Searcharxiv⌕ Search

arXiv subjects

Zhenyu Wu

Publications and source records attributed to Zhenyu Wu.

At least 19 recordsLinked to original sources
arXiv2026

Gaia DR3 Compact Binary Candidates Constrained by Radial Velocity Measurements

Astrometry enables the most direct determination of the orbital elements and component masses of compact binaries. The potential of the current Gaia data for this purpose has not yet been fully explored. In this work, we combine Gaia astrometric and radial velocity (RV) measurements to perform joint orbital fitting for 1,835 Gaia astrometric binaries with available RV data. Among the 518 systems with at least three effective RV epochs, we identify 171 compact binary candidates, including four with inferred dark companion masses exceeding 2.1 $M_{\odot}$ and 22 between 1.4 and 2.1 $M_{\odot}$. These candidates provide valuable targets for future follow-up observations and Gaia DR4 analyses. We also flag 1,317 additional potential compact binary candidates with effective RV epochs less than three, including 84 with inferred companion masses above 2.1 $M_{\odot}$ and 173 between 1.4 and 2.1 $M_{\odot}$. More importantly, our results point to a large, currently hidden population of black hole and neutron star binaries that can be further explored with the forthcoming Gaia DR4 data and complementary follow-up observations.

astro-ph.SR↗
arXiv2026

OS-Wizard: Scaling Computer-Using Agents to Specialized Software via Knowledge-Guided Exploration

Computer-using agents are increasingly capable of performing complex tasks in the digital world. However, high-quality interaction trajectories remain scarce for specialized desktop applications, where limited application knowledge, cumbersome GUI workflows, and incomplete exploration hinder automated collection. We introduce OS-Wizard, an automated trajectory collection framework. It combines tutorial-guided exploration with a hybrid GUI--Code action space, providing application-specific knowledge and simplifying GUI workflows through direct execution of application-native commands. The framework also checkpoints and revisits pivotal states during tree exploration to explore alternative workflows, complementing linear rollouts for broader coverage. We collect 10.9K trajectories comprising 54.9K state--action samples. Fine-tuning open-source base models on this corpus yields a best-performing model that improves performance on a benchmark of specialized desktop workflows by 58.2 percentage points to 64.1\%, surpassing reported human performance of 59.8\% and establishing a new state of the art. The fine-tuned model also improves performance on general desktop workflows using GUI actions only, suggesting that the collected supervision benefits general desktop tasks as well. Comparative experiments and ablations demonstrate the efficiency of trajectory collection and the importance of code actions.

cs.AI↗
arXiv2026

LoopWAM: Closed-Loop Generalist Policy Improvement from Deployment Rollouts

Scaling expert demonstrations alone does not enable generalist robot policies to improve from deployment. Real-world rollouts contain successes, suboptimal behavior, and failures: imitating every recorded action can reinforce poor behavior, while filtering out failures discards evidence of physical interactions and their consequences. We introduce LoopWAM, a closed-loop world-action model that learns directly from this full quality spectrum, using rollouts from both itself and other deployed policies. Source-dependent noise sampling unifies behavior learning and action-conditioned world modeling, while quality steering and quality-weighted supervision retain full video supervision across qualities and regulate action imitation. Video and action tokens interact throughout parallel denoising, allowing learned consequences to inform action generation directly. On RoboChallenge Table30-V2, three deployment-and-training rounds raise success from 23.67% to 43.33%, a 19.7-percentage-point gain, with a final score of 56.80. LoopWAM also achieves 94.5% average success on RoboTwin 2.0 and 49.5% overall success on RoboCasa365 without additional embodied pretraining, obtaining the best overall results among the compared methods across all three benchmarks. On challenging shirt folding task, two rounds of deployment-driven improvement after single-task post-training raise success from 13.20% to 55.56%. These results demonstrate closed-loop improvement from mixed-quality experience under both multi-task learning and single-task post-training. Project page: https://loop-wam.github.io/

cs.RO↗
arXiv2026

OS-Helix: Unifying GUI Interaction and Code Execution in Native Computer-Using Agents

