean score improvements over the corresponding No Skill baseline are 4.67--10.54 points on MyPCBench Bounded Action and 3.11--7.50 points on a 286-task subset of OSWorld-Verified. Our code and data are available at https://github.com/china-qijizhifeng/Work2Skill.

cs.AI↗

arXiv2026

NucleoNet: A Nuclear-Aware Determinant Ansatz for Multi-Geometry Neural Variational Monte Carlo

Neural-network variational Monte Carlo provides an accurate route to solving the electronic Schrödinger equation, but many neural wavefunctions are optimized for a single nuclear geometry, requiring separate calculations to map out a potential energy surface. We introduce NucleoNet, a nuclear-aware determinant ansatz in which electrons and nuclei form a unified particle set propagated through a shared electron--nucleus backbone that jointly evolves one-particle and two-particle representations. Nuclei carry learned representations that evolve alongside the electronic representations and condition orbital construction, while antisymmetry is imposed only by electronic determinants. On equilibrium small-molecule benchmarks, NucleoNet achieves accuracy comparable to established neural-wavefunction baselines with substantially fewer parameters. For each molecular system, a single model optimized jointly across geometries describes the N$_2$ dissociation curve, Li displacement in LiCN, and the ethane and formamide minimum-energy paths for bond dissociation. Sparse-geometry training generalizes to unseen geometries for N$_2$ and LiCN without further optimization, and N$_2$ ablations show that nuclear streams substantially improve this interpolation. These results show that a compact wavefunction with jointly evolved nuclear representations can describe the geometries of a molecular system with a single set of parameters.

physics.chem-ph↗

arXiv2026

Scan-Parallel Variational Inference for event-driven Latent Jump Models

Continuous-time latent state-space models are widely used for irregularly observed time series, typically representing uncertainty through a random initial state followed by smooth ordinary differential equation (ODE) dynamics, or Brownian perturbations accumulated by a stochastic differential equation (SDE). While these constructions are natural for smooth or diffusive systems, they do not directly encode a finite set of unobserved shocks whose timing, magnitude, and effect on memory are themselves uncertain. We instead introduce latent jump models (LJMs): the latent state follows closed-form deterministic dynamics between Poisson-driven event times and undergoes a learned instantaneous jump at each event, with state-dependent intensity and mark distributions. This defines a time-homogeneous Markov process whose jumps are intrinsic state transitions rather than numerical integration steps. For training, we derive a tractable event-augmented path-space evidence lower bound (ELBO) and introduce Scan-Parallel Variational Inference (SPVI): an observation-conditioned variational posterior samples the complete latent event configuration before the state recurrence is evaluated using triangular affine scans. This enables fast and scalable training for LJMs in logarithmic temporal depth. Across synthetic stochastic systems and real irregular time series, LJMs remain competitive with continuous-time latent baselines in forecast accuracy, often with fewer latent updates than numerical solver evaluations, while training one to two orders of magnitude faster than the sequential baselines.

stat.ML↗