arXiv subjects

Bo Zhang

Publications and source records attributed to Bo Zhang.

At least 289 records · Page 16Linked to original sources

arXiv2025

GDI-Bench: A Benchmark for General Document Intelligence with Vision and Reasoning Decoupling

The rapid advancement of multimodal large language models (MLLMs) has profoundly impacted the document domain, creating a wide array of application scenarios. This progress highlights the need for a comprehensive benchmark to evaluate these models' capabilities across various document-specific tasks. However, existing benchmarks often fail to locate specific model weaknesses or guide systematic improvements. To bridge this gap, we introduce a General Document Intelligence Benchmark (GDI-Bench), featuring 2.3k images across 9 key scenarios and 19 document-specific tasks. By decoupling visual complexity and reasoning complexity, the GDI-Bench structures graded tasks that allow performance assessment by difficulty, aiding in model weakness identification and optimization guidance. We evaluate various open-source and closed-source models on GDI-Bench, conducting decoupled analyses in the visual and reasoning domains, revealing their strengths and weaknesses. To address the diverse tasks and domains in the GDI-Bench, we propose a GDI-Model that mitigates catastrophic forgetting during the supervised fine-tuning (SFT) process through an intelligence-preserving training strategy, thereby reinforcing the inherent weaknesses of the base model. Our model achieves state-of-the-art performance on previous benchmarks and the GDI-Bench. Both our benchmark and models are or will be open-sourced on https://huggingface.co/GDIBench.

cs.CL↗

arXiv2025

Chiral Valley Edge States

Valleytronics has emerged as a promising paradigm, enabling comprehensive control of the valley degree of freedom (DoF) for energy-efficient and high-speed information processing. However, backscattering-induced valley depolarization remains a fundamental limitation, stemming from the weak topological protection of the valley Hall phase. Here, we propose and demonstrate the concept of chiral valley edge states, which integrate the robust unidirectional chiral edge states with valley DoF. By controlling the valley Dirac masses, we selectively confine the chiral edge band around a single valley, enabling back-scattering-free propagation while imparting valley polarization. Our strategy not only addresses the valley depolarization issue but also introduces a unique functionality--valley multiplexing--allowing independent and arbitrary control over waves associated with different valley polarizations. We demonstrate our concept experimentally within hybrid topological photonic crystal systems composed of Chern and valley photonic crystals. Moreover, two key components for valley multiplexing are demonstrated: a valley (de-)multiplexer and a valley-locked waveguide crossing, facilitating non-interfering signal routing. Our results establish a novel interplay between the topological quantum Hall and valley Hall phases, offering a new framework for robust valley-based information processing.

physics.optics↗

arXiv2025

Pseudo-Likelihood Ratio Screening based on Network Data with Applications

Social network platforms today generate vast amounts of data, including network structures and a large number of user-defined tags, which reflect users' interests. The dimensionality of these personalized tags can be ultra-high, posing challenges for model analysis in targeted preference analysis. Traditional categorical feature screening methods overlook the network structure, which can lead to incorrect feature set and suboptimal prediction accuracy. This study focuses on feature screening for network-involved preference analysis based on ultra-high-dimensional categorical tags. We introduce the concepts of self-related features and network-related features, defined as those directly related to the response and those related to the network structure, respectively. We then propose a pseudo-likelihood ratio feature screening procedure that identifies both types of features. Theoretical properties of this procedure under different scenarios are thoroughly investigated. Extensive simulations and real data analysis on Sina Weibo validate our findings.

stat.ME↗

arXiv2025

Jailbreaking LLMs' Safeguard with Universal Magic Words for Text Embedding Models

