arXiv subjects

Dong Zheng

Publications and source records attributed to Dong Zheng.

At least 19 recordsLinked to original sources

arXiv2026

A framework for modeling aerosol-cloud-lightning interactions: Validation of charge structure and aerosol effects

This study develops a novel framework within the Weather Research and Forecast Model for modeling aerosol-cloud-lightning interactions. The framework explicitly represents aerosol-cloud interactions by prescribing aerosols with two configurations: an idealized setup, where both cloud condensation nuclei (CCN) and ice nucleating particles (IN) are assumed to have a single chemical composition and spatially uniform distributions; and a quasi-realistic configuration, with multi-species aerosols assigned spatially varying distributions, where hygroscopic components act as CCN, dust particles act as IN, and all aerosol species influence radiative transfer. Cloud microphysics is coupled with detailed charge separation and discharge processes to enable the lightning simulation. The framework is evaluated using two thunderstorms in Guangdong, China. For an isolated storm, the model successfully reproduces the observed tripolar charge structure (positive-negative-positive), demonstrating its capability in simulating cloud electrification. For a frontal storm, it captures well the observed precipitation and lightning, and shows that increasing CCN suppresses the rainfall while enhancing the lightning. Higher CCN concentrations produce more numerous but smaller cloud droplets, which suppresses the coalescence into rain droplets, allows a greater number of droplets to loft into the upper troposphere, and forms more but smaller cloud ice particles. This boosts graupel-ice collisions, intensifies non-inductive charging, strengthens the upper positive charge and the vertical electric-field gradient, ultimately increasing the lightning frequency. In contrast, no significant aerosol-induced invigoration of updrafts is observed. These results highlight the dominant role of aerosol microphysical effects over dynamical invigoration in modulating thunderstorm electrification and lightning activity.

physics.ao-ph↗

arXiv2026

High-energy neutrino emission from the Type~IIn supernova SN~2017hcd

Neutrino astronomy provides another window to exploring the Universe, exemplified by the detection of a megaelectronvolt neutrino burst from the core-collapse supernova (CCSN) SN~1987A (refs.~\citenum{hir+87,bio+87}). Commonly discussed theories suggest that some CCSNe could produce neutrinos with energies a thousand times more than those of SN~1987A \cite{tm18}, which has been probed with new-generation facilities \cite{abb+12,aar+15,abb+23}. The interaction of SN ejecta with a dense circumstellar medium (CSM) or a jet, launched in a CCSN, being choked in the stellar envelope of the progenitor or an outside CSM are both well-accepted scenarios for the high-energy neutrino production. Here we report the detection of a high-energy neutrino flare at a 3.9$σ$ significance from SN~2017hcd, made by our analysis of the public track-like neutrino data taken by the IceCube Neutrino Observatory \cite{IceCube17}. A Type IIn SN with optical emissions arising from the ejecta--CSM interaction, SN~2017hcd's neutrino flare lasted $\sim$1--2 month, with its central time $\sim$14-day prior to the SN's optical discovery time. Its estimated isotropic neutrino energy (all flavors) is approximately two orders of magnitude higher than the energy ($\sim 10^{50}$\,erg) carried in the SN's ejecta, too high to be explained with the ejecta--CSM scenario. Thus, a choked jet may be the source of the neutrino flare.

astro-ph.HE↗

arXiv2025

Event-Customized Image Generation

Customized Image Generation, generating customized images with user-specified concepts, has raised significant attention due to its creativity and novelty. With impressive progress achieved in subject customization, some pioneer works further explored the customization of action and interaction beyond entity (i.e., human, animal, and object) appearance. However, these approaches only focus on basic actions and interactions between two entities, and their effects are limited by insufficient ''exactly same'' reference images. To extend customized image generation to more complex scenes for general real-world applications, we propose a new task: event-customized image generation. Given a single reference image, we define the ''event'' as all specific actions, poses, relations, or interactions between different entities in the scene. This task aims at accurately capturing the complex event and generating customized images with various target entities. To solve this task, we proposed a novel training-free event customization method: FreeEvent. Specifically, FreeEvent introduces two extra paths alongside the general diffusion denoising process: 1) Entity switching path: it applies cross-attention guidance and regulation for target entity generation. 2) Event transferring path: it injects the spatial feature and self-attention maps from the reference image to the target image for event generation. To further facilitate this new task, we collected two evaluation benchmarks: SWiG-Event and Real-Event. Extensive experiments and ablations have demonstrated the effectiveness of FreeEvent.

