arXiv subjects

Wei Zhu

Publications and source records attributed to Wei Zhu.

At least 73 records · Page 4Linked to original sources

arXiv2025

Four Giant Planets from 2024 KMTNet Microlensing Campaign

In this work, we present analyses of four newly discovered planetary microlensing events from the 2024 KMTNet survey season: KMT-2024-BLG-0176, KMT-2024-BLG-0349, KMT-2024-BLG-1870, and KMT-2024-BLG-2087. In each case, the planetary nature was revealed through distinct types of anomalies in the lensing light curves: a positive bump near the peak for KMT-2024-BLG-0176, an asymmetric peak for KMT-2024-BLG-0349, a short-duration central dip for KMT-2024-BLG-1870, and a caustic-crossing feature for KMT-2024-BLG-2087. Detailed modeling of the light curves confirms that these anomalies are produced by planetary companions with planet-to-host mass ratios in the range of $(1.5\text{--}17.9)\times 10^{-3}$. Despite the diversity of signal morphologies, all planets detected in these events are giant planets with masses comparable to or exceeding that of Jupiter in the Solar System. Each planet orbits a host star less massive than the Sun, emphasizing the strength of microlensing in uncovering planetary systems around low-mass stellar hosts.

astro-ph.EP↗

arXiv2025

Latent Nonlinear Denoising Score Matching for Enhanced Learning of Structured Distributions

We present latent nonlinear denoising score matching (LNDSM), a novel training objective for score-based generative models that integrates nonlinear forward dynamics with the VAE-based latent SGM framework. This combination is achieved by reformulating the cross-entropy term using the approximate Gaussian transition induced by the Euler-Maruyama scheme. To ensure numerical stability, we identify and remove two zero-mean but variance exploding terms arising from small time steps. Experiments on variants of the MNIST dataset demonstrate that the proposed method achieves faster synthesis and enhanced learning of inherently structured distributions. Compared to benchmark structure-agnostic latent SGMs, LNDSM consistently attains superior sample quality and variability.

stat.ML↗

arXiv2025

TRACE: Time SeRies PArameter EffiCient FinE-tuning

We propose an efficient fine-tuning method for time series foundation models, termed TRACE: Time Series Parameter Efficient Fine-tuning. While pretrained time series foundation models are gaining popularity, they face the following challenges: (1) Unlike natural language tasks, time series data vary in frequency, channel numbers, historical/prediction lengths. For long-term forecasting tasks in particular, tailored fine-tuning can significantly enhance performance.(2) Existing parameter-efficient tuning methods like LoRA remain applicable but require adaptation to temporal characteristics. To address these challenges, our TRACE framework introduces two key innovations: (1) Gated DSIC (Gated Dynamic Simulation Importance Calculation), an unbiased LoRA module importance selection mechanism that ensures conditional parameter consistency before and after masking. Experiments demonstrate that Gated DSIC outperforms common fine-tuning. (2) Reconstructed prediction heads for long-term forecasting tasks, which achieve comparable or superior performance to linear probing heads while drastically reducing parameter counts. Extensive experiments on long-/short-term forecasting, anomaly detection and natural language tasks across diverse datasets, coupled with ablation studies, validate the effectiveness of our method.

cs.LG↗

arXiv2025

Exciting Stellar Eccentricity in Gaia BH3 via a Hidden Black Hole Binary

We propose that the high eccentricity of the stellar orbit in Gaia BH3 system could be excited through a secular resonance effect if the inner dark object is, in effect, a tight and eccentric black hole binary (BHB). During the orbital decay of the inner BHB, the apsidal precession rate of the inner binary matches that of the outer stellar orbit, and this resonance advection can drive the outer eccentricity into some extreme values. For a Gaia BH3-like system, we show that a near equal-mass ($q=0.8$) BHB with an initial semi-major axis of 1--3 au and an initial eccentricity $\gtrsim 0.95$ is able to excite the outer orbit to the observed value, leaving a current BHB with semi-major axis 0.25--0.5 au and eccentricity $\sim 0.8$. The eccentric inner BHB imprints two observable signatures on the outer star: (1) short-term RV modulations with an amplitude $\lesssim 100$ m/s and (2) long-term apsidal precession with a rate $\lesssim 0.1^{\circ}$/yr. Although neither of these is detected in the currently available astrometry and RV data, we show that these signals are detectable with the full Gaia astrometry data and dedicated high-precision and/or long-term RV observations. Our work provides a new perspective on the dynamical formation of Gaia BH3, and the methodology is readily applicable to similar systems such as HD 130298, Gaia BH1, and Gaia BH2.