The security issue of large language models (LLMs) has gained wide attention recently, with various defense mechanisms developed to prevent harmful output, among which safeguards based on text embedding models serve as a fundamental defense. Through testing, we discover that the output distribution of text embedding models is severely biased with a large mean. Inspired by this observation, we propose novel, efficient methods to search for **universal magic words** that attack text embedding models. Universal magic words as suffixes can shift the embedding of any text towards the bias direction, thus manipulating the similarity of any text pair and misleading safeguards. Attackers can jailbreak the safeguards by appending magic words to user prompts and requiring LLMs to end answers with magic words. Experiments show that magic word attacks significantly degrade safeguard performance on JailbreakBench, cause real-world chatbots to produce harmful outputs in full-pipeline attacks, and generalize across input/output texts, models, and languages. To eradicate this security risk, we also propose defense methods against such attacks, which can correct the bias of text embeddings and improve downstream performance in a train-free manner.

cs.CL↗

arXiv2025

Convergence Analysis of the Last Iterate in Distributed Stochastic Gradient Descent with Momentum

Distributed stochastic gradient methods are widely used to preserve data privacy and ensure scalability in large-scale learning tasks. While existing theory on distributed momentum Stochastic Gradient Descent (mSGD) mainly focuses on time-averaged convergence, the more practical last-iterate convergence remains underexplored. In this work, we analyze the last-iterate convergence behavior of distributed mSGD in non-convex settings under the classical Robbins-Monro step-size schedule. We prove both almost sure convergence and $L_2$ convergence of the last iterate, and derive convergence rates. We further show that momentum can accelerate early-stage convergence, and provide experiments to support our theory.

math.OC↗

arXiv2025

Flash-VL 2B: Optimizing Vision-Language Model Performance for Ultra-Low Latency and High Throughput

In this paper, we introduce Flash-VL 2B, a novel approach to optimizing Vision-Language Models (VLMs) for real-time applications, targeting ultra-low latency and high throughput without sacrificing accuracy. Leveraging advanced architectural enhancements and efficient computational strategies, Flash-VL 2B is designed to maximize throughput by reducing processing time while maintaining competitive performance across multiple vision-language benchmarks. Our approach includes tailored architectural choices, token compression mechanisms, data curation, training schemes, and a novel image processing technique called implicit semantic stitching that effectively balances computational load and model performance. Through extensive evaluations on 11 standard VLM benchmarks, we demonstrate that Flash-VL 2B achieves state-of-the-art results in both speed and accuracy, making it a promising solution for deployment in resource-constrained environments and large-scale real-time applications.

cs.CV↗

arXiv2025

Bipolar doping in van der Waals semiconductor through Flexo-doping

Doping plays a key role in functionalizing semiconductor devices, yet traditional chemical approaches relying on foreign-atom incorporation suffer from doping-asymmetry, pronounced lattice disorder and constrained spatial resolution. Here, we demonstrate a physical doping technique to directly write nanoscale doping patterns into layered semiconductors (MoS2). By applying localized tensile and compressive stress via an atomic force microscopy probe, p and n type conductance are simultaneously written into the designed area with sub-100-nm resolution, as verified by spatially resolved capacitance and photocurrent experiments. Density functional theory calculations reveal strain-driven shifts of donor and acceptor levels, as large as several hundreds of meV, linking mechanical stress to semiconductor doping. Fabricated strain-engineered junction efficiently rectifies the current flow and performs logic operations with stable dynamic response. This strain-driven approach enables spatially precise doping in van der Waals materials without degrading crystallinity, offering a versatile platform for nanoscale semiconductor devices.

cond-mat.mtrl-sci↗

arXiv2025

On Reducing the Execution Latency of Superconducting Quantum Processors via Quantum Job Scheduling

