arXiv subjects

Z. Zhang

Publications and source records attributed to Z. Zhang.

At least 19 recordsLinked to original sources

arXiv2026

Apodex 1.1: Scaling Agentic Intelligence for Complex Work

General-purpose language models can reason and synthesize knowledge, but complex work also requires sustained interaction with files, information sources, and executable code, together with state maintenance, failure recovery, and verifiable delivery. We call this \emph{working capability}: sustained, verifiable progress toward a real-world objective. Apodex 1.1 develops this capability along two complementary dimensions. \emph{Environment Scaling} expands the diversity and verifiability of executable file, search, and code environments, while \emph{Agentic Coordination Scaling} trains agents to decompose long-horizon tasks, delegate parallel work, integrate asynchronous results, and replan. A shared execution harness and AgentOS maintain task state and provenance across tools and agents, and training turns environment trajectories and coordination traces into reliable behavior. Across complex professional work, finance, scientific research, mathematics, coding, and search, Apodex 1.1 reaches the leading performance band despite using a substantially smaller model than many frontier systems. The 35B-parameter Apodex 1.1 Mini further retains strong working capability in a locally deployable form. These results ground agentic intelligence in useful, verifiable work completed over time and advance our goal of building a \emph{Heavy-Duty Solver} for ambitious, long-running tasks.

cs.AI↗

arXiv2026

The G347.3-0.5 outlier from O3: a follow-up case study for continuous gravitational-wave candidates

We report a multi-pipeline follow-up study of the continuous gravitational-wave (CW) outlier identified by the Ming et al. Einstein@Home directed search for the young supernova remnant G347.3-0.5 (containing the central compact object RXJ1713.7-3946). The outlier was initially identified in LIGO data from the O3a observing run and followed up with the addition of O3b and O4a data by the Einstein@Home pipeline, with O4a producing weaker evidence. Here, we show extended data quality checks and results from three pipelines that have independently investigated this outlier, including the newly released O4b data set. We can robustly recover the outlier in O3, while we find no evidence for a related standard CW signal in O4a (consistent with Ming et al.) and O4b data. This work serves as a useful test case for future multi-pipeline follow-ups of CW candidates that aim to investigate their astrophysical or noise nature, and it is therefore an important step in preparation for the first CW detection.

gr-qc↗

arXiv2026

Laser-intensity-spike-dominated hot electron generation from two-plasmon decay instability driven by moderate-bandwidth pulses

Our direct-drive-relevant experiments on the low-coherence Kunwu laser facility identify two-plasmon decay (TPD) as the primary source of hot electrons, and demonstrate for the first time that broadband laser pulses enhance TPD. Using particle-in-cell simulations, we attribute this TPD enhancement and the consequent hot electron production to stochastic intensity spikes inherent in broadband laser fields, robust in both weakly- and strongly-driven regimes. These findings suggest that mitigating hot electron generation requires suppressing these intensity spikes.

physics.plasm-ph↗

arXiv2026

Narrow spectral artifact investigation and mitigation in LIGO data from the fourth LIGO-Virgo-KAGRA observing run

We present efforts to identify, characterize, and mitigate narrow spectral artifacts in LIGO detector data during the fourth LIGO-Virgo-KAGRA observing run. Narrow spectral artifacts in gravitational-wave detectors are non-astrophysical noise sources that can degrade searches for narrowband persistent gravitational waves. Identifying and, where possible, mitigating these noise sources is one of the core efforts of the LIGO Detector Characterization group. Key software tools have been updated and new tools deployed for the fourth LIGO-Virgo-KAGRA observing run to facilitate investigations and data monitoring. We discuss these tool upgrades, and present several identified narrowband artifacts that have been successfully investigated and mitigated in LIGO data. Regardless of whether artifacts are mitigated or not, narrowband persistent gravitational-wave searches require information on which frequency bands contain non-astrophysical artifacts. Minimizing the number of bands containing non-astrophysical artifacts is essential to maximize the potential for discovery of a new class of gravitational-wave signals.

