BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Yifan Zhou

Yifan Zhou contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
astro-ph.EPastro-ph.SRArtificial IntelligenceComputer VisionRoboticsComputation and LanguageMachine Learningastro-ph.IMcond-mat.mtrl-scieess.IVeess.SPeess.SYGraphicsInformation Theorymath.ITMethodologySystems and Control

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
25works
0followers
17topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

25 published item(s)

preprint2026arXiv

3D Skew Gaussian Splatting with Any Camera Trajectory Visualization Engine

While 3D Gaussian Splatting (3DGS) has revolutionized real-time photorealistic view synthesis, its fundamental reliance on symmetric Gaussian distributions introduces visual artifacts that hinder accurate spatial data exploration. Specifically, symmetric kernels struggle to capture shape and color discontinuities , which cause blurriness and primitive redundancy that mislead human perception during visual analysis. To address these visualization barriers, we introduce 3D Skew Gaussian Splatting (3DSGS), a novel framework that significantly enhances the structural fidelity and compactness of explicit scene representations. Our key insight lies in extending the standard primitive to a general Skew Gaussian counterpart. This generalized primitive inherits the highly efficient rasterization properties of standard Gaussians while gaining intrinsic asymmetric modeling capabilities. We couple this with an enhanced opacity representation to better handle complex transparency, alongside a depth-aware densification strategy that intelligently manages primitive allocation. Furthermore, to make these advancements actionable for real-world visual analytics, we re-derive the CUDA rasterization pipeline to universally support both symmetric and skew Gaussians, integrating it into a decoupled, free-camera interactive visualization engine. Extensive experiments demonstrate that 3DSGS achieves superior rendering quality and structural compactness, particularly in regions with intricate details, while maintaining the real-time frame rates necessary for fluid interactive exploration. Supplementary derivations and visual results are available at \textbf{\textit{https://3d-skew-gs.github.io/}}.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Instruction Tuning Changes How Upstream State Conditions Late Readout: A Cross-Patching Diagnostic

Recent interpretability work has identified model-internal handles on post-trained behavior, including refusal directions, assistant/persona axes, and sparse chat-tuning features. These results localize where behaviors can be read out or controlled, often in middle-to-late layers. We ask how earlier computation and the late stack cooperate to turn those differences into next-token margins. To test this, we introduce first-divergence cross-patching: at the first token where pretrained base (PT) and instruction-tuned (IT) checkpoints disagree, we cross each model's earlier-layer state with each model's late stack. The diagnostic separates training recipes: same-base instruction-following descendants show late effects that depend on their own earlier-layer state, while OpenMath2 math-domain SFT and controlled code/biomed CPT controls with verified domain learning do not; for OpenMath2, the late effect is already largely portable from base earlier-layer state. Across five dense families (4B-32B), the IT late stack adds +0.76 logits from PT upstream and +2.44 from IT upstream, giving a +1.68 interaction that is positive in every family. Thus the late stack has a real PT-upstream effect, but its larger effect in the IT checkpoint appears only when it reads its own post-trained upstream state. Sparse features in final MLP layers partially mediate the effect and are driven by upstream patches, supporting a handoff from earlier state to final-layer feature activation to IT-token margin. Forced-token scoring shows that the local token choice can change later exact-answer success. Operationally, paired-checkpoint studies that localize a difference to late layers should test whether it survives under the other checkpoint's upstream state before treating the late stack as self-contained.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

SciEvalKit: An Open-source Evaluation Toolkit for Scientific General Intelligence

We introduce SciEvalKit, a unified benchmarking toolkit designed to evaluate AI models for science across a broad range of scientific disciplines and task capabilities. Unlike general-purpose evaluation platforms, SciEvalKit focuses on the core competencies of scientific intelligence, including Scientific Multimodal Perception, Scientific Multimodal Reasoning, Scientific Multimodal Understanding, Scientific Symbolic Reasoning, Scientific Code Generation, Science Hypothesis Generation and Scientific Knowledge Understanding. It supports six major scientific domains, spanning from physics and chemistry to astronomy and materials science. SciEvalKit builds a foundation of expert-grade scientific benchmarks, curated from real-world, domain-specific datasets, ensuring that tasks reflect authentic scientific challenges. The toolkit features a flexible, extensible evaluation pipeline that enables batch evaluation across models and datasets, supports custom model and dataset integration, and provides transparent, reproducible, and comparable results. By bridging capability-based evaluation and disciplinary diversity, SciEvalKit offers a standardized yet customizable infrastructure to benchmark the next generation of scientific foundation models and intelligent agents. The toolkit is open-sourced and actively maintained to foster community-driven development and progress in AI4Science.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