Quantum computing has gained considerable attention, especially after the arrival of the Noisy Intermediate-Scale Quantum (NISQ) era. Quantum processors and cloud services have been made world-wide increasingly available. Unfortunately, jobs on existing quantum processors are often executed in series, and the workload could be heavy to the processor. Typically, one has to wait for hours or even longer to obtain the result of a single quantum job on public quantum cloud due to long queue time. In fact, as the scale grows, the qubit utilization rate of the serial execution mode will further diminish, causing the waste of quantum resources. In this paper, to our best knowledge for the first time, the Quantum Job Scheduling Problem (QJSP) is formulated and introduced, and we accordingly aim to improve the utility efficiency of quantum resources. Specifically, a noise-aware quantum job scheduler (NAQJS) concerning the circuit width, number of measurement shots, and submission time of quantum jobs is proposed to reduce the execution latency. We conduct extensive experiments on a simulated Qiskit noise model, as well as on the Xiaohong (from QuantumCTek) superconducting quantum processor. Numerical results show the effectiveness in both the QPU time and turnaround time.

cs.AR↗

arXiv2025

Mantodea phylogenomics provides new insights into X-chromosome progression and evolutionary radiation

Background: Praying mantises, members of the order Mantodea, play important roles in agriculture, medicine, bionics, and entertainment. However, the scarcity of genomic resources has hindered extensive studies on mantis evolution and behaviour. Results: Here, we present the chromosome-scale reference genomes of five mantis species: the European mantis (Mantis religiosa), Chinese mantis (Tenodera sinensis), triangle dead leaf mantis (Deroplatys truncata), orchid mantis (Hymenopus coronatus), and metallic mantis (Metallyticus violaceus). We found that transposable element expansion is the major force governing genome size in Mantodea. Based on whole-alignments, we deduced that the Mantodea ancestor may have had only one X chromosome and that translocations between the X chromosome and an autosome may have occurred in the lineage of the superfamily Mantoidea. Furthermore, we found a lower evolutionary rate for the metallic mantis than for the other mantises. We also found that Mantodea underwent rapid radiation after the K-Pg mass extinction event, which could have contributed to the confusion in species classification. Conclusions: We present the chromosome-scale reference genomes of five mantis species to reveal the X-chromosome evolution, clarify the phylogeny relationship, and transposable element expansion.

q-bio.PE↗

arXiv2025

ChartX & ChartVLM: A Versatile Benchmark and Foundation Model for Complicated Chart Reasoning

Recently, many versatile Multi-modal Large Language Models (MLLMs) have emerged continuously. However, their capacity to query information depicted in visual charts and engage in reasoning based on the queried contents remains under-explored. In this paper, to comprehensively and rigorously benchmark the ability of the off-the-shelf MLLMs in the chart domain, we construct ChartX, a multi-modal evaluation set covering 18 chart types, 7 chart tasks, 22 disciplinary topics, and high-quality chart data. Besides, we develop ChartVLM to offer a new perspective on handling multi-modal tasks that strongly depend on interpretable patterns, such as reasoning tasks in the field of charts or geometric images. We evaluate the chart-related ability of mainstream MLLMs and our ChartVLM on the proposed ChartX evaluation set. Extensive experiments demonstrate that ChartVLM surpasses both versatile and chart-related large models, achieving results comparable to GPT-4V. We believe that our study can pave the way for further exploration in creating a more comprehensive chart evaluation set and developing more interpretable multi-modal models. Both ChartX and ChartVLM are available at: https://github.com/Alpha-Innovator/ChartVLM

cs.CV↗

arXiv2025

Stochastic Gradient Descent in Non-Convex Problems: Asymptotic Convergence with Relaxed Step-Size via Stopping Time Methods

Stochastic Gradient Descent (SGD) is widely used in machine learning research. Previous convergence analyses of SGD under the vanishing step-size setting typically require Robbins-Monro conditions. However, in practice, a wider variety of step-size schemes are frequently employed, yet existing convergence results remain limited and often rely on strong assumptions. This paper bridges this gap by introducing a novel analytical framework based on a stopping-time method, enabling asymptotic convergence analysis of SGD under more relaxed step-size conditions and weaker assumptions. In the non-convex setting, we prove the almost sure convergence of SGD iterates for step-sizes $ \{ ε_t \}_{t \geq 1} $ satisfying $\sum_{t=1}^{+\infty} ε_t = +\infty$ and $\sum_{t=1}^{+\infty} ε_t^p < +\infty$ for some $p > 2$. Compared with previous studies, our analysis eliminates the global Lipschitz continuity assumption on the loss function and relaxes the boundedness requirements for higher-order moments of stochastic gradients. Building upon the almost sure convergence results, we further establish $L_2$ convergence. These significantly relaxed assumptions make our theoretical results more general, thereby enhancing their applicability in practical scenarios.

