arXiv subjects

Qin Zhang

Publications and source records attributed to Qin Zhang.

At least 37 records · Page 2Linked to original sources

arXiv2026

Frequency Moments in Noisy Streaming and Distributed Data under Mismatch Ambiguity

We propose a novel framework for statistical estimation on noisy datasets. Within this framework, we focus on the frequency moments ($F_p$) problem and demonstrate that it is possible to approximate $F_p$ of the unknown ground-truth dataset using sublinear space in the data stream model and sublinear communication in the coordinator model, provided that the approximation ratio is parameterized by a data-dependent quantity, which we call the $F_p$-mismatch-ambiguity. We also establish a set of lower bounds, which are tight in terms of the input size. Our results yield several interesting insights: (1) In the data stream model, the $F_p$ problem is inherently more difficult in the noisy setting than in the noiseless one. In particular, while $F_2$ can be approximated in logarithmic space in terms of the input size in the noiseless setting, any algorithm for $F_2$ in the noisy setting requires polynomial space. (2) In the coordinator model, in sharp contrast to the noiseless case, achieving polylogarithmic communication in the input size is generally impossible for $F_p$ under noise. However, when the $F_p$ mismatch ambiguity falls below a certain threshold, it becomes possible to achieve communication that is entirely independent of the input size.

cs.DS↗

arXiv2026

VisRef: Visual Refocusing while Thinking Improves Test-Time Scaling in Multi-Modal Large Reasoning Models

Advances in large reasoning models have shown strong performance on complex reasoning tasks by scaling test-time compute through extended reasoning. However, recent studies observe that in vision-dependent tasks, extended textual reasoning at inference time can degrade performance as models progressively lose attention to visual tokens and increasingly rely on textual priors alone. To address this, prior works use reinforcement learning (RL)-based fine-tuning to route visual tokens or employ refocusing mechanisms during reasoning. While effective, these methods are computationally expensive, requiring large-scale data generation and policy optimization. To leverage the benefits of test-time compute without additional RL fine-tuning, we propose VisRef, a visually grounded test-time scaling framework. Our key idea is to actively guide the reasoning process by re-injecting a coreset of visual tokens that are semantically relevant to the reasoning context while remaining diverse and globally representative of the image, enabling more grounded multi-modal reasoning. Experiments on three visual reasoning benchmarks with state-of-the-art multi-modal large reasoning models demonstrate that, under fixed test-time compute budgets, VisRef consistently outperforms existing test-time scaling approaches by up to 6.4%.

cs.CV↗

arXiv2026

Towards Explicit Acoustic Evidence Perception in Audio LLMs for Speech Deepfake Detection

Speech deepfake detection (SDD) focuses on identifying whether a given speech signal is genuine or has been synthetically generated. Existing audio large language model (LLM)-based methods excel in content understanding; however, their predictions are often biased toward semantically correlated cues, which results in fine-grained acoustic artifacts being overlooked during the decisionmaking process. Consequently, fake speech with natural semantics can bypass detectors despite harboring subtle acoustic anomalies; this suggests that the challenge stems not from the absence of acoustic data, but from its inadequate accessibility when semantic-dominant reasoning prevails. To address this issue, we investigate SDD within the audio LLM paradigm and introduce SDD with Auditory Perception-enhanced Audio Large Language Model (SDD-APALLM), an acoustically enhanced framework designed to explicitly expose fine-grained time-frequency evidence as accessible acoustic cues. By combining raw audio with structured spectrograms, the proposed framework empowers audio LLMs to more effectively capture subtle acoustic inconsistencies without compromising their semantic understanding. Experimental results indicate consistent gains in detection accuracy and robustness, especially in cases where semantic cues are misleading. Further analysis reveals that these improvements stem from a coordinated utilization of semantic and acoustic information, as opposed to simple modality aggregation.