astro-ph.HE↗

arXiv2025

AMAS: Adaptively Determining Communication Topology for LLM-based Multi-Agent System

Although large language models (LLMs) have revolutionized natural language processing capabilities, their practical implementation as autonomous multi-agent systems (MAS) for industrial problem-solving encounters persistent barriers. Conventional MAS architectures are fundamentally restricted by inflexible, hand-crafted graph topologies that lack contextual responsiveness, resulting in diminished efficacy across varied academic and commercial workloads. To surmount these constraints, we introduce AMAS, a paradigm-shifting framework that redefines LLM-based MAS through a novel dynamic graph designer. This component autonomously identifies task-specific optimal graph configurations via lightweight LLM adaptation, eliminating the reliance on monolithic, universally applied structural templates. Instead, AMAS exploits the intrinsic properties of individual inputs to intelligently direct query trajectories through task-optimized agent pathways. Rigorous validation across question answering, mathematical deduction, and code generation benchmarks confirms that AMAS systematically exceeds state-of-the-art single-agent and multi-agent approaches across diverse LLM architectures. Our investigation establishes that context-sensitive structural adaptability constitutes a foundational requirement for high-performance LLM MAS deployments.

cs.CL↗

arXiv2025

FT-MDT: Extracting Decision Trees from Medical Texts via a Novel Low-rank Adaptation Method

Knowledge of the medical decision process, which can be modeled as medical decision trees (MDTs), is critical to building clinical decision support systems. However, current MDT construction methods rely heavily on time-consuming and laborious manual annotation. To address this challenge, we propose PI-LoRA (Path-Integrated LoRA), a novel low-rank adaptation method for automatically extracting MDTs from clinical guidelines and textbooks. We integrate gradient path information to capture synergistic effects between different modules, enabling more effective and reliable rank allocation. This framework ensures that the most critical modules receive appropriate rank allocations while less important ones are pruned, resulting in a more efficient and accurate model for extracting medical decision trees from clinical texts. Extensive experiments on medical guideline datasets demonstrate that our PI-LoRA method significantly outperforms existing parameter-efficient fine-tuning approaches for the Text2MDT task, achieving better accuracy with substantially reduced model complexity. The proposed method achieves state-of-the-art results while maintaining a lightweight architecture, making it particularly suitable for clinical decision support systems where computational resources may be limited.

cs.CL↗

arXiv2025

Unexpected Near-Resonant and Metastable States of Young Multi-Planet Systems

Recent observations suggest that the incidence of near-resonant planets declines as planetary systems age, making young planetary systems key signposts of early dynamical evolution. Here we investigate the dynamical states of three of the youngest multi-transiting planetary systems: AU Mic (3-planet, $\sim$20-Myr-old), V1298 Tau (4-planet, $\sim$23-Myr-old), and TOI-2076 (4-planet, $\sim$200-Myr-old). We find that most planet pairs in these systems lie near resonance with circulating rather than librating resonant angles. As a result, they are more susceptible to dynamical chaos than systems that are either securely locked in resonance or far removed from it. Even modest eccentricities of 0.04 to 0.08 may drive them to instability on timescales of tens to hundreds of Myr. Moreover, the observed orbital architectures are vulnerable to eccentricity excitation through mechanisms such as divergent resonance crossing triggered by planetesimal scattering. The observed near-resonant state may represent a transitional phase between a librating resonant chains and a mature non-resonant planetary system. Finally, we briefly discuss mechanisms that could give rise to the observed near-resonant configurations, including overstable libration, disk turbulence, and receding disk inner edge.