cs.CV↗

arXiv2025

ALTo: Adaptive-Length Tokenizer for Autoregressive Mask Generation

While humans effortlessly draw visual objects and shapes by adaptively allocating attention based on their complexity, existing multimodal large language models (MLLMs) remain constrained by rigid token representations. Bridging this gap, we propose ALTo, an adaptive length tokenizer for autoregressive mask generation. To achieve this, a novel token length predictor is designed, along with a length regularization term and a differentiable token chunking strategy. We further build ALToLLM that seamlessly integrates ALTo into MLLM. Preferences on the trade-offs between mask quality and efficiency is implemented by group relative policy optimization (GRPO). Experiments demonstrate that ALToLLM achieves state-of-the-art performance with adaptive token cost on popular segmentation benchmarks. Code and models are released at https://github.com/yayafengzi/ALToLLM.

cs.CV↗

arXiv2025

A Disk-Originated 329-day Quasi-Periodic Oscillation in the Seyfert 1 Galaxy J1626+5120

The Seyfert 1 galaxy J1626+5120 is estimated to host a $10^8 M_{\odot}$ black hole (BH) accreting at Eddington ratio $\dot{m}_{\text{Edd}} \approx 0.043$. Its long-term multi-band light curve data show flicker-like variations, but in a well-sampled $g$-band light curve, we are able to determine a $\simeq 329$\,d quasi-periodic oscillation (QPO) at a $\sim$4.53$σ$ significance. Six optical spectra were obtained for the source, three of which were taken by us. The spectra show that the variations were mainly because of flux changes blueward of 4000\,Å. We also analyze X-ray and ultraviolet (UV) data obtained with {\it the Neil Gehrels Swift Observatory (Swift)}, which targeted the source in the past two years. X-ray and UV emissions of the source show variations correlated with optical. Time lags of four UV bands and four optical bands are determined with respect to the X-ray emission, which are consistent with a continuum reprocessing disk model. These properties point out a disk origin for the QPO, likely due to Lense-Thirring (LT) precession of the accretion flow at $\sim$20 gravitational radii of the BH. This QPO could be a key case linking sub-year long QPOs in jets, which have more cases reported, to LT precession.

astro-ph.HE↗

arXiv2025

Possible Neutrino Emission from the Pulsar Wind Nebula G63.7+1.1

We report on our finding of an excess of $54^{+16}_{-15}$ neutrinos at the location of the pulsar wind nebula (PWN) G63.7+1.1. By analyzing the IceCube track-like neutrino data for a group of 14 PWNe, which are selected as the targets because of their reportedly association with molecular clouds, G63.7+1.1 is found to be the only one detected with neutrino emission and the post-trail significance for the detection is 3.2$σ$. Previously, this PWN was estimated to have an age of $\gtrsim$8\,kyr, contain a candidate pulsar detected in X-rays, and have a distance of $\sim$6\,kpc. More importantly and related to the PWN's possible neutrino emission, surrounding molecular materials are seen to interact with the PWN. On the basis of these properties, we examine the proton-proton interactions as the process for the neutrino production. The PWN (or the pulsar) can collectively provide sufficient energy to power the required high-energy (HE) protons. This possibly first neutrino-emitting case in our Galaxy, with problems or other possibilities to be solved or examined, may reveal to us that PWNe are the significant Galactic HE neutrino sources.

astro-ph.HE↗

arXiv2025

WanJuanSiLu: A High-Quality Open-Source Webtext Dataset for Low-Resource Languages

This paper introduces the open-source dataset WanJuanSiLu, designed to provide high-quality training corpora for low-resource languages, thereby advancing the research and development of multilingual models. To achieve this, we have developed a systematic data processing framework tailored for low-resource languages. This framework encompasses key stages such as data extraction, corpus cleaning, content deduplication, security filtering, quality evaluation, and theme classification. Through the implementation of this framework, we have significantly improved both the quality and security of the dataset, while maintaining its linguistic diversity. As of now, data for all five languages have been fully open-sourced. The dataset can be accessed at https://opendatalab.com/applyMultilingualCorpus, and GitHub repository is available at https://github.com/opendatalab/WanJuan3.0

cs.CL↗

arXiv2024

PKS~2332$-$017 and PMN J1916$-$1519: Candidate Blazar Counterparts to Two High-energy Neutrino Events