SciFig: Towards Automating Scientific Figure Generation

Creating high-quality figures and visualizations for scientific papers is a time-consuming task that requires both deep domain knowledge and professional design skills. Despite over 2.5 million scientific papers published annually, the figure generation process remains largely manual. We introduce $\textbf{SciFig}$, an end-to-end AI agent system that generates publication-ready pipeline figures directly from research paper texts. SciFig uses a hierarchical layout generation strategy, which parses research descriptions to identify component relationships, groups related elements into functional modules, and generates inter-module connections to establish visual organization. Furthermore, an iterative chain-of-thought (CoT) feedback mechanism progressively improves layouts through multiple rounds of visual analysis and reasoning. We introduce a rubric-based evaluation framework that analyzes 2,219 real scientific figures to extract evaluation rubrics and automatically generates comprehensive evaluation criteria. SciFig demonstrates remarkable performance: achieving 70.1$\%$ overall quality on dataset-level evaluation and 66.2$\%$ on paper-specific evaluation, and consistently high scores across metrics such as visual clarity, structural organization, and scientific accuracy. SciFig figure generation pipeline and our evaluation benchmark will be open-sourced.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

SkillGenBench: Benchmarking Skill Generation Pipelines for LLM Agents

As LLM agents are increasingly built around reusable skills, a central challenge is no longer only whether agents can use provided skills, but whether they can generate correct, reusable, and executable skills from repositories and documents. Existing benchmarks primarily evaluate the efficacy of given skills or the ability of agents to solve downstream tasks from raw context, but they do not isolate skill generation itself as the object of study. We introduce SkillGenBench, a benchmark for evaluating skill generation pipelines under a unified and controlled protocol. In SkillGenBench, a generator receives raw corpora and produces standardized skill artifacts, which are then executed under fixed harnesses and assessed with unified evaluation procedures. The benchmark covers two generation regimes: task-conditioned generation, where a task-specific skill is synthesized after the task is revealed, and task-agnostic generation, where a reusable skill library must be distilled before downstream tasks are known. It also spans two complementary procedural sources: repository-grounded instances, where procedures are distributed across code, configuration, and scripts, and document-grounded instances, where procedures and constraints must be distilled from long-form text. We provide standardized task specifications, pinned environments, and evaluation protocols centered on deterministic execution-based checks, supplemented by auxiliary signals for diagnosis. Experiments across a range of skill-generation methods and backbones show substantial performance variation, highlight the difficulty of reusable skill distillation, and reveal distinct failure modes in skill generation from software repositories versus long-form documents. SkillGenBench establishes a reproducible testbed for studying skill generation as an independent research problem in agent systems.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

StraTA: Incentivizing Agentic Reinforcement Learning with Strategic Trajectory Abstraction

Large language models (LLMs) are increasingly used as interactive agents, but optimizing them for long-horizon decision making remains difficult because current methods are largely purely reactive, which weakens both exploration and credit assignment over extended trajectories. In this work, we present Strategic Trajectory Abstraction (StraTA), a simple framework that introduces an explicit trajectory-level strategy into agentic reinforcement learning (RL). StraTA samples a compact strategy from the initial task state, conditions subsequent actions on that strategy, and trains strategy generation and action execution jointly with a hierarchical GRPO-style rollout design, further enhanced by diverse strategy rollout and critical self-judgment. Experiments on ALFWorld, WebShop, and SciWorld show that StraTA consistently improves both sample efficiency and final performance over strong baselines. StraTA reaches success rates of 93.1% on ALFWorld and 84.2% on WebShop. On SciWorld, StraTA attains a 63.5% overall score, outperforming frontier closed-source models.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