cs.LG↗

arXiv2025

Object Placement for Anything

Object placement aims to determine the appropriate placement (\emph{e.g.}, location and size) of a foreground object when placing it on the background image. Most previous works are limited by small-scale labeled dataset, which hinders the real-world application of object placement. In this work, we devise a semi-supervised framework which can exploit large-scale unlabeled dataset to promote the generalization ability of discriminative object placement models. The discriminative models predict the rationality label for each foreground placement given a foreground-background pair. To better leverage the labeled data, under the semi-supervised framework, we further propose to transfer the knowledge of rationality variation, \emph{i.e.}, whether the change of foreground placement would result in the change of rationality label, from labeled data to unlabeled data. Extensive experiments demonstrate that our framework can effectively enhance the generalization ability of discriminative object placement models.

cs.CV↗

arXiv2025

Two-dimensional perovskites with maximum symmetry enable exciton diffusion length exceeding 2 micrometers

Realizing semiconductors with high symmetry of their crystallographic structures has been a virtue of inorganic materials and has resulted in novel physical behaviors. In contrast, hybrid (organic and inorganic) crystals such as two-dimensional metal halide perovskites exhibit much lower crystal symmetry due to in-plane or out of plane octahedral distortions. Despite their amazing ability for photoinduced light emission at room temperature, the Achilles' heel of this attractive class of 2D materials for optoelectronics remains the poor control and lack of performance for charge carrier transport. Inspired by the tremendous charge carrier properties of the 3D cubic perovskite phase of FAPbI3 and combining the use of the appropriate cage cation, the spacer molecule and the temperature and rate of crystallization, we report a new series of FA-based layered two-dimensional perovskites that exhibits the highest theoretically predicted symmetry with a tetragonal P4/mmm space group, resulting in no octahedral distortion in both in-plane and out-of-plane directions. These 2D perovskites present the shortest interlayer distances (4 angstrom), which results in systematically lower bandgaps (1.7 to 1.8 eV). Finally, the absence of octahedral distortions, results in an exciton diffusion length of 2.5 μm, and a diffusivity of 4.4 cm2s-1, both of which are an order of magnitude larger compared to previously reported 2D perovskites and on par with monolayer transition metal dichalcogenides.

cond-mat.mtrl-sci↗

arXiv2025

Efficient Tuning of Large Language Models for Knowledge-Grounded Dialogue Generation

Large language models (LLMs) demonstrate remarkable text comprehension and generation capabilities but often lack the ability to utilize up-to-date or domain-specific knowledge not included in their training data. To address this gap, we introduce KEDiT, an efficient method for fine-tuning LLMs for knowledge-grounded dialogue generation. KEDiT operates in two main phases: first, it employs an information bottleneck to compress retrieved knowledge into learnable parameters, retaining essential information while minimizing computational overhead. Second, a lightweight knowledge-aware adapter integrates these compressed knowledge vectors into the LLM during fine-tuning, updating less than 2\% of the model parameters. The experimental results on the Wizard of Wikipedia and a newly constructed PubMed-Dialog dataset demonstrate that KEDiT excels in generating contextually relevant and informative responses, outperforming competitive baselines in automatic, LLM-based, and human evaluations. This approach effectively combines the strengths of pretrained LLMs with the adaptability needed for incorporating dynamic knowledge, presenting a scalable solution for fields such as medicine.