cs.SD↗

arXiv2025

Quantum thermal diode with additional control by auxiliary atomic states

A quantum thermal diode, similar to an electronic diode, allows for unidirectional heat transmission. In this paper, we study a quantum thermal diode composed of two two-level atoms coupled to auxiliary two-level atoms. We find that the excited auxiliary atoms can weaken heat current and enhance the rectification effect, but the ground-state auxiliary atoms can enhance heat current and weaken the rectification effect. The more auxiliary atoms are coupled, the stronger the enhancing or weakening impact is. If the auxiliary atom is in a superposition state, we find that only the fraction that projects onto the excited state plays a significant role. In particular, if we properly design the coupling of the auxiliary atoms, the rectification effect can be eliminated. This provides the potential to control the heat current and the rectification performance by the states of the auxiliary atoms.

quant-ph↗

arXiv2025

Magnetic control of nonlinear transport induced by the quantum metric

The quantum geometry plays a crucial role in the nonlinear transport of quantum materials. Here, we use the Boltzmann transport formalism to study the magnetic control of nonlinear transport induced by the quantum metric in two-dimensional systems with different types of spin-orbit coupling (SOC). It is shown that the nonlinear conductivity is strongly dependent on the direction of a field and reveals significant spatial anisotropy. Moreover, the field-direction dependent relations are distinct for different SOCs. In addition, it is demonstrated that the contributions from the quantum metric and Drude mechanism are distinguishable due to their opposite signs or distinct anisotropy relations. We further derive the analytical formulas for the anisotropic nonlinear conductivity, in exact agreement with numerical results. Our work shines more light on the interplay between the nonlinear transport and quantum geometry.

cond-mat.mes-hall↗

arXiv2025

Anisotropic nonlinear transport in two-dimensional ferroelectrics

The longitudinal nonlinear response plays a crucial role in the nonreciprocal charge transport and may provide a simple electrical means to probe the spin-orbit coupling, magnetic order and polarization states, etc. Here, we report on a study on the polarization and magnetic field control of longitudinal nonlinear transport in two-dimensional (2D) ferroelectrics with in-plane polarization. Based on the Boltzmann transport theory, we first study that using a general Hamiltonian model and show that the nonlinear conductivity can be significantly tuned by the polarization and magnetic field. In addition, the nonlinear conductivity reveals a strong spatial anisotropy. We further derive the analytical formulas for the anisotropic nonlinear conductivity in exact accordance with numerical results. Then, we exemplify those phenomena in the 2D ferroelectric SnTe monolayer in the presence of an external magnetic field based on the density functional theory calculations. It is also revealed that the polarity of nonlinear conductivity is locked to the direction of the polarization, thus pointing to the possibility of the nonlinear detection of polarization states. Our work uncovers intriguing features of the longitudinal nonlinear transport in 2D ferroelectrics and provides guidelines for designing the polarization control of rectifying devices.

cond-mat.mes-hall↗

arXiv2025

DeepServe: Serverless Large Language Model Serving at Scale

In this paper, we propose DEEPSERVE, a scalable and serverless AI platform designed to efficiently serve large language models (LLMs) at scale in cloud environments. DEEPSERVE addresses key challenges such as resource allocation, serving efficiency, and cold start latencies through four main design components. First, DEEPSERVE uses a simple serverless abstraction called the request-job-task model, which helps manage diverse AI workloads across posttraining and model-serving tasks. Second, DEEPSERVE integrates an in-house serving engine named FLOWSERVE using a microkernel-inspired design, NPU-centric execution, and SPMD-based parallelism to optimize LLM serving. Third, DEEPSERVE includes novel scheduling policies tailored for a configuration with both PD-disaggregated and PD-colocated instances. Fourth, DEEPSERVE includes optimizations such as pre-warmed pods, DRAM pre-loading, and NPU-fork, which allow DEEPSERVE to scale up to 64 instances in seconds. DEEPSERVE has been in production for over a year, operating on a large Ascend NPU cluster and providing industrystandard APIs for fine-tuning, agent serving, and model serving to our customers.