astro-ph.IM↗

arXiv2026

No Blue without Red: Evolutionary Properties of Super-Early Galaxies

The discovery of numerous luminous, super-early galaxies at $z>10$ by JWST has revealed a striking diversity in their ultraviolet (UV) properties, ranging from extremely blue, dust-poor systems to a smaller population of significantly reddened sources. We investigate the physical origin of this diversity within the framework of the Attenuation-Free Model (AFM), in which radiation-driven outflows redistribute dust to large galactic radii, reducing the effective attenuation. Applying the model to a sample of 32 spectroscopically confirmed super-early galaxies, we derive their key physical properties, including halo mass, star formation efficiency, metallicity, and outflow extent. We find that these systems reside in massive halos ($\log M/M_\odot \sim 10.7$) and exhibit moderate ($0.01 \lesssim \epsilon_* \lesssim 0.05$) star formation efficiencies, while frequently reaching super-Eddington conditions that trigger powerful outflows. Within this framework, we propose an evolutionary sequence in which galaxies transition from a dust-obscured ``Red Monster'' phase to a UV-bright ``Blue Monster'' phase as outflows clear their central regions. The recently confirmed red galaxy EGS-z11-R0 at $z=11.45$ is naturally interpreted as a system observed during this obscured phase. Compact ($r_e \lesssim 150$ pc) sources are instead difficult to reconcile within AFM; we speculate that their emission is dominated by an AGN. Our results provide a unified interpretation of super-early galaxy properties and highlight the key role of radiation-driven outflows in shaping galaxy evolution at cosmic dawn. Future observations with JWST and ALMA will be essential to test these predictions and further constrain the nature of the earliest galaxies.

astro-ph.GA↗

arXiv2026

Oxygen and nitrogen isotopologues on cold COCONUTS-2b observed with MIRI/MRS

Linking the composition of gas giant planets to their formation paths has long been a goal in exoplanet science. Especially, cold gas giants with temperatures below $\sim$500K have been out of reach for detailed atmospheric characterization. With JWST, however, we can reach high signal-to-noise (S/N) spectra for such cool worlds and can can measure not only their main trace gas abundances, but even their isotopic content unlocking new possibilities in linking them to their formation paths. In this study, we present the spectrum of one of the coldest planetary-mass companions COCONUTS-2b ($\mathrm{T_{eff}}\approx$480K, separation of $\sim$6400 au from its M dwarf host star) obtained with the Mid-InfraRed Instrument Medium Resolution Spectrometer (MIRI/MRS). Combining the MIRI and archival Gemini/FLAMINGOS-2 data sets, we aim to characterize the chemical composition and physical structure of its frigid atmosphere, setting the stage to uncover insights on the formation of COCONUTS-2b. For the first time on a MIRI/MRS data set, we use the full spectral resolution of MIRI/MRS and perform atmospheric retrievals to unlock the search for faint absorption features by rare molecules and isotopologues. The latter are identified using a leave-one-out analysis and Bayes factor comparison. We robustly detect three isotopologues, namely $^{15}$NH$_3$, H$_2^{18}$O and H$_2^{17}$O in the atmosphere of COCONUTS-2b. We find the first clear evidence of oxygen isotopes in water in a cold companion. This data set demonstrates the capability of MIRI/MRS to characterize such cold planetary-mass companion's atmospheres with respect to their compositional and isotopic content. In the future, the constrained elemental and isotope ratios provide a unique avenue in comparing with the host star's abundances and eventually in tracing formation scenarios.

astro-ph.EP↗

arXiv2026

Advances in photocathode development for PICOSEC Micromegas precise-timing detectors

