arXiv subjects

Xiaowei Zhang

Publications and source records attributed to Xiaowei Zhang.

At least 37 records · Page 2Linked to original sources

arXiv2025

AgentGit: A Version Control Framework for Reliable and Scalable LLM-Powered Multi-Agent Systems

With the rapid progress of large language models (LLMs), LLM-powered multi-agent systems (MAS) are drawing increasing interest across academia and industry. However, many current MAS frameworks struggle with reliability and scalability, especially on complex tasks. We present AgentGit, a framework that brings Git-like rollback and branching to MAS workflows. Built as an infrastructure layer on top of LangGraph, AgentGit supports state commit, revert, and branching, allowing agents to traverse, compare, and explore multiple trajectories efficiently. To evaluate AgentGit, we designed an experiment that optimizes target agents by selecting better prompts. We ran a multi-step A/B test against three baselines -- LangGraph, AutoGen, and Agno -- on a real-world task: retrieving and analyzing paper abstracts. Results show that AgentGit significantly reduces redundant computation, lowers runtime and token usage, and supports parallel exploration across multiple branches, enhancing both reliability and scalability in MAS development. This work offers a practical path to more robust MAS design and enables error recovery, safe exploration, iterative debugging, and A/B testing in collaborative AI systems.

cs.MA↗

arXiv2025

Hierarchical AI Multi-Agent Fundamental Investing: Evidence from China's A-Share Market

We present a multi-agent, AI-driven framework for fundamental investing that integrates macro indicators, industry-level and firm-specific information to construct optimized equity portfolios. The architecture comprises: (i) a Macro agent that dynamically screens and weights sectors based on evolving economic indicators and industry performance; (ii) four firm-level agents -- Fundamental, Technical, Report, and News -- that conduct in-depth analyses of individual firms to ensure both breadth and depth of coverage; (iii) a Portfolio agent that uses reinforcement learning to combine the agent outputs into a unified policy to generate the trading strategy; and (iv) a Risk Control agent that adjusts portfolio positions in response to market volatility. We evaluate the system on the constituents by the CSI 300 Index of China's A-share market and find that it consistently outperforms standard benchmarks and a state-of-the-art multi-agent trading system on risk-adjusted returns and drawdown control. Our core contribution is a hierarchical multi-agent design that links top-down macro screening with bottom-up fundamental analysis, offering a robust and extensible approach to factor-based portfolio construction.

q-fin.PM↗

arXiv2025

"Over-optimizing" for Normality: Budget-constrained Uncertainty Quantification for Contextual Decision-making

We study uncertainty quantification for contextual stochastic optimization, focusing on weighted sample average approximation (wSAA), which uses machine-learned relevance weights based on covariates. Although wSAA is widely used for contextual decisions, its uncertainty quantification remains limited. In addition, computational budgets tie sample size to optimization accuracy, creating a coupling that standard analyses often ignore. We establish central limit theorems for wSAA and construct asymptotic-normality-based confidence intervals for optimal conditional expected costs. We analyze the statistical--computational tradeoff under a computational budget, characterizing how to allocate resources between sample size and optimization iterations to balance statistical and optimization errors. These allocation rules depend on structural parameters of the objective; misspecifying them can break the asymptotic optimality of the wSAA estimator. We show that ``over-optimizing'' (running more iterations than the nominal rule) mitigates this misspecification and preserves asymptotic normality, at the expense of a slight slowdown in the convergence rate of the budget-constrained estimator. The common intuition that ``more data is better'' can fail under computational constraints: increasing the sample size may worsen statistical inference by forcing fewer algorithm iterations and larger optimization error. Our framework provides a principled way to quantify uncertainty for contextual decision-making under computational constraints. It offers practical guidance on allocating limited resources between data acquisition and optimization effort, clarifying when to prioritize additional optimization iterations over more data to ensure valid confidence intervals for conditional performance.

math.OC↗

arXiv2025

CPCL: Cross-Modal Prototypical Contrastive Learning for Weakly Supervised Text-based Person Retrieval