astro-ph.EP↗

arXiv2025

Data-driven Soliton Manifold Approximations for Dark and Bright Waves: Some Prototypical 1d Case Examples

In this paper, we revisit the investigation of solitary-wave interactions in the nonlinear Schrödinger model, both in the presence and absence of a parabolic trapping potential. While approximate dynamics, based on variational or similar methods, governed by a system of ordinary differential equations (ODEs) for both bright and dark-soliton interactions have been well established in the literature based on physical expert considerations, this study focuses on a data-driven approach, the so-called Sparse Identification of Nonlinear Dynamics (SINDy). Accordingly, our purpose is to use PDE time-series of select waveform diag- nostics in order to numerically reconstruct such approximate dynamics, without prior knowledge thereof. The purpose is not only to verify the robustness of the dynamical approximated ODEs, but also to shed light on the application of such a data-driven methodology in the study of soliton interactions and to formu- late a complementary approach, more reliant on the wealth of PDE data and less so on expert theoretical constructs.

nlin.PS↗

arXiv2025

Gluon Condensation as a Unifying Mechanism for Special Spectra of Cosmic Gamma Rays and Low-Momentum Pion Enhancement at the Large Hadron Collider

Decoding the internal structure of the proton is a fundamental challenge in physics. Historically, any new discovery about the proton has fuelled advances in several scientific fields. We have reported that gluons inside the proton accumulate near the critical momentum due to chaotic phenomena, forming gluon condensation. Surprisingly, the pion distribution predicted by this gluon distribution for the production of high-energy proton collisions could answer two puzzles in astronomy and high-energy physics. We find that during ultrahigh-energy cosmic ray collisions, gluon condensation may abruptly produce a large number of low-momentum pions, whose electromagnetic decays have the typical breakout properties appearing in various cosmic gamma-ray spectra. On the other hand, the Large Hadron Collider (LHC), which is well below the cosmic ray energy scale, also shows weak but recognisable signs of gluon condensation, which had been mistaken for BEC pions. The connection between these two phenomena, which occur at different scales in the Universe, supports the existence of a new structure within the proton-gluon condensation.

hep-ph↗

arXiv2025

Early Stellar Flybys are Unlikely: Improved Constraints from Sednoids and Large-$q$ TNOs

Sedna-like objects (a.k.a. sednoids) are transneptunian objects (TNOs) characterized by large semimajor axes and exceptionally high perihelia. Their high-$q$ orbits are detached from the influence of the four giant planets and need extra perturbation to form. One hypothesis posits that close stellar flybys could have perturbed objects from the primordial scattering disk, generating the sednoid population. In this study, we run N-body simulations with different stellar encounter configurations to explore whether such a close stellar flyby can satisfy new constraints identified from sednoid (and detached extreme TNO) observation, including the low-inclination ($i<30^\circ$) profile and primordial orbital alignment. Our results suggest that flybys with field stars are unable to generate a sufficient population, whereas flybys within the birth cluster fail to produce the primordial orbital alignment. To meet the inclination constraint of detached extreme TNOs, flybys have to be either coplanar ($i_\star \sim 0^\circ$) or symmetric about the ecliptic plane ($ω_\star \sim 0^\circ, i_\star \sim 90^\circ$). After taking into account their occurrence rate at the early stage of the Solar System, we conclude that close-in stellar flybys ($q_\star \le 1000$~au) that satisfy all constraints are unlikely to happen ($\lesssim$5\%). Future discoveries of additional sednoids with precise orbital determinations are crucial to confirm the existence of the low-inclination tendency and the primordial alignment, and to further constrain the early dynamical evolution of the Solar System.