Real desktop automation is neither purely visual nor purely programmatic. Agents must understand user-facing interfaces while efficiently manipulating files, scripts, and system state. However, native GUI/code coordination cannot be obtained by prompt-level tool composition. In our preliminary experiments, naively adding a code tool degrades a GUI agent, revealing that GUI/code routing must be internalized through training on high-quality hybrid data. We present OS-Helix, an end-to-end trained multimodal computer-using agent that unifies GUI actions and code execution in one action space. To construct the required supervision, we introduce application-artifact composition, pairing desktop applications with authentic file artifacts to instantiate executable Ubuntu states, and a decoupled instruction/evaluator generation pipeline that produces environment-grounded tasks whose final states can be verified regardless of the solution modality. We cold-start the policy with SFT on filtered hybrid data and further improve it through online RL in live desktop environments. For RL, we propose Step-balanced GSPO, which aggregates importance weight by action step rather than raw assistant tokens to prevent long code actions from dominating heterogeneous-action optimization. On OSWorld-Verified, OS-Helix-27B and OS-Helix-35B-A3B achieve relative improvements of 35.0% and 33.7% over their matched Qwen3.5 hybrid baselines, while OS-Helix-27B-RL further improves by 42.9% to reach 66.44% success. Ablations provide evidence consistent with contributions from learned GUI/code coordination, scalable verifiable supervision, and action step balanced optimization, while demonstrating cross-platform transfer on WindowsAgentArena. Our project homepage is at https://os-copilot.github.io/OS-Helix/.

cs.LG↗
arXiv2026

TAPNAV: Humanoid Navigation through Tactile Active Perception

Navigation in vision-denied environments is challenging for humanoid robots because proprioceptive odometry drifts and localization uncertainty accumulates rapidly. We present TAPNAV, a tactile active-perception framework that enables humanoid navigation toward a goal by actively probing surrounding structures without relying on vision. TAPNAV maintains a pose belief from odometry, IMU, and tactile contact observations, and couples uncertainty-aware global route planning with information-gain-driven local probing. The global planner searches for routes that keep predicted localization uncertainty bounded by exploiting opportunities for tactile correction, while the local planner selects probe actions that maximize expected information gain. A whole-body controller coordinates the humanoid's locomotion and end-effector contact to execute the planned navigation and probe motions. We evaluate TAPNAV in simulation and on a Unitree G1 across different floor plans and obstacle geometries. TAPNAV achieves lower state estimation error and a higher task completion rate than baselines. These results demonstrate that actively planning physical interactions with the environment can provide localization cues for reliable humanoid navigation without vision.

cs.RO↗
arXiv2026

Dataset-Free Compliant Humanoid Loco-Manipulation with Dynamic Online Posture

Most humanoid loco-manipulation controllers require human motion data to learn whole-body coordination and posture, leaving policies reliant on external sources to provide this data. We present OCLO (Online-posture Compliant LOco-manipulation), a humanoid loco-manipulation system trained without human motion data and commanded only through two end-effector targets. Because these targets do not uniquely determine whole-body posture, OCLO generates pelvis height and torso orientation online using an analytic reachability prior, further refined through policy-in-the-loop sampling with a task-agnostic cost. OCLO also learns whole-body compliance by displacing end-effector references according to measured forces through a spring-damper model, encouraging the legs, waist, and pelvis to yield to external loads. In simulation, using the reachability prior leads to a 77.8% success rate in acquiring the commanded reference, a vast improvement over the 37.8% success rate accomplished without the prior. Further, refinement reduces end-effector orientation error across all evaluated tasks. The same posture module improves a pretrained SONIC controller on four of five tasks. Without compliance training, policies tend to lose balance under disturbances rather than sacrifice tracking. On a Unitree G1, OCLO maintains balance under end-effector disturbances that cause its ablations to fail and performs seven loco-manipulation tasks, including crouched walking and picking up a box from a low surface. Project website: https://oclo-humanoid.github.io/

cs.RO↗
arXiv2026

High-Priority Compact-Object Candidates in Young Clusters from the \textit{Gaia} DR3 Non-Single-Star Catalog