Weakly supervised text-based person retrieval seeks to retrieve images of a target person using textual descriptions, without relying on identity annotations and is more challenging and practical. The primary challenge is the intra-class differences, encompassing intra-modal feature variations and cross-modal semantic gaps. Prior works have focused on instance-level samples and ignored prototypical features of each person which are intrinsic and invariant. Toward this, we propose a Cross-Modal Prototypical Contrastive Learning (CPCL) method. In practice, the CPCL introduces the CLIP model to weakly supervised text-based person retrieval to map visual and textual instances into a shared latent space. Subsequently, the proposed Prototypical Multi-modal Memory (PMM) module captures associations between heterogeneous modalities of image-text pairs belonging to the same person through the Hybrid Cross-modal Matching (HCM) module in a many-to-many mapping fashion. Moreover, the Outlier Pseudo Label Mining (OPLM) module further distinguishes valuable outlier samples from each modality, enhancing the creation of more reliable clusters by mining implicit relationships between image-text pairs. We conduct extensive experiments on popular benchmarks of weakly supervised text-based person retrieval, which validate the effectiveness, generalizability of CPCL.

cs.CV↗

arXiv2025

Universal Magnetic Phases in Twisted Bilayer MoTe$_2$

Twisted bilayer MoTe$_2$ (tMoTe$_2$) has emerged as a robust platform for exploring correlated topological phases, notably supporting fractional Chern insulator (FCI) states at zero magnetic field across a wide range of twist angles. The evolution of magnetism and topology with twist angle remains an open question. Here, we systematically map the magnetic phase diagram of tMoTe$_2$ using local optical spectroscopy and scanning nanoSQUID-on-tip (nSOT) magnetometry. We identify spontaneous ferromagnetism at moiré filling factors $ν= -1$ and $-3$ over a twist angle range from 2.1$^\circ$ to 3.7$^\circ$, revealing a universal, twist-angle-insensitive ferromagnetic phase. At 2.1$^\circ$, we further observe robust ferromagnetism at $ν= -5$, absent in the devices with larger twist angle -- a signature of the flattening of higher bands in this twist angle range. Temperature-dependent measurements reveal a contrasting twist-angle dependence of the Curie temperatures between $ν= -1$ and $ν= -3$, indicating distinct interplay between exchange interaction and bandwidth for the two Chern bands. Despite spontaneous time-reversal symmetry breaking, we find no evidence of a topological gap at $ν= -3$; however, fragile correlated topological phases could be obscured by the device disorder evident in our spatially resolved measurements. Our results establish a global framework for understanding and controlling magnetic order in tMoTe$_2$ and highlight its potential for accessing correlated topological phases in higher energy Chern band.

cond-mat.mes-hall↗

arXiv2025

A Survey on Vision-Language-Action Models: An Action Tokenization Perspective

The remarkable advancements of vision and language foundation models in multimodal understanding, reasoning, and generation has sparked growing efforts to extend such intelligence to the physical world, fueling the flourishing of vision-language-action (VLA) models. Despite seemingly diverse approaches, we observe that current VLA models can be unified under a single framework: vision and language inputs are processed by a series of VLA modules, producing a chain of \textit{action tokens} that progressively encode more grounded and actionable information, ultimately generating executable actions. We further determine that the primary design choice distinguishing VLA models lies in how action tokens are formulated, which can be categorized into language description, code, affordance, trajectory, goal state, latent representation, raw action, and reasoning. However, there remains a lack of comprehensive understanding regarding action tokens, significantly impeding effective VLA development and obscuring future directions. Therefore, this survey aims to categorize and interpret existing VLA research through the lens of action tokenization, distill the strengths and limitations of each token type, and identify areas for improvement. Through this systematic review and analysis, we offer a synthesized outlook on the broader evolution of VLA models, highlight underexplored yet promising directions, and contribute guidance for future research, hoping to bring the field closer to general-purpose intelligence.