The Convergence Gap: Instruction-Tuned Language Models Stabilize Later in the Forward Pass

Final outputs hide when a checkpoint commits to its next-token prediction. We introduce the convergence gap, a model-diffing diagnostic that decodes each layer's next-token distribution and measures its distance to the model's own final distribution. Across six paired pretrained and instruction-tuned checkpoints in native prompting regimes, instruction-tuned checkpoints remain farther from their final predictions later into the stack. The effect persists under endpoint-matched raw and tuned readouts, endpoint-free same-history checks, and fixed-history template replay. Matched-prefix interventions identify late MLP windows as the largest tested leverage point: late IT grafts into PT hosts increase late KL by +0.34 nats, while PT-late swaps into IT hosts reduce it by -0.51 nats; matched random late perturbations give only +0.003 versus +0.327 for the true late graft. A preselected Gemma case study provides behavior-facing plausibility for the same late swap, without serving as a benchmark claim. These results identify a robust predictiondynamics signature of post-training: released instruction-following checkpoints tend to settle later, and late MLP computation is the strongest tested bidirectional handle on that delay under matched histories.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

WorldMem: Long-term Consistent World Simulation with Memory

World simulation has gained increasing popularity due to its ability to model virtual environments and predict the consequences of actions. However, the limited temporal context window often leads to failures in maintaining long-term consistency, particularly in preserving 3D spatial consistency. In this work, we present WorldMem, a framework that enhances scene generation with a memory bank consisting of memory units that store memory frames and states (e.g., poses and timestamps). By employing a memory attention mechanism that effectively extracts relevant information from these memory frames based on their states, our method is capable of accurately reconstructing previously observed scenes, even under significant viewpoint or temporal gaps. Furthermore, by incorporating timestamps into the states, our framework not only models a static world but also captures its dynamic evolution over time, enabling both perception and interaction within the simulated world. Extensive experiments in both virtual and real scenarios validate the effectiveness of our approach.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Atmospheric Characterization of Hot Jupiter CoRoT-1 b Using the Wide Field Camera 3 on the Hubble Space Telescope

Exoplanet CoRoT-1 b is intriguing because we predict it to be a transitional planet between hot Jupiters (equilibrium temperatures ~ 1500 K) and ultra-hot Jupiters (equilibrium temperatures > 2000 K). In 2012, observations of CoRoT-1 b included one primary transit and three secondary eclipses with the Hubble Space Telescope (HST) Wide Field Camera 3 (WFC3) combined with the G141 grism (1.1-1.7 $μ$m) in stare mode. We aimed to further investigate CoRoT-1 b through its secondary eclipses, producing spectrophotometric light curves corrected for charge trapping, also known as the ramp effect in time-series observations with the WFC3. We found that, when correcting for the ramp effect and using the typically discarded first orbit, we are better capable of constraining and optimizing the emission and transmission spectra. We did a grid retrieval in this transitional temperature regime and found the spectra for CoRoT-1 b to be featureless and to agree with an inverted temperature-pressure (T-P) profile. We note, however, that the contribution function for the WFC3 indicates pressures probed near $10^{-3}$ to $10^{0}$ bar, which correspond to a nearly isothermal region in our T-P profile, thereby indicating that the inversion at high altitude is model-dependent. Despite no distinct features, the analysis done on CoRoT-1 b paves the way to high-precision results with stare mode spectroscopy. As a new generation of observations from the James Webb Space Telescope (JWST) approaches, CoRoT-1 b might be an interesting follow-up target because the time-series spectroscopic modes of JWST's NIRSpec, MIRI, and NIRCam instruments will be analogous to HST's stare mode.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