We report our counterpart identification study for two high-energy neutrino events IC-130127A and IC-131204A listed in the IceCube Event Catalog of Alert Tracks. These two events belong to Gold alerts, which have a significant probability of being of astrophysical origin.Within the events' 90\% positional uncertainty regions, we respectively find PKS~2332$-$017 and PMN J1916$-$1519. The first source is a flat-spectrum radio quasar at redshift $z= 1.18$ and the second a blazar of an uncertain type with photometric $z= 0.968$. As they correspondingly had a $γ$-ray flare temporally coincident with the arrival times of IC-130127A and IC-131204A, we identify them as the respective neutrino emitters. Detailed analysis of the $γ$-ray data for the two blazars, obtained with the Large Area Telescope (LAT) onboard {\it the Fermi Gamma-ray Space Telescope (Fermi)}, is conducted. The two flares respectively from PKS~2332$-$017 and PMN~J1916$-$1519 lasted $\sim$4\,yr and $\sim$4\,month, and showed possible emission hardening by containing high-energy $\sim$2--10\,GeV photons in the emissions. Accompanying the flare of PKS~2332$-$017, optical and MIR brightening variations were also observed. We discuss the properties of the two sources and compare the properties with those of the previously reported (candidate) neutrino-emitting blazars.

astro-ph.HE↗

arXiv2024

Revisiting Gamma-Ray Emission of the Supernova Remnant RCW 103

We analyze more than 15 years of the \gr\ data, obtained with the Large Area Telescope (LAT) onboard {\it the Fermi Gamma-ray Space Telescope (Fermi)}, for the region of the young supernova remnant (SNR) RCW~103, since the nearby source 4FGL J1616.2$-$5054e, counterpart to HESS~J1616$-$518 and $\simeq$13\,arcmin away from the SNR, is determined to be extended in the more recent Fermi-LAT source catalog. Different templates for 4FGL J1616.2$-$5054e and RCW~103 are tested, and we find that a point source with a power-law (PL) spectrum at the southern limb of the SNR best describes the detected gamma-ray emission. The photon index of the PL emission is $Γ\simeq 2.31$(or $α\simeq 2.4$ in a Log-Parabola model) , softer than the previously reported $Γ\simeq 2.0$ when the counterpart to HESS~J1616$-$518 was considered to be a point source (which likely caused mis-identification of extended emission at RCW~103). In order to produce the \gr\ emission in a hadronic scenario, we estimate that protons with an index$\sim$2.4 PL energy distribution are needed. These results fit with those from multi-wavelength observations that have indicated the remnant at the southern limb is interacting with a molecular cloud.

astro-ph.HE↗

arXiv2024

Finding Candidate TeV Halos among Very-High Energy Sources

We search for possible pulsar TeV halos among the very-high-energy (VHE) sources reported in different VHE surveys, among which in particular we use the results from the first Large High Altitude Air Shower Observatory (LHAASO) catalog of $γ$-ray sources. Six candidates are found. They share similar properties of containing a middle-aged, gamma-ray--bright pulsar in their positional error circles (the respective pulsars are J0248+6021, J0359+5414, J0622+3749, J0633+0632, J2006+3102, and J2238+5903), being in a rather clean field without any common Galactic VHE-emitting supernova remnants or (bright) pulsar wind nebulae (PWNe), and showing an absence of any gamma-ray emissions in 0.1--500\,GeV after removing the pulsars' emissions. Combining these candidates with several reported (candidate) TeV halos, we obtain the relationships between their luminosity at 50\,TeV ($L_{\rm 50TeV}$) and the corresponding pulsars' spin-down energy ($\dot{E}$), which are $L_{\rm 50TeV}\sim \dot{E}^{0.9}$ and $L_{\rm 50TeV}/\dot{E}\sim 6.4\times 10^{-4}$. The relationships are nearly identical to previously reported ones. We probe possible connections between the extension sizes of the VHE sources and the pulsars' ages, and find a weak older-and-smaller trend. By comparing to the VHE detection results for PWNe, it is clear that the (candidate) TeV halos have hard emissions by either having their power-law indices be smaller than 2 in 1--25\,TeV or by only being detected in 25--100\,TeV. In addition, we also consider seven other VHE sources as possible TeV halos based on the results from different studies of them, but they do not fit cleanly with the properties listed above, indicating their potentially complex nature.

astro-ph.HE↗

arXiv2024

Covert Communication for Untrusted UAV-Assisted Wireless Systems

Wireless systems are of paramount importance for providing ubiquitous data transmission for smart cities. However, due to the broadcasting and openness of wireless channels, such systems face potential security challenges. UAV-assisted covert communication is a supporting technology for improving covert performances and has become a hot issue in the research of wireless communication security. This paper investigates the performance of joint covert and security communication in a tow-hop UAV-assisted wireless system, where a source transmits the covert message to a destination with the help of an untrusted UAV. We first design a transmission scheme such that use UAVs to assist in covert communications while ensuring the security of covert messages. Then, we develop a theoretical model to derive the expressions for the detection error probability of the warden and the covert and security rate, and the maximum covert and security rate is optimized by power control under a given covertness and security requirements. Finally, numerical results are provided to illustrate our theoretical analysis and the performance of covert and security communication in such systems.