cs.RO↗

arXiv2025

Detection of subsurface structures with a vehicle-based atom gravity gradiometer

High-precision mobile gravity gradiometers are very useful in geodesy and geophysics. Atom gravity gradiometers (AGGs) could be among the most accurate mobile gravity gradiometers but are currently constrained by the trade-off between portability and sensitivity. Here, we present a high-sensitivity mobile AGG featuring an ultra-compact sensor head with a volume of only 94 L. In the laboratory, it achieves a sensitivity of 77 E/$\sqrt{Hz}$ (1 E=1$\times10^{-9}$/s$^2$) and a long-term stability of better than 0.5 E. We integrated the instrument in a minivan, enabling efficient mobile field surveys with excellent maneuverability in confined spaces. Using this vehicular system, we surveyed the gravitational field over a set of subsurface structures within a small wooded area, successfully resolving their structural signatures with a signal-to-noise ratio of 57 and quantifying the water depth in a reservoir with an accuracy of $\pm$0.23 m. Compared with previous observations using a CG-5 gravimeter, the superior spatial resolution inherent in gradiometry is clearly demonstrated. This work paves the way for bring AGGs to practical field applications.

physics.ins-det↗

arXiv2025

Adaptive Moment Estimation Optimization Algorithm Using Projection Gradient for Deep Learning

Training deep neural networks is challenging. To accelerate training and enhance performance, we propose PadamP, a novel optimization algorithm. PadamP is derived by applying the adaptive estimation of the p-th power of the second-order moments under scale invariance, enhancing projection adaptability by modifying the projection discrimination condition. It is integrated into Adam-type algorithms, accelerating training, boosting performance, and improving generalization in deep learning. Combining projected gradient benefits with adaptive moment estimation, PadamP tackles unconstrained non-convex problems. Convergence for the non-convex case is analyzed, focusing on the decoupling of first-order moment estimation coefficients and second-order moment estimation coefficients. Unlike prior work relying on , our proof generalizes the convergence theorem, enhancing practicality. Experiments using VGG-16 and ResNet-18 on CIFAR-10 and CIFAR-100 show PadamP's effectiveness, with notable performance on CIFAR-10/100, especially for VGG-16. The results demonstrate that PadamP outperforms existing algorithms in terms of convergence speed and generalization ability, making it a valuable addition to the field of deep learning optimization.

math.OC↗

arXiv2025

Diffusion-based Hierarchical Negative Sampling for Multimodal Knowledge Graph Completion

Multimodal Knowledge Graph Completion (MMKGC) aims to address the critical issue of missing knowledge in multimodal knowledge graphs (MMKGs) for their better applications. However, both the previous MMGKC and negative sampling (NS) approaches ignore the employment of multimodal information to generate diverse and high-quality negative triples from various semantic levels and hardness levels, thereby limiting the effectiveness of training MMKGC models. Thus, we propose a novel Diffusion-based Hierarchical Negative Sampling (DHNS) scheme tailored for MMKGC tasks, which tackles the challenge of generating high-quality negative triples by leveraging a Diffusion-based Hierarchical Embedding Generation (DiffHEG) that progressively conditions on entities and relations as well as multimodal semantics. Furthermore, we develop a Negative Triple-Adaptive Training (NTAT) strategy that dynamically adjusts training margins associated with the hardness level of the synthesized negative triples, facilitating a more robust and effective learning procedure to distinguish between positive and negative triples. Extensive experiments on three MMKGC benchmark datasets demonstrate that our framework outperforms several state-of-the-art MMKGC models and negative sampling techniques, illustrating the effectiveness of our DHNS for training MMKGC models. The source codes and datasets of this paper are available at https://github.com/ngl567/DHNS.

cs.AI↗

arXiv2025

Seesaw Experimentation: A/B Tests with Spillovers

This paper examines how spillover effects in A/B testing can impede organizational progress and develops strategies for mitigating these challenges. We identify a phenomenon termed ``seesaw experimentation'', where a firm's overall performance paradoxically deteriorates despite achieving continuous improvements in measured A/B testing metrics. Seesaw experimentation arises when successful innovations in primary metrics generate negative externalities in secondary, unmeasured dimensions. To address this problem, we propose implementing a positive hurdle rate for A/B test approval. We derive the optimal hurdle rate, offering a simple solution that preserves decentralized experimentation while mitigating negative spillovers.