CSFlow: Learning Optical Flow via Cross Strip Correlation for Autonomous Driving

Optical flow estimation is an essential task in self-driving systems, which helps autonomous vehicles perceive temporal continuity information of surrounding scenes. The calculation of all-pair correlation plays an important role in many existing state-of-the-art optical flow estimation methods. However, the reliance on local knowledge often limits the model's accuracy under complex street scenes. In this paper, we propose a new deep network architecture for optical flow estimation in autonomous driving--CSFlow, which consists of two novel modules: Cross Strip Correlation module (CSC) and Correlation Regression Initialization module (CRI). CSC utilizes a striping operation across the target image and the attended image to encode global context into correlation volumes, while maintaining high efficiency. CRI is used to maximally exploit the global context for optical flow initialization. Our method has achieved state-of-the-art accuracy on the public autonomous driving dataset KITTI-2015. Code is publicly available at https://github.com/MasterHow/CSFlow.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

HST/WFC3 H$α$ Direct-Imaging Detection of a Point-like Source in the Disk Cavity of AB Aur

Accreting protoplanets enable the direct characterization of planet formation. As part of a high-contrast imaging search for accreting planets with the Hubble Space Telescope (HST) Wide Field Camera 3, we present H$α$ images of AB Aurigae (AB Aur), a Herbig Ae/Be star harboring a transition disk. The data were collected in two epochs of direct-imaging observations using the F656N narrow-band filter. After subtracting the point spread function of the primary star, we identify a point-like source located at a P.A. of $182.5^{\circ}\pm1.4^{\circ}$ and a separation of $600\pm22$~mas relative to the host star. The position is consistent with the recently identified protoplanet candidate AB Aur b. The source is visible in two individual epochs separated by ${\sim}50$ days and the H$α$ intensities in the two epochs agree. The H$α$ flux density is $F_ν=1.5\pm0.4$~mJy, $3.2\pm0.9$ times of the optical continuum determined by published HST/STIS photometry. In comparison to PDS 70 b and c, the H$α$ excess emission is weak. The central star is accreting and the stellar H$α$ emission has a similar line-to-continuum ratio as seen in AB Aur b. We conclude that both planetary accretion and scattered stellar light are possible sources of the H$α$ emission, and the H$α$ detection alone does not validate AB Aur b as an accreting protoplanet. Disentangling the origin of the emission will be crucial for probing planet formation in the AB Aur disk.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Learning Ball-balancing Robot Through Deep Reinforcement Learning

The ball-balancing robot (ballbot) is a good platform to test the effectiveness of a balancing controller. Considering balancing control, conventional model-based feedback control methods have been widely used. However, contacts and collisions are difficult to model, and often lead to failure in balancing control, especially when the ballbot tilts a large angle. To explore the maximum initial tilting angle of the ballbot, the balancing control is interpreted as a recovery task using Reinforcement Learning (RL). RL is a powerful technique for systems that are difficult to model, because it allows an agent to learn policy by interacting with the environment. In this paper, by combining the conventional feedback controller with the RL method, a compound controller is proposed. We show the effectiveness of the compound controller by training an agent to successfully perform a recovery task involving contacts and collisions. Simulation results demonstrate that using the compound controller, the ballbot can keep balance under a larger set of initial tilting angles, compared to the conventional model-based controller.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Spectral Variability of VHS J1256-1257 b from 1 to 5 $μ$m

Multi-wavelength time-resolved observations of rotationally modulated variability from brown dwarfs and giant exoplanets are the most effective method for constraining their heterogeneous atmospheric structures. In a companion paper (Bowler et al. 2020), we reported the discovery of strong near-infrared variability in HST/WFC3/G141 light curves of the very red L-dwarf companion VHS J1256-1257b. In this paper, we present a follow-up 36-hr Spitzer/IRAC Channel 2 light curve together with an in-depth analysis of the HST and the Spitzer data. The combined dataset provides time-resolved light curves of VHS1256b sampling 1.1 to 4.5 $μ$m. The Spitzer light curve is best-fit with a single sine wave with a period of $22.04\pm0.05$ hr and a peak-to-peak amplitude of $5.76\pm0.04$%. Combining the period with a previously measured projected rotational velocity ($v\sin i$), we find that VHS1256b is most consistent with equatorial viewing geometry. The HST/G141+Spitzer spectral energy distribution favors a $T_{\rm eff}$ of 1000~K, low surface gravity model with disequilibrium chemistry. The spectral variability of VHS1256b is consistent with predictions from partly cloudy models, suggesting heterogeneous clouds are the dominant source of the observed modulations. We find evidence at the $3.3σ$ level for amplitude variations within the 1.67$μ$m CH$_{4}$ band, which is the first such detection for a variable L-dwarf. We compare the HST/G141 time-resolved spectra of three red L-dwarfs with high-amplitude near-infrared rotational modulations and find that although their time-averaged spectra are similar, their spectroscopic variabilities exhibit notable differences. This diversity reinforces the advantage of time-resolved spectroscopic observations for understanding the atmospheres of brown dwarfs and directly-imaged exoplanets.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Sunbathing under white light -- 3D modelling of brown dwarf - white dwarf atmospheres with strong UV irradiation