cs.CR↗

arXiv2024

State Regularized Policy Optimization on Data with Dynamics Shift

In many real-world scenarios, Reinforcement Learning (RL) algorithms are trained on data with dynamics shift, i.e., with different underlying environment dynamics. A majority of current methods address such issue by training context encoders to identify environment parameters. Data with dynamics shift are separated according to their environment parameters to train the corresponding policy. However, these methods can be sample inefficient as data are used \textit{ad hoc}, and policies trained for one dynamics cannot benefit from data collected in all other environments with different dynamics. In this paper, we find that in many environments with similar structures and different dynamics, optimal policies have similar stationary state distributions. We exploit such property and learn the stationary state distribution from data with dynamics shift for efficient data reuse. Such distribution is used to regularize the policy trained in a new environment, leading to the SRPO (\textbf{S}tate \textbf{R}egularized \textbf{P}olicy \textbf{O}ptimization) algorithm. To conduct theoretical analyses, the intuition of similar environment structures is characterized by the notion of homomorphous MDPs. We then demonstrate a lower-bound performance guarantee on policies regularized by the stationary state distribution. In practice, SRPO can be an add-on module to context-based algorithms in both online and offline RL settings. Experimental results show that SRPO can make several context-based algorithms far more data efficient and significantly improve their overall performance.

cs.LG↗

arXiv2024

Two-Stage Constrained Actor-Critic for Short Video Recommendation

The wide popularity of short videos on social media poses new opportunities and challenges to optimize recommender systems on the video-sharing platforms. Users sequentially interact with the system and provide complex and multi-faceted responses, including watch time and various types of interactions with multiple videos. One the one hand, the platforms aims at optimizing the users' cumulative watch time (main goal) in long term, which can be effectively optimized by Reinforcement Learning. On the other hand, the platforms also needs to satisfy the constraint of accommodating the responses of multiple user interactions (auxiliary goals) such like, follow, share etc. In this paper, we formulate the problem of short video recommendation as a Constrained Markov Decision Process (CMDP). We find that traditional constrained reinforcement learning algorithms can not work well in this setting. We propose a novel two-stage constrained actor-critic method: At stage one, we learn individual policies to optimize each auxiliary signal. At stage two, we learn a policy to (i) optimize the main signal and (ii) stay close to policies learned at the first stage, which effectively guarantees the performance of this main policy on the auxiliaries. Through extensive offline evaluations, we demonstrate effectiveness of our method over alternatives in both optimizing the main goal as well as balancing the others. We further show the advantage of our method in live experiments of short video recommendations, where it significantly outperforms other baselines in terms of both watch time and interactions. Our approach has been fully launched in the production system to optimize user experiences on the platform.

cs.LG↗

arXiv2023

KuaiSim: A Comprehensive Simulator for Recommender Systems

Reinforcement Learning (RL)-based recommender systems (RSs) have garnered considerable attention due to their ability to learn optimal recommendation policies and maximize long-term user rewards. However, deploying RL models directly in online environments and generating authentic data through A/B tests can pose challenges and require substantial resources. Simulators offer an alternative approach by providing training and evaluation environments for RS models, reducing reliance on real-world data. Existing simulators have shown promising results but also have limitations such as simplified user feedback, lacking consistency with real-world data, the challenge of simulator evaluation, and difficulties in migration and expansion across RSs. To address these challenges, we propose KuaiSim, a comprehensive user environment that provides user feedback with multi-behavior and cross-session responses. The resulting simulator can support three levels of recommendation problems: the request level list-wise recommendation task, the whole-session level sequential recommendation task, and the cross-session level retention optimization task. For each task, KuaiSim also provides evaluation protocols and baseline recommendation algorithms that further serve as benchmarks for future research. We also restructure existing competitive simulators on the KuaiRand Dataset and compare them against KuaiSim to future assess their performance and behavioral differences. Furthermore, to showcase KuaiSim's flexibility in accommodating different datasets, we demonstrate its versatility and robustness when deploying it on the ML-1m dataset.