The PICOSEC Micromegas detector is a~precise-timing gaseous detector that combines a Cherenkov radiator, a~semi-transparent photocathode and a Micromegas amplification stage, targeting time resolutions of tens of picoseconds for minimum ionising particles (MIPs). Initial single-pad prototypes achieved time resolutions of $\sigma<25$ ps, demonstrating strong potential for High Energy Physics (HEP) applications and beyond. The objective of this paper is a comprehensive characterisation of photocathodes, with a strong focus on robust materials while preserving excellent timing performance. The study includes laboratory measurements of optical and resistive properties, along with beam tests using 150 GeV/$c$ muons to evaluate the time resolution and photoelectron yield for various photocathodes. The best performance was obtained by a 5 nm Cesium Iodide (CsI) photocathode, reaching $\sigma = 10.9 \pm 0.3$ ps with more than 30 extracted photoelectrons, representing the most precise time resolution achieved by PICOSEC Micromegas to date. Metallic and carbon-based photocathodes, including Titanium (Ti), Boron Carbide (B$_4$C) and Diamond-Like Carbon (DLC), were also tested, with Ti and B$_4$C emerging as the most promising alternatives, achieving $\sigma \approx 30$ ps with about 5 extracted photoelectrons. These results demonstrate that improved robustness can be achieved while maintaining excellent time resolution, supporting the feasibility of using the PICOSEC Micromegas concept in future experiments.

physics.ins-det↗

arXiv2026

Timing performance of large prototype based on $\upmu$RWELL- PICOSEC detector technology with $10 \times 10\ \mathrm{cm}^{2}$ active area

The $\upmu$RWELL-PICOSEC detector, which combines a $\upmu$RWELL gaseous amplification structure with a Cherenkov radiator and photocathode, is a novel approach to acheive fast and precise timing in gaseous detectors. With timing precision at the level of tens of picoseconds, this technology is particularly suited for time-of-flight (TOF) applications in particle physics and potentially medical imaging. Beam tests with a 150~GeV/$c$ muon beam have been carried out on a large-area (10~$\times$~10~cm$^{2}$) prototype equipped with a cesium iodide (CsI) photocathode. Using an oscilloscope-based single-channel readout, timing measurements on two individual pads of the detector have yielded resolutions of $\approx$ 48 ps and $\approx$ 52 ps under different biasing conditions respectively.

physics.ins-det↗

arXiv2026

Spatial resolution improvement of PICOSEC Micromegas precise timing detectors

The combination of a Cherenkov radiator with a semi-transparent photocathode and a Micromegas based amplification stage allows PICOSEC Micromegas detectors to achieve a time resolution of better than 15ps. While tileable prototypes with 10x10 channels feature 1x1 cm^2 readout pads, finer readout granularity can be used to improve the spatial resolution. We report on the study of high readout granularity PICOSEC Micromegas prototypes which achieve around 0.5mm spatial resolution with 3.5mm large pads. No significant improvement was found when readout pad size was further reduced to 2.2mm. The timing resolution of the leading pad was found to be slightly degraded but remained better than 20ps for a medium granularity prototype. The achieved spatial resolution can enable PICOSEC Micromegas to be used as precise timing and moderate resolution tracking detector simultaneously.

physics.ins-det↗

arXiv2026

MiroThinker-1.7 & H1: Towards Heavy-Duty Research Agents via Verification

We present MiroThinker-1.7, a new research agent designed for complex long-horizon reasoning tasks. Building on this foundation, we further introduce MiroThinker-H1, which extends the agent with heavy-duty reasoning capabilities for more reliable multi-step problem solving. In particular, MiroThinker-1.7 improves the reliability of each interaction step through an agentic mid-training stage that emphasizes structured planning, contextual reasoning, and tool interaction. This enables more effective multi-step interaction and sustained reasoning across complex tasks. MiroThinker-H1 further incorporates verification directly into the reasoning process at both local and global levels. Intermediate reasoning decisions can be evaluated and refined during inference, while the overall reasoning trajectory is audited to ensure that final answers are supported by coherent chains of evidence. Across benchmarks covering open-web research, scientific reasoning, and financial analysis, MiroThinker-H1 achieves state-of-the-art performance on deep research tasks while maintaining strong results on specialized domains. We also release MiroThinker-1.7 and MiroThinker-1.7-mini as open-source models, providing competitive research-agent capabilities with significantly improved efficiency.

cs.CL↗

arXiv2026

Detailed lens modeling and kinematics of the submillimeter galaxy G09v1.97. An analysis of CO, H2O, H2O+, and dust continuum emission