The search for massive compact objects such as black holes (BHs) or neutron stars (NSs) in star clusters is longstanding challenge in astrophysics, yet no definitive identification of such compact objects in young clusters has been reported. In this Letter, we construct a sample of high-priority candidates for follow-up investigation. By cross-matching the cluster catalog with the Gaia Non-Single-Star database and estimating companion masses, we identify 13 systems with companion masses exceeding 1.4~$M_{\odot}$, three of which have masses above 2.1~$M_{\odot}$. Applying our newly developed Astrometric Mass-Ratio Function (AMRF) classification based on cluster-specific isochrones, we identify three systems whose companions cannot readily be explained by a single main-sequence star or an unresolved close binary. These systems have inferred companion masses exceeding 2.1~$M_{\odot}$, suggesting that they may be high-mass compact-object candidates. Joint Gaia and radial velocity orbit fitting for a comparison subset yields companion masses broadly consistent with the astrometric estimates, supporting the use of our framework for target prioritization. These unconfirmed systems are valuable targets for multi-epoch spectroscopy, high-resolution imaging, and future Gaia epoch astrometry.

astro-ph.GA↗
arXiv2026

EM-Based p-norm-like Constraint RLS Algorithm with Adaptive Regularization for Sparse System Identification

In this paper, the recursive least squares (RLS) algorithm is considered in the sparse system identification setting. The cost function of RLS algorithm is regularized by a p-norm-like ($0 \leq p \leq 1$) constraint of the estimated system parameters. To minimize the regularized cost function, we transform it into a penalized maximum likelihood (ML) problem that is solved by the expectation-maximization (EM) algorithm, and derive the tap-weight update through a thresholding operator whose sparsity also permits a low-complexity implementation. We further propose an error-driven adaptive regularization scheme operating on two timescales: a fast gate driven by the normalized excess a-priori error relaxes the penalty toward a lower bound, while a slow effective-support tracker sets its stationary target. Numerical simulations demonstrate the superiority of the new algorithm over the state-of-the-art sparse RLS algorithms, as well as conventional sparse RLS algorithm.

cs.IT↗
arXiv2026

ArtifactWorld: Scaling 3D Gaussian Splatting Artifact Restoration via Video Generation Models

3D Gaussian Splatting (3DGS) delivers high-fidelity real-time rendering but suffers from geometric and photometric degradations under sparse-view constraints. Current generative restoration approaches are often limited by insufficient temporal coherence, a lack of explicit spatial constraints, and a lack of large-scale training data, resulting in multi-view inconsistencies, erroneous geometric hallucinations, and limited generalization to diverse real-world artifact distributions. In this paper, we present ArtifactWorld, a framework that resolves 3DGS artifact repair through systematic data expansion and a homogeneous dual-model paradigm. To address the data bottleneck, we establish a fine-grained phenomenological taxonomy of 3DGS artifacts and construct a comprehensive training set of 107.5K diverse paired video clips to enhance model robustness. Architecturally, we unify the restoration process within a video diffusion backbone, utilizing an isomorphic predictor to localize structural defects via an artifact heatmap. This heatmap then guides the restoration through an Artifact-Aware Triplet Fusion mechanism, enabling precise, intensity-guided spatio-temporal repair within native self-attention. Extensive experiments demonstrate that ArtifactWorld achieves state-of-the-art performance in sparse novel view synthesis and robust 3D reconstruction. Code available at: https://github.com/fyting/ArtifactWorld.

cs.CV↗
arXiv2026

Does Gradient Conflict Predict the Understanding--Generation Trade-off? A Controlled Audit of Conflict-Metric Validity in Unified Multimodal Models