cs.CL↗

arXiv2025

Dolphin: Moving Towards Closed-loop Auto-research through Thinking, Practice, and Feedback

The scientific research paradigm is undergoing a profound transformation owing to the development of Artificial Intelligence (AI). Recent works demonstrate that various AI-assisted research methods can largely improve research efficiency by improving data analysis, accelerating computation, and fostering novel idea generation. To further move towards the ultimate goal (i.e., automatic scientific research), in this paper, we introduce Dolphin, a closed-loop LLM-driven framework to enhance the automation level of scientific research. Dolphin first generates novel ideas based on feedback from previous experiments and relevant papers ranked by the topic and task attributes. Then, the generated ideas can be implemented using a code template refined and debugged with the designed exception-traceback-guided local code structure. Finally, Dolphin automatically analyzes the results of each idea and feeds the results back to the next round of idea generation. Experiments are conducted on the benchmark datasets of different topics and a subset of MLE-bench. Results show that Dolphin can continuously improve the performance of the input topic in a loop. We highlight that Dolphin can automatically propose methods that are comparable to the state-of-the-art in some tasks such as 3D point classification.

cs.AI↗

arXiv2025

Leveraging Shared Factor Structures for Enhanced Matrix Completion with Nonconvex Penalty Regularization

This article investigates the problem of noisy low-rank matrix completion with a shared factor structure, leveraging the auxiliary information from the missing indicator matrix to enhance prediction accuracy. Despite decades of development in matrix completion, the potential relationship between observed data and missing indicators has largely been overlooked. To address this gap, we propose a joint modeling framework for the observed data and missing indicators within the context of a generalized factor model and derive the asymptotic limit distribution of the estimators. Furthermore, to tackle the rank estimation problem for model specification, we employ matrix nonconvex penalty regularization and establish nonasymptotic probability guarantees for the Oracle property. The theoretical results are validated through extensive simulation studies and real-world data analysis, demonstrating the effectiveness of the proposed method.

stat.ME↗

arXiv2025

Negative Control Outcome Adjustment in Early-Phase Randomized Trials: Estimating Vaccine Effects on Immune Responses in HIV Exposed Uninfected Infants

Adjustment for prognostic baseline variables can reduce bias due to covariate imbalance and increase efficiency in randomized trials. While the use of covariate adjustment in late-phase trials is justified by favorable large-sample properties, it is seldom used in small, early-phase studies, due to uncertainty in which variables are prognostic and the potential for precision loss, type I error rate inflation, and undercoverage of confidence intervals. To address this problem, we consider adjustment for a valid negative control outcome (NCO), or an auxiliary post-randomization outcome believed completely unaffected by treatment but more highly correlated with the primary outcome than baseline covariates. We articulate the assumptions that permit adjustment for NCOs without producing post-randomization selection bias, and describe plausible data generating models where NCO adjustment can improve upon adjustment for baseline covariates alone. In numerical experiments, we illustrate performance and provide practical recommendations regarding model selection and finite-sample variance corrections. We apply our methods to the reanalysis of two early-phase vaccine trials in HIV exposed uninfected (HEU) infants, where we demonstrate that adjustment for auxiliary post-baseline immunological parameters can enhance precision of vaccine effect estimates relative to standard approaches that avoid adjustment or adjust for baseline covariates alone.

stat.ME↗

arXiv2025

MureObjectStitch: Multi-reference Image Composition

Generative image composition aims to regenerate the given foreground object in the background image to produce a realistic composite image. The existing methods are struggling to preserve the foreground details and adjust the foreground pose/viewpoint at the same time. In this work, we propose an effective finetuning strategy for generative image composition model, in which we finetune a pretrained model using one or more images containing the same foreground object. Moreover, we propose a multi-reference strategy, which allows the model to take in multiple reference images of the foreground object. The experiments on MureCOM dataset verify the effectiveness of our method. The code and model have been released at https://github.com/bcmi/MureObjectStitch-Image-Composition.

cs.CV↗