cs.DC↗

arXiv2025

AuthGuard: Generalizable Deepfake Detection via Language Guidance

Existing deepfake detection techniques struggle to keep-up with the ever-evolving novel, unseen forgeries methods. This limitation stems from their reliance on statistical artifacts learned during training, which are often tied to specific generation processes that may not be representative of samples from new, unseen deepfake generation methods encountered at test time. We propose that incorporating language guidance can improve deepfake detection generalization by integrating human-like commonsense reasoning -- such as recognizing logical inconsistencies and perceptual anomalies -- alongside statistical cues. To achieve this, we train an expert deepfake vision encoder by combining discriminative classification with image-text contrastive learning, where the text is generated by generalist MLLMs using few-shot prompting. This allows the encoder to extract both language-describable, commonsense deepfake artifacts and statistical forgery artifacts from pixel-level distributions. To further enhance robustness, we integrate data uncertainty learning into vision-language contrastive learning, mitigating noise in image-text supervision. Our expert vision encoder seamlessly interfaces with an LLM, further enabling more generalized and interpretable deepfake detection while also boosting accuracy. The resulting framework, AuthGuard, achieves state-of-the-art deepfake detection accuracy in both in-distribution and out-of-distribution settings, achieving AUC gains of 6.15% on the DFDC dataset and 16.68% on the DF40 dataset. Additionally, AuthGuard significantly enhances deepfake reasoning, improving performance by 24.69% on the DDVQA dataset.

cs.CV↗

arXiv2025

EPIC: Efficient Position-Independent Caching for Serving Large Language Models

Large Language Models (LLMs) show great capabilities in a wide range of applications, but serving them efficiently becomes increasingly challenging as requests (prompts) become more complex. Context caching improves serving performance by reusing Key-Value (KV) vectors, the intermediate representations of tokens that are repeated across requests. However, existing context caching requires exact prefix matches across requests, limiting reuse cases in settings such as few-shot learning and retrieval-augmented generation, where immutable content (e.g., documents) remains unchanged across requests but is preceded by varying prefixes. Position-Independent Caching (PIC) addresses this issue by enabling modular reuse of the KV vectors regardless of prefixes. We formalize PIC and advance prior work by introducing EPIC, a serving system incorporating our new LegoLink algorithm, which mitigates the inappropriate "attention sink" effect at every document beginning, to maintain accuracy with minimal computation. Experiments show that EPIC achieves up to 8x improvements in Time-To-First-Token (TTFT) and 7x throughput gains over existing systems, with negligible or no accuracy loss.

cs.LG↗

arXiv2025

Ground-V: Teaching VLMs to Ground Complex Instructions in Pixels

This work presents a simple yet effective workflow for automatically scaling instruction-following data to elicit pixel-level grounding capabilities of VLMs under complex instructions. In particular, we address five critical real-world challenges in text-instruction-based grounding: hallucinated references, multi-object scenarios, reasoning, multi-granularity, and part-level references. By leveraging knowledge distillation from a pre-trained teacher model, our approach generates high-quality instruction-response pairs linked to existing pixel-level annotations, minimizing the need for costly human annotation. The resulting dataset, Ground-V, captures rich object localization knowledge and nuanced pixel-level referring expressions. Experiment results show that models trained on Ground-V exhibit substantial improvements across diverse grounding tasks. Specifically, incorporating Ground-V during training directly achieves an average accuracy boost of 4.4% for LISA and a 7.9% for PSALM across six benchmarks on the gIoU metric. It also sets new state-of-the-art results on standard benchmarks such as RefCOCO/+/g. Notably, on gRefCOCO, we achieve an N-Acc of 83.3%, exceeding the previous state-of-the-art by more than 20%.

cs.CV↗

arXiv2025

Phase field model of Coulomb explosion damage in solid induced by ultrashort laser

Much experimental evidence reveals that Coulomb explosion governs non-thermal material removal under femtosecond or even shorter laser pulses, and non-thermal laser damage has been a topic widely discussed. Nevertheless, there is still no continuum mechanical model capable of describing the evolution of such damage. In this study, we develop a model that characterizes solid damage through a phase field variable governed by Allen-Cahn dynamics. The parameter of the model is defined by a conceptual mechanism: during Coulomb explosion, electron pressure surpasses the interatomic barrier potential, dissociates material from the solid surface as small equivalent particles and resulting in localized damage. The numerical simulation validates the model's availability and demonstrate its ability to predict damage morphology under varying laser conditions. This work advances the understanding of non-thermal ablation and provides a tool for optimizing ultrafast laser processing.