astro-ph.EP↗

arXiv2025

Statistical Learning Guarantees for Group-Invariant Barron Functions

We investigate the generalization error of group-invariant neural networks within the Barron framework. Our analysis shows that incorporating group-invariant structures introduces a group-dependent factor $δ_{G,Γ,σ} \le 1$ into the approximation rate. When this factor is small, group invariance yields substantial improvements in approximation accuracy. On the estimation side, we establish that the Rademacher complexity of the group-invariant class is no larger than that of the non-invariant counterpart, implying that the estimation error remains unaffected by the incorporation of symmetry. Consequently, the generalization error can improve significantly when learning functions with inherent group symmetries. We further provide illustrative examples demonstrating both favorable cases, where $δ_{G,Γ,σ}\approx |G|^{-1}$, and unfavorable ones, where $δ_{G,Γ,σ}\approx 1$. Overall, our results offer a rigorous theoretical foundation showing that encoding group-invariant structures in neural networks leads to clear statistical advantages for symmetric target functions.

cs.LG↗

arXiv2025

A Comprehensive Analysis of Three Microlensing Planet Candidates with the Planet/Binary Degeneracy

We present observations and analyses of three high-magnification microlensing events: KMT-2022-BLG-0954, KMT-2024-BLG-0697, and MOA-2024-BLG-018. All three exhibit the "Planet/Binary" degeneracy, with planetary solutions corresponding to mass ratios in the range $-3.7 < \log q < -2.2$, while the binary solutions yield $\log q > -2.0$. For KMT-2022-BLG-0954, we identify a previously unrecognized degeneracy among planetary solutions, involving different mass ratios and normalized source radii. In all three cases, single-lens binary-source models are excluded. Bayesian analyses suggest that the planetary solutions correspond to gas giants orbiting M/K dwarfs beyond the snow line, while KMT-2022-BLG-0954 also admits an alternative interpretation as a super-Earth orbiting a late-type M dwarf. The binary solutions imply a diverse set of systems, including M-dwarf pairs and M-dwarf--brown-dwarf binaries. A review of known events subject to the "Planet/Binary" degeneracy shows that in most cases the degeneracy cannot be resolved through follow-up high-resolution imaging, particularly in the presence of the newly identified degeneracy.

astro-ph.EP↗

arXiv2025

Robust Generative Learning with Lipschitz-Regularized $α$-Divergences Allows Minimal Assumptions on Target Distributions

This paper demonstrates the robustness of Lipschitz-regularized $α$-divergences as objective functionals in generative modeling, showing they enable stable learning across a wide range of target distributions with minimal assumptions. We establish that these divergences remain finite under a mild condition-that the source distribution has a finite first moment-regardless of the properties of the target distribution, making them adaptable to the structure of target distributions. Furthermore, we prove the existence and finiteness of their variational derivatives, which are essential for stable training of generative models such as GANs and gradient flows. For heavy-tailed targets, we derive necessary and sufficient conditions that connect data dimension, $α$, and tail behavior to divergence finiteness, that also provide insights into the selection of suitable $α$'s. We also provide the first sample complexity bounds for empirical estimations of these divergences on unbounded domains. As a byproduct, we obtain the first sample complexity bounds for empirical estimations of these divergences and the Wasserstein-1 metric with group symmetry on unbounded domains. Numerical experiments confirm that generative models leveraging Lipschitz-regularized $α$-divergences can stably learn distributions in various challenging scenarios, including those with heavy tails or complex, low-dimensional, or fractal support, all without any prior knowledge of the structure of target distributions.

stat.ML↗

arXiv2025

Ensemble of Precision-Recall Curve (PRC) Classification Trees with Autoencoders

Anomaly detection underpins critical applications from network security and intrusion detection to fraud prevention, where recognizing aberrant patterns rapidly is indispensable. Progress in this area is routinely impeded by two obstacles: extreme class imbalance and the curse of dimensionality. To combat the former, we previously introduced Precision-Recall Curve (PRC) classification trees and their ensemble extension, the PRC Random Forest (PRC-RF). Building on that foundation, we now propose a hybrid framework that integrates PRC-RF with autoencoders, unsupervised machine learning methods that learn compact latent representations, to confront both challenges simultaneously. Extensive experiments across diverse benchmark datasets demonstrate that the resulting Autoencoder-PRC-RF model achieves superior accuracy, scalability, and interpretability relative to prior methods, affirming its potential for high-stakes anomaly-detection tasks.