econ.GN↗

arXiv2025

Conjugate-Gradient-like Based Adaptive Moment Estimation Optimization Algorithm for Deep Learning

Training deep neural networks is a challenging task. In order to speed up training and enhance the performance of deep neural networks, we rectify the vanilla conjugate gradient as conjugate-gradient-like and incorporate it into the generic Adam, and thus propose a new optimization algorithm named CG-like-Adam for deep learning. Specifically, both the first-order and the second-order moment estimation of generic Adam are replaced by the conjugate-gradient-like. Convergence analysis handles the cases where the exponential moving average coefficient of the first-order moment estimation is constant and the first-order moment estimation is unbiased. Numerical experiments show the superiority of the proposed algorithm based on the CIFAR10/100 dataset.

cs.LG↗

arXiv2024

DEMONet: Underwater Acoustic Target Recognition based on Multi-Expert Network and Cross-Temporal Variational Autoencoder

Building a robust underwater acoustic recognition system in real-world scenarios is challenging due to the complex underwater environment and the dynamic motion states of targets. A promising optimization approach is to leverage the intrinsic physical characteristics of targets, which remain invariable regardless of environmental conditions, to provide robust insights. However, our study reveals that while physical characteristics exhibit robust properties, they may lack class-specific discriminative patterns. Consequently, directly incorporating physical characteristics into model training can potentially introduce unintended inductive biases, leading to performance degradation. To utilize the benefits of physical characteristics while mitigating possible detrimental effects, we propose DEMONet in this study, which utilizes the detection of envelope modulation on noise (DEMON) to provide robust insights into the shaft frequency or blade counts of targets. DEMONet is a multi-expert network that allocates various underwater signals to their best-matched expert layer based on DEMON spectra for fine-grained signal processing. Thereinto, DEMON spectra are solely responsible for providing implicit physical characteristics without establishing a mapping relationship with the target category. Furthermore, to mitigate noise and spurious modulation spectra in DEMON features, we introduce a cross-temporal alignment strategy and employ a variational autoencoder (VAE) to reconstruct noise-resistant DEMON spectra to replace the raw DEMON features. The effectiveness of the proposed DEMONet with cross-temporal VAE was primarily evaluated on the DeepShip dataset and our proprietary datasets. Experimental results demonstrated that our approach could achieve state-of-the-art performance on both datasets.

cs.SD↗

arXiv2024

Realization of cold atom gyroscope in space

High-precision gyroscopes in space are essential for fundamental physics research and navigation. Due to its potential high precision, the cold atom gyroscope is expected to be the next generation of gyroscopes in space. Here, we report the first realization of a cold atom gyroscope, which was demonstrated by the atom interferometer installed in the China Space Station (CSS) as a payload. By compensating for CSS's high dynamic rotation rate using a built-in piezoelectric mirror, spatial interference fringes in the interferometer are successfully obtained. Then, the optimized ratio of the Raman laser's angles is derived, the coefficients of the piezoelectric mirror are self-calibrated in orbit, and various systemic effects are corrected. We achieve a rotation measurement resolution of 50*10^-6 rad/s for a single shot and 17*10^-6 rad/s for an average number of 32. The measured rotation is (-1142+/-29)*10^-6 rad/s and is compatible with that recorded by the classical gyroscope of the CSS. This study paves the way for developing high-precision cold atom gyroscopes in space.

physics.atom-ph↗

arXiv2024

Helmsman of the Masses? Evaluate the Opinion Leadership of Large Language Models in the Werewolf Game