physics.optics↗

arXiv2025

Optimal Transport-Guided Source-Free Adaptation for Face Anti-Spoofing

Developing a face anti-spoofing model that meets the security requirements of clients worldwide is challenging due to the domain gap between training datasets and diverse end-user test data. Moreover, for security and privacy reasons, it is undesirable for clients to share a large amount of their face data with service providers. In this work, we introduce a novel method in which the face anti-spoofing model can be adapted by the client itself to a target domain at test time using only a small sample of data while keeping model parameters and training data inaccessible to the client. Specifically, we develop a prototype-based base model and an optimal transport-guided adaptor that enables adaptation in either a lightweight training or training-free fashion, without updating base model's parameters. Furthermore, we propose geodesic mixup, an optimal transport-based synthesis method that generates augmented training data along the geodesic path between source prototypes and target data distribution. This allows training a lightweight classifier to effectively adapt to target-specific characteristics while retaining essential knowledge learned from the source domain. In cross-domain and cross-attack settings, compared with recent methods, our method achieves average relative improvements of 19.17% in HTER and 8.58% in AUC, respectively.

cs.CV↗

arXiv2025

A vision-based robotic system for precision pollination of apples

Global food production depends upon successful pollination, a process that relies on natural and managed pollinators. However, natural pollinators are declining due to factors such as climate change, habitat loss, and pesticide use. This paper presents an integrated robotic system for precision pollination in apples. The system consisted of a machine vision system to identify target flower clusters and estimate their positions and orientations, and a manipulator motion planning and actuation system to guide the sprayer to apply charged pollen suspension to the target flower clusters. The system was tested in the lab, followed by field evaluation in Honeycrisp and Fuji orchards. In the Honeycrisp variety, the robotic pollination system achieved a fruit set of 34.8% of sprayed flowers with 87.5% of flower clusters having at least one fruit when a 2 gm/l pollen suspension was used. In comparison, the natural pollination technique achieved a fruit set of 43.1% with 94.9% of clusters with at least one fruit. In Fuji apples, the robotic system achieved lower pollination success, with 7.2% of sprayed flowers setting fruit and 20.6% of clusters having at least one fruit, compared to 33.1% and 80.6%, respectively, with natural pollination. Fruit quality analysis showed that robotically pollinated fruits were comparable to naturally pollinated fruits in terms of color, weight, diameter, firmness, soluble solids, and starch content. Additionally, the system cycle time was 6.5 seconds per cluster. The results showed a promise for robotic pollination in apple orchards. However, further research and development is needed to improve the system and assess its suitability across diverse orchard environments and apple cultivars.

cs.RO↗

arXiv2025

Faster Approximation Algorithms for k-Center via Data Reduction

We study efficient algorithms for the Euclidean $k$-Center problem, focusing on the regime of large $k$. We take the approach of data reduction by considering $α$-coreset, which is a small subset $S$ of the dataset $P$ such that any $β$-approximation on $S$ is an $(α+ β)$-approximation on $P$. We give efficient algorithms to construct coresets whose size is $k \cdot o(n)$, which immediately speeds up existing approximation algorithms. Notably, we obtain a near-linear time $O(1)$-approximation when $k = n^c$ for any $0 < c < 1$. We validate the performance of our coresets on real-world datasets with large $k$, and we observe that the coreset speeds up the well-known Gonzalez algorithm by up to $4$ times, while still achieving similar clustering cost. Technically, one of our coreset results is based on a new efficient construction of consistent hashing with competitive parameters. This general tool may be of independent interest for algorithm design in high dimensional Euclidean spaces.

cs.DS↗

arXiv2025

Breaking the $\log(1/Δ_2)$ Barrier: Better Batched Best Arm Identification with Adaptive Grids

We investigate the problem of batched best arm identification in multi-armed bandits, where we aim to identify the best arm from a set of $n$ arms while minimizing both the number of samples and batches. We introduce an algorithm that achieves near-optimal sample complexity and features an instance-sensitive batch complexity, which breaks the $\log(1/Δ_2)$ barrier. The main contribution of our algorithm is a novel sample allocation scheme that effectively balances exploration and exploitation for batch sizes. Experimental results indicate that our approach is more batch-efficient across various setups. We also extend this framework to the problem of batched best arm identification in linear bandits and achieve similar improvements.