The formation mechanisms of intensely starbursting galaxies at high redshift remain unknown. One possible mechanism for triggering these starbursts is mergers and interactions, but detecting these at high redshift remains a challenge. Observations of high-redshift gravitationally lensed galaxies enable studies of the interstellar medium and environment of these extreme starbursts in detail. We used high angular resolution observations of dust continuum, CO(6-5), H2O(211-202), and H2O+(202-111) emission to constrain the ongoing processes in the z = 3.63 gravitationally lensed submillimeter galaxy H-ATLAS J083051.0+013224 (G09v1.97). We used PyAutoLens to create a de-magnified source plane CO(6-5) emission line cube and performed kinematic modeling using 3DBarolo. Additionally, we investigated the properties of the continuum and molecular line emission in the source plane. We find that the regions of CO(6-5) and H2O(211-202) emission match closely in the source plane but that the dust continuum emission is more compact. We find that our lens modeling results do not require more than one source, contrary to what has been found in previous studies. Instead, we find that G09v1.97 resembles a rotating disk with Vmax/sigma = 2.8 +/- 0.4 with evidence for residual emission indicative of non-circular motions such as outflows, tidal tails, or an additional background galaxy. We suggest that the origin of the non-circular motions may be associated with a bi-conical outflow, a tidal tail from an interaction, or indicate the possible presence of an additional galaxy. We calculate the dynamical mass, gas mass, star-formation rate, and depletion time for G09v1.97 and find a high star-formation rate and low gas depletion time. In combination, this suggests that G09v1.97 has recently undergone an interaction, triggering intense star formation, and is in the process of settling into a disk.

astro-ph.GA↗

arXiv2025

Deep Eigenspace Network for Parametric Non-self-adjoint Eigenvalue Problems

We consider operator learning for efficiently solving parametric non-self-adjoint eigenvalue problems. To overcome the spectral instability and mode switching associated with non-self-adjoint operators, we choose to learn the eigenspace rather than individual eigenfunctions. In particular, we propose a Deep Eigenspace Network (DEN) architecture integrating Fourier Neural Operators, geometry-adaptive POD bases, and explicit banded cross-mode mixing mechanism to capture complex spectral dependencies. We apply DEN to the non-self-adjoint Steklov eigenvalue problem and prove the Lipschitz continuity of the eigenspace with respect to the parameter. Furthermore, we derive error bounds for the eigenvalues. Numerical experiments validate that DEN is highly effective and efficient.

math.NA↗

arXiv2025

Structure of Jupiter's High-Latitude Storms: Folded Filamentary Regions Revealed by Juno

Sprawling, turbulent cloud formations dominate the meteorology of Jupiter's mid-to-high latitudes, known as Folded Filamentary Regions (FFRs). A multi-wavelength characterisation by Juno reveals the spatial distribution, vertical structure, and energetics of the FFRs. The cloud tops display multiple lobes of stratiform aerosols, separated by darker, cloud-free lanes, and embedded with smaller eddies and high-altitude cumulus clouds. These cyclonic FFRs are microwave-bright in shallow-sounding wavelengths ($p<5$ bars) and microwave-dark in deep-sounding wavelengths ($p>10$ bars), with the transition potentially associated with the water condensation layer (6-7 bars). Associating microwave contrasts with temperature anomalies, this implies despinning of cyclonic eddies above/below their mid-planes. Despite deep roots (being detectable in wavelengths sounding $\sim100$ bars), they are ``pancake vortices'' with horizontal extents at least an order of magnitude larger than their depth. In the northern hemisphere, FFRs are most common in cyclonic belts poleward of $40^\circ$N (all latitudes are planetocentric), particularly a North Polar Filamentary Belt (NPFB) near $66-70^\circ$N that defines the transition from organised belts/zones to the chaotic polar domain. This distribution explains the high lightning rates from $45-80^\circ$N, peaking in a belt poleward of $52.3^\circ$N, which may trace the availability of water for moist convection. Many observed lightning flashes can be associated to specific FFRs containing bright storms, but some FFRs display no activity, suggesting quiescent periods during storm evolution. Analogies to Earth's oceanic eddies suggest that cyclones deform isentropic surfaces at their midplanes, raising deep water-rich layers upwards to promote moist convection, release latent heat, and inject clouds into the upper troposphere.