The atmospheres of brown dwarfs orbiting in close proximity to their parent white dwarf represent some of the most extreme irradiated environments known. Understanding their complex dynamical mechanisms pushes the limits of theoretical and modelling efforts, making them valuable objets to study to test contemporary understanding of irradiated atmospheres. We use the Exo-FMS GCM to simulate the brown dwarfs WD0137-349B, SDSS J141126.20+200911.1B and EPIC212235321B, first coupled to a multi-banded grey radiative-transfer scheme then a spectral correlated-k scheme with high temperature opacity tables. We then post-process the GCM results using gCMCRT to compare to available observational data. Our GCM models predict strongly temperature inverted atmospheres, spanning many decades in pressure due to impact of UV band heating. Post-processing of our models suggest that the day-night contrast is too small in the GCM results. We therefore suggest that the formation of cloud particles as well as atmospheric drag effects such as magnetic drag are important considerations in setting the day-night temperature contrast for these objects.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Efficiently Imaging Accreting Protoplanets from Space: Reference Star Differential Imaging of the PDS 70 Planetary System using the HST/WFC3 Archival PSF Library

Accreting protoplanets provide key insights into how planets assemble from their natal protoplanetary disks. Recently, Zhou et al. (2021) used angular differential imaging (ADI) with Hubble Space Telescope's Wide Field Camera 3 (HST/WFC3) to recover the young accreting planet PDS 70 b in F656N ($\mathrm{H}α$) at a S/N of 7.9. In this paper, we demonstrate a promising approach to efficiently imaging accreting planets by applying reference star differential imaging (RDI) to the same dataset. We compile a reference library from the database of WFC3 point-spread functions (PSFs) provided by Space Telescope Science Institute and develop a set of morphology-significance criteria for pre-selection of reference frames to improve RDI subtraction. RDI with this PSF library results in a detection of PDS 70 b at a S/N of 5.3. Astrometry and photometry of PDS 70 b are calibrated using a forward-modeling method and injection-recovery tests, resulting in a separation of $186 \pm 13$ mas, a position angle of $142 \pm 5^\circ$, and an H$α$ flux of $(1.7 \pm 0.3)\times10^{-15}$ erg s$^{-1}$ cm$^{-2}$. The lower detection significance with RDI can be attributed to the $\sim$100 times lower peak-to-background ratios of the reference PSFs compared to the ADI PSFs. Building a high-quality reference library with WFC3 will provide unique opportunities to study accretion variability on short timescales not limited by roll angle scheduling constraints and efficiently search for actively accreting protoplanets in $\mathrm{H}α$ around targets inaccessible to ground-based adaptive optics systems, such as faint transition disk hosts.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

HST/WFC3 Complete Phase-resolved Spectroscopy of White Dwarf-Brown Dwarf Binaries WD 0137 and EPIC 2122