cs.LG↗

arXiv2025

AgRegNet: A Deep Regression Network for Flower and Fruit Density Estimation, Localization, and Counting in Orchards

One of the major challenges for the agricultural industry today is the uncertainty in manual labor availability and the associated cost. Automated flower and fruit density estimation, localization, and counting could help streamline harvesting, yield estimation, and crop-load management strategies such as flower and fruitlet thinning. This article proposes a deep regression-based network, AgRegNet, to estimate density, count, and location of flower and fruit in tree fruit canopies without explicit object detection or polygon annotation. Inspired by popular U-Net architecture, AgRegNet is a U-shaped network with an encoder-to-decoder skip connection and modified ConvNeXt-T as an encoder feature extractor. AgRegNet can be trained based on information from point annotation and leverages segmentation information and attention modules (spatial and channel) to highlight relevant flower and fruit features while suppressing non-relevant background features. Experimental evaluation in apple flower and fruit canopy images under an unstructured orchard environment showed that AgRegNet achieved promising accuracy as measured by Structural Similarity Index (SSIM), percentage Mean Absolute Error (pMAE) and mean Average Precision (mAP) to estimate flower and fruit density, count, and centroid location, respectively. Specifically, the SSIM, pMAE, and mAP values for flower images were 0.938, 13.7%, and 0.81, respectively. For fruit images, the corresponding values were 0.910, 5.6%, and 0.93. Since the proposed approach relies on information from point annotation, it is suitable for sparsely and densely located objects. This simplified technique will be highly applicable for growers to accurately estimate yields and decide on optimal chemical and mechanical flower thinning practices.

cs.CV↗

arXiv2025

Quantum Data Sketches

Recent advancements in quantum technologies, particularly in quantum sensing and simulation, have facilitated the generation and analysis of inherently quantum data. This progress underscores the necessity for developing efficient and scalable quantum data management strategies. This goal faces immense challenges due to the exponential dimensionality of quantum data and its unique quantum properties such as no-cloning and measurement stochasticity. Specifically, classical storage and manipulation of an arbitrary n-qubit quantum state requires exponential space and time. Hence, there is a critical need to revisit foundational data management concepts and algorithms for quantum data. In this paper, we propose succinct quantum data sketches to support basic database operations such as search and selection. We view our work as an initial step towards the development of quantum data management model, opening up many possibilities for future research in this direction.

cs.DB↗

arXiv2024

Computing Approximate Graph Edit Distance via Optimal Transport

Given a graph pair $(G^1, G^2)$, graph edit distance (GED) is defined as the minimum number of edit operations converting $G^1$ to $G^2$. GED is a fundamental operation widely used in many applications, but its exact computation is NP-hard, so the approximation of GED has gained a lot of attention. Data-driven learning-based methods have been found to provide superior results compared to classical approximate algorithms, but they directly fit the coupling relationship between a pair of vertices from their vertex features. We argue that while pairwise vertex features can capture the coupling cost (discrepancy) of a pair of vertices, the vertex coupling matrix should be derived from the vertex-pair cost matrix through a more well-established method that is aware of the global context of the graph pair, such as optimal transport. In this paper, we propose an ensemble approach that integrates a supervised learning-based method and an unsupervised method, both based on optimal transport. Our learning method, GEDIOT, is based on inverse optimal transport that leverages a learnable Sinkhorn algorithm to generate the coupling matrix. Our unsupervised method, GEDGW, models GED computation as a linear combination of optimal transport and its variant, Gromov-Wasserstein discrepancy, for node and edge operations, respectively, which can be solved efficiently without needing the ground truth. Our ensemble method, GEDHOT, combines GEDIOT and GEDGW to further boost the performance. Extensive experiments demonstrate that our methods significantly outperform the existing methods in terms of the performance of GED computation, edit path generation, and model generalizability.

cs.LG↗