astro-ph.EP↗

arXiv2025

Design and Performance of a 96-channel Resistive PICOSEC Micromegas Detector for ENUBET

The PICOSEC-Micromegas (PICOSEC-MM) detector is a fast gaseous detector that achieves picosecond-level timing by coupling a Cherenkov radiator, typically an MgF2 crystal, to a Micromegas-based photodetector with a photocathode. This configuration allows the fast photoelectron-induced signal to suppress the intrinsic time jitter of gaseous detectors, enabling sub-20 ps timing precision while preserving the robustness and scalability of micro-pattern gaseous detector technologies. The 96-pad PICOSEC-MM detector is a large-area demonstrator optimized for precision timing in high-energy physics, building on research and development insights from earlier 7-pad resistive prototypes to validate scalability, uniformity, and robustness for the ENUBET project. It employs a 2.5 nm diamond-like carbon photocathode and a Micromegas board with a surface resistivity of 10 megaohms per square, and was characterized using 150 GeV/c muons at the CERN SPS beamline, with one-third of the active area instrumented per run. A dedicated alignment procedure for multi-pad PICOSEC-MM systems was used to reconstruct pad centers and merge measurements across regions, yielding a timing resolution of 43 ps and uniform signal arrival time distributions over the tested area. Mechanical flatness was identified as a key factor, with planarity tolerances within 10 micrometers required to maintain good timing resolution, and the successful operation of the 96-pad demonstrator confirms the scalability of the PICOSEC-MM concept toward robust, high-granularity, picosecond-level gaseous timing detectors for monitored neutrino beam experiments such as ENUBET.

physics.ins-det↗

arXiv2025

Performance Optimization and Characterization of 7-pad Resistive PICOSEC Micromegas Detectors

We present a comprehensive characterization of resistive PICOSEC Micromegas detector prototypes, tested under identical conditions, constant drift gap, field configurations, and photocathode at the CERN SPS H4 beam line. This work provides a proof of concept for the use of resistive layer technology in gaseous timing detectors, demonstrating that robustness can be improved without compromising the excellent timing performance of PICOSEC Micromegas. Different resistive architectures and values were explored to optimize stability and ensure reliable long-term operation in challenging experimental environments. The prototype with a 10M{\Omega} resistive layer achieved the best overall performance, with a timing resolution of 22.900 {\pm} 0.002 ps and a spatial resolution of 1.190 {\pm} 0.003 mm, while charge sharing across multiple pads enabled combined timing resolutions below 28 ps. A lower-resistivity (200k{\Omega}) configuration exhibited enhanced charge spread, leading to minor systematic offsets in reconstructed pad centers, yet maintained robust timing and spatial performance. Capacitive charge-sharing architectures improved spatial resolution in some regions but suffered from signal attenuation and nonuniform charge distributions, resulting in slightly degraded timing (33.300 {\pm} 0.002 ps) and complex localization patterns. Mechanical precision, particularly readout planarity and photocathode alignment, was identified as critical for uniform detector response. These studies benchmark the potential of resistive layers for gaseous timing detectors and provide a foundation for scalable designs with optimized timing and spatial resolution across diverse experimental applications.

physics.ins-det↗

arXiv2025

Efficient Dynamic and Momentum Aperture Optimization for Lattice Design Using Multipoint Bayesian Algorithm Execution