Brown dwarfs in close-in orbits around white dwarfs offer an excellent opportunity to investigate properties of fast-rotating, tidally-locked, and highly-irradiated atmospheres. We present Hubble Space Telescope Wide Field Camera 3 G141 phase-resolved observations of two brown dwarf-white dwarf binaries: WD 0137-349 and EPIC 212235321. Their 1.1 to 1.7 $μ$m phase curves demonstrate rotational modulations with semi-amplitudes of $5.27\pm0.02$% and $29.1\pm0.1$%; both can be well fit by multi-order Fourier series models. The high-order Fourier components have the same phase as the first order and are likely caused by hot spots located at the substellar points, suggesting inefficient day/night heat transfer. Both brown dwarfs' phase-resolved spectra can be accurately represented by linear combinations of their day- and night-side spectra. Fitting the irradiated brown dwarf model grids to the day-side spectra require a filling factor of ~50%, further supporting a hot spot dominating the emission of the day-sides. The night-side spectrum of WD 0137-349B is reasonably well fit by non-irradiated substellar models and the one of EPIC 212235321B can be approximated by a Planck function. We find strong spectral variations in the brown dwarfs' day/night flux and brightness temperature contrasts, which highlights the limitations of band-integrated measurements in probing heat transfer in irradiated objects. On the color-magnitude diagram, WD 0137-349B evolves along a cloudless model track connecting the early-L and mid-T spectral types, demonstrating that clouds and disequilibrium chemistry have a negligible effect on this object. A full interpretation of these high-quality phase-resolved spectra calls for new models that couple atmospheric circulation and radiative transfer under high-irradiation conditions.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Neuro-Reachability of Networked Microgrids

A neural ordinary differential equations network (ODE-Net)-enabled reachability method (Neuro-Reachability) is devised for the dynamic verification of networked microgrids (NMs) with unidentified subsystems and heterogeneous uncertainties. Three new contributions are presented: 1) An ODENet-enabled dynamic model discovery approach is devised to construct the data-driven state-space model which preserves the nonlinear and differential structure of the NMs system; 2) A physics-data-integrated (PDI) NMs model is established, which empowers various NM analytics; and 3) A conformance-empowered reachability analysis is developed to enhance the reliability of the PDI-driven dynamic verification. Extensive case studies demonstrate the efficacy of the ODE-Net-enabled method in microgrid dynamic model discovery, and the effectiveness of the Neuro-Reachability approach in verifying the NMs dynamics under multiple uncertainties and various operational scenarios.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Cloud Atlas: High-precision HST/WFC3/IR Time-Resolved Observations of Directly-Imaged Exoplanet HD106906b

HD106906b is an ~11$M_{\mathrm{Jup}}$, ~15Myr old directly-imaged exoplanet orbiting at an extremely large distance from its host star. The wide separation (7.11 arcsec) between HD106906b and its host star greatly reduces the difficulty in direct-imaging observations, making it one of the most favorable directly-imaged exoplanets for detailed characterization. In this paper, we present HST/WFC3/IR time-resolved observations of HD106906b in the F127M, F139M, and F153M bands. We have achieved ~1% precision in the lightcurves in all three bands. The F127M lightcurve demonstrates marginally-detectable ($2.7σ$ significance) variability with a best-fitting period of 4 hr, while the lightcurves in the other two bands are consistent with flat lines. We construct primary-subtracted deep images and use these images to exclude additional companions to HD106906 that are more massive than 4$M_{\mathrm{Jup}}$ and locate at projected distances of more than ~500 au. We measure the astrometry of HD106906b in two HST/WFC3 epochs and achieve precisions better than 2.5 mas. The position angle and separation measurements do not deviate from those in the 2004 HST/ACS/HRC images for more than $1σ$ uncertainty. We provide the HST/WFC3 astrometric results for 25 background stars that can be used as reference sources in future precision astrometry studies. Our observations also provide the first 1.4-micron water band photometric measurement for HD106906b. HD106906b's spectral energy distribution and the best-fitting BT-Settl model have an inconsistency in the 1.4-micron water absorption band, which highlights the challenges in modeling atmospheres of young planetary-mass objects.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Cluster-Adaptive Network A/B Testing: From Randomization to Estimation