Large language models (LLMs) have exhibited memorable strategic behaviors in social deductive games. However, the significance of opinion leadership exhibited by LLM-based agents has been largely overlooked, which is crucial for practical applications in multi-agent and human-AI interaction settings. Opinion leaders are individuals who have a noticeable impact on the beliefs and behaviors of others within a social group. In this work, we employ the Werewolf game as a simulation platform to assess the opinion leadership of LLMs. The game includes the role of the Sheriff, tasked with summarizing arguments and recommending decision options, and therefore serves as a credible proxy for an opinion leader. We develop a framework integrating the Sheriff role and devise two novel metrics based on the critical characteristics of opinion leaders. The first metric measures the reliability of the opinion leader, and the second assesses the influence of the opinion leader on other players' decisions. We conduct extensive experiments to evaluate LLMs of different scales. In addition, we collect a Werewolf question-answering dataset (WWQA) to assess and enhance LLM's grasp of the game rules, and we also incorporate human participants for further analysis. The results suggest that the Werewolf game is a suitable test bed to evaluate the opinion leadership of LLMs, and few LLMs possess the capacity for opinion leadership.

cs.CL↗

arXiv2024

A First-Principle Approach to X-ray Active Optics: Design and Verification

This paper presents the first-principle design approach for X-ray active optics, using the simulation-modulation cycle in place of the measurement-modulation feedback loops used in traditional active optics. Hence, the new active optics have the potential to outperform the accuracy of surface-shape metrology instruments. We apply an X-ray mirror with localized thermal elastic deformation to validate the idea. Both the finite element simulations and surface shape measurements have demonstrated that the active optics modulation accuracy limit can be achieved at the atomic layer level. It is believed that the implementation of the first-principle design strategy has the capacity to revolutionize both the manufacturing processes of X-ray mirrors and the beamline engineering of synchrotron radiation.

physics.optics↗

arXiv2023

Smooth Nested Simulation: Bridging Cubic and Square Root Convergence Rates in High Dimensions

Nested simulation concerns estimating functionals of a conditional expectation via simulation. In this paper, we propose a new method based on kernel ridge regression to exploit the smoothness of the conditional expectation as a function of the multidimensional conditioning variable. Asymptotic analysis shows that the proposed method can effectively alleviate the curse of dimensionality on the convergence rate as the simulation budget increases, provided that the conditional expectation is sufficiently smooth. The smoothness bridges the gap between the cubic root convergence rate (that is, the optimal rate for the standard nested simulation) and the square root convergence rate (that is, the canonical rate for the standard Monte Carlo simulation). We demonstrate the performance of the proposed method via numerical examples from portfolio risk management and input uncertainty quantification.

stat.ME↗

arXiv2023

Dynamic Selection in Algorithmic Decision-making

This paper identifies and addresses dynamic selection problems in online learning algorithms with endogenous data. In a contextual multi-armed bandit model, a novel bias (self-fulfilling bias) arises because the endogeneity of the data influences the choices of decisions, affecting the distribution of future data to be collected and analyzed. We propose an instrumental-variable-based algorithm to correct for the bias. It obtains true parameter values and attains low (logarithmic-like) regret levels. We also prove a central limit theorem for statistical inference. To establish the theoretical properties, we develop a general technique that untangles the interdependence between data and actions.

econ.EM↗

arXiv2023

How Does Risk Hedging Impact Operations? Insights from a Price-Setting Newsvendor Model

If a financial asset's price movement impacts a firm's product demand, the firm can respond to the impact by adjusting its operational decisions. For example, in the automotive industry, car makers decrease the selling prices of fuel-inefficient cars when the oil price rises. Meanwhile, the firm can implement a risk-hedging strategy using the financial asset jointly with its operational decisions. Motivated by this, we develop and solve a general risk-management model integrating risk hedging into a price-setting newsvendor. The optimal hedging strategy is calculated analytically, which leads to an explicit objective function for optimizing price and ``virtual production quantity'' (VPQ). (The latter determines the service level, i.e., the demand fulfillment probability.) We find that hedging generally reduces the optimal price {when the firm sets the target mean return as its production-only maximum expected profit. With the same condition on the target mean return}, hedging also reduces the optimal VPQ when the asset price trend positively impacts product demand; meanwhile, it may increase the VPQ by a small margin when the impact is negative. We construct the return-risk efficient frontier that characterizes the optimal return-risk trade-off. Our numerical study using data from a prominent automotive manufacturer shows that the markdowns in price and reduction in VPQ are small under our model and that the hedging strategy substantially reduces risk without materially reducing operational profit.

q-fin.RM↗