Unified multimodal models (UMMs) are increasingly designed around gradient conflict between understanding and generation objectives. The premise that reducing these metrics improves the downstream understanding-generation trade-off has never been tested directly. We audit it in a controlled testbed, GRIDUMM, which mirrors key structural ingredients of UMM training while making the ground-truth trade-off exactly computable. Across 63 configurations and 372 measured checkpoints, no directional conflict metric reaches an absolute Spearman correlation of 0.3 with a confidence interval excluding zero for conflict measured during training against the eventual trade-off. A dose-response intervention that monotonically suppresses conflict leaves the trade-off flat, separating correlation from causation. The norm ratio is a generation-failure detector and becomes null among configurations that master generation. Functional interference measures outperform directional conflict metrics, while training loss tracks the trade-off strongly. Our results do not show that conflict is useless; they show that its validity as a diagnostic target must be established, not assumed, and we release the audit protocol as a reusable standard.

cs.LG↗
arXiv2026

Beyond Layers: Position-Resolved Gradient Conflict and Position-Aware Modulation for Unified Multimodal Models

Unified multimodal models (UMMs) train image understanding and autoregressive image generation on shared parameters, and the two objectives are known to interfere. Existing diagnoses and remedies operate at the resolution of layers or experts, measuring conflict per layer and resolving it by separating parameters. We argue that this resolution hides an orthogonal axis. Generation in a UMM is next-token prediction over a raster sequence of visual tokens whose roles vary systematically with position, so how strongly a generation gradient interferes with understanding should depend on where in the sequence it originates. We introduce a position-resolved interference map that attributes understanding-generation gradient conflict to visual-token positions within every layer, computed from a single backward pass at $1.2\times$ the cost of a standard backward pass. On Show-o and Janus-Pro, position explains a large share of conflict variance after controlling for depth (partial $η^2=0.31$ vs. $0.35$ for layer on Show-o; $0.15$ vs. $0.30$ on Janus-Pro): the first quarter of the sequence has a mean gradient cosine of $-0.18$ against understanding, the last quarter $-0.02$. The dependence survives per-position gradient-norm normalization, retaining $80%$ of its effect size, and conflict strength tracks semantic content (Spearman $ρ=0.64$). Building on the map, we propose position-aware modulation (PAM), which removes the anti-aligned component of generation gradients only at high-conflict positions without changing the architecture. Under a matched trainable-parameter budget, PAM improves over layer-wise separation by $+21$ MME and $+2.4$ GenEval points on Show-o while matching it on POPE and overall FID; a random-position control recovers about $31%$ of the gain. Position-based and layer-based separation are complementary degrees of freedom and can be combined.

cs.CV↗
arXiv2026

Galactic Archaeology with the Subaru `Ōnohi`ula Prime Focus Spectrograph Strategic Program

The recently commissioned Subaru `Ōnohi`ula Prime Focus Spectrograph (PFS) will obtain spectra from nearly 2,400 fibers that cover 1.24 square degrees. The 360 night Subaru Strategic Program for PFS is dedicating approximately one-third of its allocation (130 nights) to study the structure and evolution of galaxies in the Local Group. This Galactic Archaeological survey has three pillars. (1) We will determine whether the mass density profiles of dwarf galaxies are consistent with cusps, as expected for cold dark matter, or cores, as expected from alternative dark matter theories or baryonic feedback. We will deduce the density profiles as a function of radius from modeling of the full line-of-sight velocity and abundance distributions for six dwarf galaxies. Our total sample will consist of 18,000 member stars to beyond the nominal tidal radius of each system. (2) From measurements of the [alpha/Fe] abundance ratio, we will learn the difference in assembly history of the two most massive galaxies in the Local Group: M31 and the Milky Way. We will observe 30,000 member stars over 45 square degrees of M31's halo and outer disk. (3) We will uncover how the most fragile (outer) part of the Milky Way responded to accretion events both in the distant past (such as Gaia-Sausage Enceladus) and in more recent history (such as the Sagittarius dwarf spheroidal galaxy). To support this study, PFS will provide velocities and metallicities--from which, in combination with photometry, we will deduce ages--for tens of thousands of main-sequence stars out to a Galactocentric distance of ~30 kpc.

astro-ph.GA↗
arXiv2026

When Depth Hurts: Reliability-Aware Geometry Distillation for Depth-Free RGB-D Salient Object Detection