A/B testing is an important decision-making tool in product development for evaluating user engagement or satisfaction from a new service, feature or product. The goal of A/B testing is to estimate the average treatment effects (ATE) of a new change, which becomes complicated when users are interacting. When the important assumption of A/B testing, the Stable Unit Treatment Value Assumption (SUTVA), which states that each individual's response is affected by their own treatment only, is not valid, the classical estimate of the ATE usually leads to a wrong conclusion. In this paper, we propose a cluster-adaptive network A/B testing procedure, which involves a sequential cluster-adaptive randomization and a cluster-adjusted estimator. The cluster-adaptive randomization is employed to minimize the cluster-level Mahalanobis distance within the two treatment groups, so that the variance of the estimate of the ATE can be reduced. In addition, the cluster-adjusted estimator is used to eliminate the bias caused by network interference, resulting in a consistent estimation for the ATE. Numerical studies suggest our cluster-adaptive network A/B testing achieves consistent estimation with higher efficiency. An empirical study is conducted based on a real world network to illustrate how our method can benefit decision-making in application.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Indications for very high metallicity and absence of methane for the eccentric exo-Saturn WASP-117b

We investigate the atmospheric composition of the long period ($P_{\rm orb}=$ 10 days), eccentric exo-Saturn WASP-117b. WASP-117b could be in atmospheric temperature and chemistry similar to WASP-107b. In mass and radius WASP-117b is similar to WASP-39b, which allows a comparative study of these planets. We analyze a near-infrared transmission spectrum of WASP-117b taken with Hubble Space Telescope/WFC3 G141, which was reduced with two independent pipelines. High resolution measurements were taken with VLT/ESPRESSO in the optical. We report the robust ($3σ$) detection of a water spectral feature. Using a 1D atmosphere model with isothermal temperature, uniform cloud deck and equilibrium chemistry, the Bayesian evidence of a retrieval analysis of the transmission spectrum indicates a preference for a high atmospheric metallicity ${\rm [Fe/H]}=2.58^{+0.26}_{-0.37}$ and clear skies. The data are also consistent with a lower-metallicity composition ${\rm [Fe/H]}<1.75$ and a cloud deck between $10^{-2.2} - 10^{-5.1}$ bar, but with weaker Bayesian preference. We retrieve a low CH$_4$ abundance of $<10^{-4}$ volume fraction within $1 σ$ and $<2\cdot 10^{-1}$ volume fraction within $3 σ$. We cannot constrain the equilibrium temperature between theoretically imposed limits of 700 and 1000~K. Further observations are needed to confirm quenching of CH$_4$ with $K_{zz}\geq 10^8$~cm$^2$/s. We report indications of Na and K in the VLT/ESPRESSO high resolution spectrum with substantial Bayesian evidence in combination with HST data.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Reliability Validation of Learning Enabled Vehicle Tracking

This paper studies the reliability of a real-world learning-enabled system, which conducts dynamic vehicle tracking based on a high-resolution wide-area motion imagery input. The system consists of multiple neural network components -- to process the imagery inputs -- and multiple symbolic (Kalman filter) components -- to analyse the processed information for vehicle tracking. It is known that neural networks suffer from adversarial examples, which make them lack robustness. However, it is unclear if and how the adversarial examples over learning components can affect the overall system-level reliability. By integrating a coverage-guided neural network testing tool, DeepConcolic, with the vehicle tracking system, we found that (1) the overall system can be resilient to some adversarial examples thanks to the existence of other components, and (2) the overall system presents an extra level of uncertainty which cannot be determined by analysing the deep learning components only. This research suggests the need for novel verification and validation methods for learning-enabled systems.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Robust Trajectory and Transmit Power Optimization for Secure UAV-Enabled Cognitive Radio Networks

Cognitive radio is a promising technology to improve spectral efficiency. However, the secure performance of a secondary network achieved by using physical layer security techniques is limited by its transmit power and channel fading. In order to tackle this issue, a cognitive unmanned aerial vehicle (UAV) communication network is studied by exploiting the high flexibility of a UAV and the possibility of establishing line-of-sight links. The average secrecy rate of the secondary network is maximized by robustly optimizing the UAV&#39;s trajectory and transmit power. Our problem formulation takes into account two practical inaccurate location estimation cases, namely, the worst case and the outage-constrained case. In order to solve those challenging non-convex problems, an iterative algorithm based on $\mathcal{S}$-Procedure is proposed for the worst case while an iterative algorithm based on Bernstein-type inequalities is proposed for the outage-constrained case. The proposed algorithms can obtain effective suboptimal solutions of the corresponding problems. Our simulation results demonstrate that the algorithm under the outage-constrained case can achieve a higher average secrecy rate with a low computational complexity compared to that of the algorithm under the worst case. Moreover, the proposed schemes can improve the secure communication performance significantly compared to other benchmark schemes.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Strong Near-Infrared Spectral Variability of the Young Cloudy L Dwarf Companion VHS J1256-1257 b