cs.LG↗

arXiv2025

Learn to Teach: Sample-Efficient Privileged Learning for Humanoid Locomotion over Diverse Terrains

Humanoid robots promise transformative capabilities for industrial and service applications. While recent advances in Reinforcement Learning (RL) yield impressive results in locomotion, manipulation, and navigation, the proposed methods typically require enormous simulation samples to account for real-world variability. This work proposes a novel one-stage training framework-Learn to Teach (L2T)-which unifies teacher and student policy learning. Our approach recycles simulator samples and synchronizes the learning trajectories through shared dynamics, significantly reducing sample complexities and training time while achieving state-of-the-art performance. Furthermore, we validate the RL variant (L2T-RL) through extensive simulations and hardware tests on the Digit robot, demonstrating zero-shot sim-to-real transfer and robust performance over 12+ challenging terrains without depth estimation modules.

cs.RO↗

arXiv2025

CorrMoE: Mixture of Experts with De-stylization Learning for Cross-Scene and Cross-Domain Correspondence Pruning

Establishing reliable correspondences between image pairs is a fundamental task in computer vision, underpinning applications such as 3D reconstruction and visual localization. Although recent methods have made progress in pruning outliers from dense correspondence sets, they often hypothesize consistent visual domains and overlook the challenges posed by diverse scene structures. In this paper, we propose CorrMoE, a novel correspondence pruning framework that enhances robustness under cross-domain and cross-scene variations. To address domain shift, we introduce a De-stylization Dual Branch, performing style mixing on both implicit and explicit graph features to mitigate the adverse influence of domain-specific representations. For scene diversity, we design a Bi-Fusion Mixture of Experts module that adaptively integrates multi-perspective features through linear-complexity attention and dynamic expert routing. Extensive experiments on benchmark datasets demonstrate that CorrMoE achieves superior accuracy and generalization compared to state-of-the-art methods. The code and pre-trained models are available at https://github.com/peiwenxia/CorrMoE.

cs.CV↗

arXiv2025

Nonlinear denoising score matching for enhanced learning of structured distributions

We present a novel method for training score-based generative models which uses nonlinear noising dynamics to improve learning of structured distributions. Generalizing to a nonlinear drift allows for additional structure to be incorporated into the dynamics, thus making the training better adapted to the data, e.g., in the case of multimodality or (approximate) symmetries. Such structure can be obtained from the data by an inexpensive preprocessing step. The nonlinear dynamics introduces new challenges into training which we address in two ways: 1) we develop a new nonlinear denoising score matching (NDSM) method, 2) we introduce neural control variates in order to reduce the variance of the NDSM training objective. We demonstrate the effectiveness of this method on several examples: a) a collection of low-dimensional examples, motivated by clustering in latent space, b) high-dimensional images, addressing issues with mode imbalance, small training sets, and approximate symmetries, the latter being a challenge for methods based on equivariant neural networks, which require exact symmetries, c) latent space representation of high-dimensional data, demonstrating improved performance with greatly reduced computational cost. Our method learns score-based generative models with less data by flexibly incorporating structure arising in the dataset.

stat.ML↗

arXiv2025

DynamicFace: High-Quality and Consistent Face Swapping for Image and Video using Composable 3D Facial Priors

Face swapping transfers the identity of a source face to a target face while retaining the attributes like expression, pose, hair, and background of the target face. Advanced face swapping methods have achieved attractive results. However, these methods often inadvertently transfer identity information from the target face, compromising expression-related details and accurate identity. We propose a novel method DynamicFace that leverages the power of diffusion models and plug-and-play adaptive attention layers for image and video face swapping. First, we introduce four fine-grained facial conditions using 3D facial priors. All conditions are designed to be disentangled from each other for precise and unique control. Then, we adopt Face Former and ReferenceNet for high-level and detailed identity injection. Through experiments on the FF++ dataset, we demonstrate that our method achieves state-of-the-art results in face swapping, showcasing superior image quality, identity preservation, and expression accuracy. Our framework seamlessly adapts to both image and video domains. Our code and results will be available on the project page: https://dynamic-face.github.io/

cs.CV↗