Depth can resolve appearance ambiguity in RGB-D salient object detection (SOD), yet sensor depth is not uniformly reliable. Missing regions, blurred boundaries, and structural artifacts can propagate through multimodal fusion and make an RGB-D detector less accurate than its RGB-only counterpart. Existing quality-aware approaches regulate observed depth but remain dependent on the same potentially defective modality. We propose \method, a reliability-aware geometry distillation framework developed for RGB-D SOD benchmarks without using dataset-provided depth during training or inference. A frozen Depth Anything V2 model serves only as a training-time teacher, transferring dense relative geometry, hierarchical spatial attention, and boundary structure to a compact edge-aware geometry branch. Pooled bidirectional interaction aligns geometry with appearance, and a pixel-wise reliability estimator selectively injects geometry that is compatible with the current RGB representation. The teacher is removed after training, leaving an RGB-only inference network. Trained on 2,985 RGB-mask pairs, \method{} achieves the best or tied-best result in 26 of 36 metric-dataset comparisons against ten recent RGB-D SOD methods, including a 13.4\% relative MAE reduction on ReDWeb-S. When retrained on DUTS-TR, it also improves the strongest prior $F$-measure by 4.2\% on PASCAL-S, showing that the distilled geometry transfers beyond a particular sensor or dataset domain. Code will be released upon publication.

cs.CV↗
arXiv2026

TRACE: An Evidence-Grounded Benchmark for Safety Evaluation of Large Reasoning Models

Large Reasoning Models (LRMs) generate intermediate reasoning traces that may contain unsafe content, even when their final responses appear safe. Guardrail models are designed to detect and block unsafe content, yet existing benchmarks for unsafe content detection focus primarily on prompts and final responses, leaving reasoning traces largely unexamined. Moreover, these benchmarks typically provide only binary safety labels, without evidence annotations that justify the judgments. To address these limitations, we introduce TRACE, an evidence-grounded safety evaluation benchmark that covers the entire LRM inference pipeline: prompts, reasoning traces, and final responses. TRACE includes prompts in two languages spanning nine risk categories and ten attack strategies. For each prompt, four LRMs generate reasoning traces and final responses, and we annotate the safety of each component and extract supporting evidence from the corresponding source text. Evaluating 18 guardrail models on TRACE reveals that safety judgment for reasoning traces is substantially more challenging than for prompts or final responses, and that current models struggle to accurately extract supporting evidence. These findings highlight the need for guardrail models that can reliably detect and precisely localize unsafe content across the LRM inference pipeline.

cs.AI↗
arXiv2026

E2HiL: Entropy-Guided Sample Selection for Efficient Real-World Human-in-the-Loop Reinforcement Learning

Human-in-the-loop guidance has emerged as an effective approach for accelerating online reinforcement learning (RL) in real-world manipulation. However, existing human-in-the-loop RL (HiL-RL) frameworks often suffer from low sample efficiency, requiring substantial human interventions to achieve convergence and thereby leading to high labor costs. To address this, we propose a sample-efficient real-world human-in-the-loop RL framework named E2HiL, which requires fewer human interventions by actively selecting informative samples. Specifically, stable reduction of policy entropy enables improved trade-off between exploration and exploitation with higher sample efficiency. We first build influence functions of different samples on the policy entropy, which is efficiently estimated by the covariance of action probabilities and soft advantages of policies. Then we select samples with moderate values of influence functions, where shortcut samples that induce sharp entropy drops and noisy samples with negligible effect are pruned. Extensive experiments across 10 real-world manipulation tasks, spanning multiple embodiments and learning frameworks, demonstrate that E2HiL improves success rates by 24.9% while reducing human interventions by 9.3% compared to state-of-the-art HiL-RL baselines. These results validate its effectiveness as a policy- and embodiment-agnostic plug-and-play module for efficient real-world RL. The project page can be found at https://e2hil.github.io/.

cs.RO↗
arXiv2026

Impact of subhalo dynamical friction heating on the formation of the first structures in the universe