cs.IR↗

arXiv2023

Two Candidate Pulsar TeV Halos Identified from Property-Similarity Studies

TeV halos have been suggested as a common phenomenon associated with middle-aged pulsars. Based on our recent work on PSR~J0631+1036, which is the only known source positionally coincident with a hard TeV gamma-ray source and likely powers the latter as a TeV halo, we select 3 candidate TeV halos from the first Large High Altitude Air Shower Observatory (LHAASO) catalog of gamma-ray sources. The corresponding pulsars, given by the positional coincidences and property similarities, are PSR J1958+2846, PSR J2028+3332, and PSR J1849$-$0001. We analyze the GeV $γ$-ray data obtained with the Large Area Telescope (LAT) onboard {\it the Fermi Gamma-ray Space Telescope} for the first two pulsars, as the last is gamma-ray quiet. We remove the pulsed emissions of the pulsars from the source regions from timing analysis, and determine that there are no residual GeV emissions in the regions as any possible counterparts to the TeV sources. Considering the previous observational results for the source regions and comparing the two pulsars to Geminga (and Monogem), the LHAASO-detected TeV sources are likely the pulsars' respective TeV halos. We find that the candidate and identified TeV halos, including that of PSR~J1849$-$0001, have luminosites at 50 TeV (estimated from the differential fluxes) approximately proportional to the spin-down energy $\dot{E}$ of the pulsars, and the ratios of the former to the latter are $\sim 6\times 10^{-4}$.

astro-ph.HE↗

arXiv2023

3HWC J0631+107/LHAASO J0631+1040: a TeV halo powered by the pulsar J0631+1036?

PSR~J0631+1036 is a middle-aged pulsar with properties similar to those of the nearby Geminga pulsar. It is bright in $γ$-rays, and has been noted as the only source possibly associated with the TeV source 3HWC J0631+107 (also the LHAASO J0631+1040). For understanding the nature of the TeV source, we analyze the GeV $γ$-ray data obtained with the Large Area Telescope (LAT) onboard {\it the Fermi Gamma-ray Space Telescope} for the source region. We are able to remove the pulsar's emission from the region from timing analysis, and find that the region is rather clean without possible GeV $γ$-ray emission present as the counterpart to the TeV source. By comparing this pulsar to Geminga and considering the spectral feature of the TeV source, we argue that it is likely the TeV halo powered by the pulsar.

astro-ph.HE↗

arXiv2023

ResAct: Reinforcing Long-term Engagement in Sequential Recommendation with Residual Actor

Long-term engagement is preferred over immediate engagement in sequential recommendation as it directly affects product operational metrics such as daily active users (DAUs) and dwell time. Meanwhile, reinforcement learning (RL) is widely regarded as a promising framework for optimizing long-term engagement in sequential recommendation. However, due to expensive online interactions, it is very difficult for RL algorithms to perform state-action value estimation, exploration and feature extraction when optimizing long-term engagement. In this paper, we propose ResAct which seeks a policy that is close to, but better than, the online-serving policy. In this way, we can collect sufficient data near the learned policy so that state-action values can be properly estimated, and there is no need to perform online exploration. ResAct optimizes the policy by first reconstructing the online behaviors and then improving it via a Residual Actor. To extract long-term information, ResAct utilizes two information-theoretical regularizers to confirm the expressiveness and conciseness of features. We conduct experiments on a benchmark dataset and a large-scale industrial dataset which consists of tens of millions of recommendation requests. Experimental results show that our method significantly outperforms the state-of-the-art baselines in various long-term engagement optimization tasks.

cs.IR↗

arXiv2023

Generative Flow Network for Listwise Recommendation

Personalized recommender systems fulfill the daily demands of customers and boost online businesses. The goal is to learn a policy that can generate a list of items that matches the user's demand or interest. While most existing methods learn a pointwise scoring model that predicts the ranking score of each individual item, recent research shows that the listwise approach can further improve the recommendation quality by modeling the intra-list correlations of items that are exposed together. This has motivated the recent list reranking and generative recommendation approaches that optimize the overall utility of the entire list. However, it is challenging to explore the combinatorial space of list actions and existing methods that use cross-entropy loss may suffer from low diversity issues. In this work, we aim to learn a policy that can generate sufficiently diverse item lists for users while maintaining high recommendation quality. The proposed solution, GFN4Rec, is a generative method that takes the insight of the flow network to ensure the alignment between list generation probability and its reward. The key advantages of our solution are the log scale reward matching loss that intrinsically improves the generation diversity and the autoregressive item selection model that captures the item mutual influences while capturing future reward of the list. As validation of our method's effectiveness and its superior diversity during active exploration, we conduct experiments on simulated online environments as well as an offline evaluation framework for two real-world datasets.

cs.IR↗