We demonstrate that multipoint Bayesian algorithm execution can overcome fundamental computational challenges in storage ring design optimization. Dynamic (DA) and momentum (MA) optimization is a multipoint, multiobjective design task for storage rings, ultimately informing the flux of x-ray sources and luminosity of colliders. Current state-of-art black-box optimization methods require extensive particle-tracking simulations for each trial configuration; the high computational cost restricts the extent of the search to $\sim 10^3$ configurations, and therefore limits the quality of the final design. We remove this bottleneck using multipointBAX, which selects, simulates, and models each trial configuration at the single particle level. We demonstrate our approach on a novel design for a fourth-generation light source, with neural-network powered multipointBAX achieving equivalent Pareto front results using more than two orders of magnitude fewer tracking computations compared to genetic algorithms. The significant reduction in cost positions multipointBAX as a promising alternative to black-box optimization, and we anticipate multipointBAX will be instrumental in the design of future light sources, colliders, and large-scale scientific facilities.

physics.acc-ph↗

arXiv2025

Threshold $J/\psi$ Photoproduction as a Probe of Nuclear Gluon Structure

The nuclear EMC effect is the observation that quark distributions in bound nucleons experience significant modification at large $x$ relative to free nucleons. Despite decades of measurements verifying the presence of this effect in quarks across a wide range of nuclei, behavior of large-$x$ gluons in nuclei remains almost completely unknown. As the nuclear physics community seeks out new observables to try to elucidate the mechanisms behind the EMC effect, it becomes striking that we remain ignorant regarding the impact of nuclear effects on gluonic behavior. Recent photonuclear data using the Hall D photon beam have enabled the first measurement of $J/\psi$ photoproduction from nuclei near and below the energy threshold, with the results highlighted in Physical Review Letters as an Editors' Suggestion. These data have placed the first, and currently only, constraints on the behavior of large-$x$ gluons within bound nucleons. However, compared to the quantity of data which currently informs our knowledge of the quark-sector EMC effect, these data are extremely limited, and remain unable to conclusively observe or exclude large modification of gluon distributions. A high-luminosity photonuclear experiment will enable a precision measurement of incoherent $J/\psi$ photoproduction at and below the threshold region. This data will provide the first stringent constraints on nuclear modification of gluon structure or other exotic effects which could impact the production of $J/\psi$ from nuclei. We request 85 PAC days at Hall D using the GlueX detector with a 12 GeV electron beam energy and a coherent photon peak energy of $8$ GeV, split into 80 days using a $^4$He target and 5 calibration days using a $^2$H target.

nucl-ex↗

arXiv2025

FAUST. XXVIII. High-Resolution ALMA Observations of Class 0/I Disks: Structure, Optical Depths, and Temperatures

We present high-resolution (~7.5 au) ALMA observations at 1.3 and 3 mm of 16 disks around Class 0/I protostars across multiple star-forming regions and a variety of multiplicities, showing a range of disk sizes (~2-100 au) and including circumbinary disks (CBDs) in binaries with separations <100 au. The disk properties show similarities to Class II disks, including (a) low spectral index (SI) values (alpha=2.1) that increase with disk radius, (b) 3 mm disk sizes only marginally smaller than at 1.3 mm (<10%), and (c) radial intensity profiles well described by modified self-similar profiles. We also find key differences: (i) SI values increasing with radius, but exceeding 2 only at the disk edge (ii) higher brightness temperatures Tb, in some cases higher than the predicted temperatures due to irradiation, and (iii) ~10x higher luminosity at a given size compared to the Class II disks. These results confirm significant optical depth in the observed Class 0/I disks, at both 1.3 and 3 mm, helping to explain their higher luminosities, but higher temperatures are also required for the most compact (< 40 au) disks, suggesting additional viscous heating. Considering optical depth, most disk dust masses are estimated in the range 30-900 Mearth (0.01-0.3 Msun in gas), resulting in some disks reaching marginal gravitational instability. The median location of the water iceline is ~3 au, but it can extend beyond 10-20 au for the hottest disks. CBDs exhibit lower optical depths at both wavelengths and hence higher SI values (alpha=3.0), dust masses of 100 Mearth, and beta~1.5 (2 Class 0 CBDs) and beta~1 (1 Class I CBD), suggesting substantial grain growth only in the more evolved CBD. The inferred high optical depths provide a compelling explanation for the apparent scarcity of dust substructures in the younger disks at ~ 1 mm, despite mounting evidence for early planet formation (ABRIDGED).

astro-ph.SR↗