We present a model for gas heating, driven by dynamical friction from orbiting subhalos within dark matter halos. Using data from the TNG50 simulation, we derive the subhalo mass function and calculate the dynamical friction heating rate for a wide range of halo masses and redshifts from $z = 15$ to 0. Our results show that, by converting gravitational potential energy into thermal energy, dynamical friction is an important mechanism for galaxy quenching in massive halos at low redshifts, consistent with previous studies. Additionally, we find that in the early universe at $z \sim 15$, heating rates can be comparable to the molecular hydrogen cooling rates in metal-free minihalos. This can suppress gas cooling and fragmentation and does increase the critical molecular fraction for Pop III star formation by up to one order of magnitude, thereby making Pop III star formation more difficult. In combination with the Lyman-Werner background, the dynamical friction heating mechanism favors the formation of direct-collapse black hole (DCBH) seeds in atomic cooling halos, even when the average H$_2$ fraction is $\sim 10^{-5}$ during the minihalo progenitor phase. Dynamical friction heating at a fixed host halo mass can vary by two orders of magnitude due to the scatter in the number of subhalos. To capture dynamical friction heating in simulations, it is necessary to resolve subhalos with a subhalo to host halo mass ratio $ψ\gtrsim 0.05$.

astro-ph.GA↗
arXiv2026

Intern-S2-Preview: Scientific Agentic Foundation Model

Scientific discovery increasingly requires AI systems that can reason over scientific evidence of heterogeneous modalities, interact with scientific tools and environments, and sustain progress across long task horizons. We present Intern-S2-Preview, a series of scientific agentic foundation models designed to support multimodal scientific understanding, reasoning, generation, and long-horizon tasks. The training pipeline begins with scientific multimodal pre-training over rendered scientific documents, interleaved image-text data, and diverse scientific corpora. Starting from the pretrained checkpoint, we apply a unified post-training pipeline consisting of supervised fine-tuning, scalable multi-task reinforcement learning (RL), black- and white-box agentic RL, and on-policy distillation. This pipeline is supported by practical techniques that improve rollout and training stability and efficiency, including partial rollout with off-policy correction, adaptive length regularization, online speculative decoding, robust multi-task optimization, and trace-aware experience assembly for agentic tasks. At the architecture level, Intern-S2-Preview-397B extends time series modelling from efficient long-sequence understanding to numerical forecasting, while Memory Decoder is studied as a separate memory-augmented path for rapid scientific specialization without modifying the frozen 397B backbone. Evaluations across scientific, multimodal, agentic, and general-purpose benchmarks show that Intern-S2-Preview-397B achieves competitive or leading results in multiple settings. The time series modules improve scientific signal understanding and forecasting on SciTS, while the separate Intern-MemDec-4B extension improves the Biology-Instructions average score from 56.92 to 60.32 without modifying the frozen 397B backbone.

cs.LG↗
arXiv2026

MDB-Link: Hierarchical Schema Linking for Multi-Database Text-to-SQL

Traditional Text-to-SQL research and benchmarks assume a known target database, overlooking settings in which a query must be routed within a large, heterogeneous database collection. We therefore study schema linking in a multi-database setting, where the system must first locate the target database and then construct a compact, SQL-relevant schema for generation. We propose MDB-Link, a hierarchical schema-linking framework that retrieves question-relevant columns from a global index, aggregates retrieval evidence to shortlist databases, and uses a budget-aware large language model (LLM) for database reranking, table selection, and column grounding. With Qwen2.5-14B, MDB-Link outperforms LinkAlign on MMQA, Spider2-Snow, and BIRD-dev in database localization and column selection while producing schema subsets close in size to the gold schemas. Exact match improves from 16.88 to 51.41 on MMQA, 2.50 to 9.17 on Spider2-Snow, and 12.52 to 38.01 on BIRD-dev. MDB-Link also runs faster than LinkAlign and AutoLink, demonstrating the effectiveness of hierarchical schema reduction for downstream SQL generation.

cs.CL↗