Rotationally-modulated variability of brown dwarfs and giant planets provides unique information about their surface brightness inhomogeneities, atmospheric circulation, cloud evolution, vertical atmospheric structure, and rotational angular momentum. We report results from Hubble Space Telescope/Wide Field Camera 3 near-infrared time-series spectroscopic observations of three companions with masses in or near the planetary regime: VHS J125601.92-125723.9 b, GSC 6214-210 B, and ROXs 42 B b. VHS J1256-1257 b exhibits strong total intensity and spectral variability with a brightness difference of 19.3% between 1.1-1.7 $μ$m over 8.5 hours and even higher variability at the 24.7% level at 1.27 $μ$m. The light curve of VHS J1256-1257 b continues to rise at the end of the observing sequence so these values represent lower limits on the full variability amplitude at this epoch. This observed variability rivals (and may surpass) the most variable brown dwarf currently known, 2MASS J21392676+0220226. The implied rotation period of VHS J1256-1257 b is $\approx$21-24 hr assuming sinusoidal modulations, which is unusually long for substellar objects. No significant variability is evident in the light curves of GSC 6214-210 B ($<$1.2%) and ROXs 42 B b ($<$15.6%). With a spectral type of L7, an especially red spectrum, and a young age, VHS J1256-1257 b reinforces emerging patterns between high variability amplitude, low surface gravity, and evolutionary phase near the L/T transition.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

Cloud Atlas: Weak color modulations due to rotation in the planetary-mass companion GU Psc b and 11 other brown dwarfs

Among the greatest challenges in understanding ultra-cool brown dwarf and exoplanet atmospheres is the evolution of cloud structure as a function of temperature and gravity. In this study, we present the rotational modulations of GU Psc b -- a rare mid-T spectral type planetary-mass companion at the end of the L/T spectral type transition. Based on the HST/WFC3 1.1-1.67$\rm\, μm$ time-series spectra, we observe a quasi-sinusoidal light curve with a peak-to-trough flux variation of 2.7 % and a minimum period of eight hours. The rotation-modulated spectral variations are weakly wavelength-dependent, or largely gray between 1.1-1.67$\rm\,μ$m. The gray modulations indicate that heterogeneous clouds are present in the photosphere of this low-gravity mid-T dwarf. We place the color and brightness variations of GU Psc b in the context of rotational modulations reported for mid-L to late-T dwarfs. Based on these observations, we report a tentative trend: mid-to-late T dwarfs become slightly redder in $J-H$ color with increasing $J$-band brightness, while L dwarfs become slightly bluer with increasing brightness. If this trend is verified with more T-dwarf samples, it suggests that in addition to the mostly gray modulations, there is a second-order spectral-type dependence on the nature of rotational modulations.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

Electronic and Magnetic Characterization of Epitaxial VSe$_2$ Monolayers on Superconducting NbSe$_2$

Vertical integration of two-dimensional (2D) van der Waals (vdW) materials with different quantum ground states is predicted to lead to novel electronic properties that are not found in the constituent layers. Here, we present the direct synthesis of superconductor-magnet hybrid heterostructures by combining superconducting niobium diselenide (NbSe$_2$) with the monolayer (ML) vanadium diselenide (VSe$_2$). More significantly, the in-situ growth in ultra-high vacuum (UHV) allows to produce a clean and an atomically sharp interfaces. Combining different characterization techniques and density-functional theory (DFT) calculations, we investigate the electronic and magnetic properties of VSe$_2$ on NbSe$_2$. Low temperature scanning tunneling microscopy (STM) measurements demonstrate a reduction of the superconducting gap on VSe$_2$ layer. This together with the lack of charge density wave signatures indicates magnetization of the sheet, but not of a conventional itinerant ferromagnet.

0 citations0 reviewsNo code linkedOpen access