arXiv subjects

Anqi Zhang

Publications and source records attributed to Anqi Zhang.

18 recordsLinked to original sources

arXiv2026

Quantum-enhanced ghost imaging recognition via joint optimization of speckle patterns and quantum network parameters

Ghost imaging enables nonlocal image reconstruction and exhibits strong robustness against interference, but achieving high-fidelity recognition at ultra-low sampling rates remains challenging. Quantum machine learning offers a novel approach for efficient feature extraction on noisy medium-scale quantum devices; however, existing methods generally suffer from low recognition accuracy and weak noise resistance. This paper proposes a ghost imaging recognition method based on the simultaneous optimization of speckle patterns and quantum network parameters. By leveraging the mathematical equivalence between classical convolution and speckle-object dot product operations in ghost imaging, a speckle consistency regularization mechanism is introduced to achieve end-to-end joint optimization of optical coding and quantum feature extractors. A parallel 8-qubit quantum circuit employing block coding and a star-shaped entanglement structure is designed to extract higher-order features from bucket signals. Simulation results on the MNIST and Fashion-MNIST datasets show that at an ultra-low sampling rate of 1.5625%, the proposed framework achieves recognition accuracies of 90.1% and 81.7%, respectively, representing a 2.6% improvement over classical convolutional neural networks and a maximum improvement of 14.2% over traditional hybrid quantum machine learning models. This method also exhibits strong robustness to quantum noise and has been validated on a real optical ghost imaging system, achieving an average recognition accuracy of 84.8%. These results confirm that the joint optimization of speckle patterns and quantum network parameters provides a reliable and practical solution for low-sampling ghost imaging recognition.

quant-ph↗

arXiv2026

Quantum correlations of tripartite mixed states in the black hole quantum atmosphere

We investigate quantum state texture, genuine multipartite entanglement, and tripartite nonlocality of a tripartite mixed state in the black hole quantum atmosphere. By introducing the Hartle-Hawking local temperature into the Bogoliubov coefficients, we characterize the influence of the local Hawking effect on both physically accessible and inaccessible reduced states. We find that the extrema of these three quantities all lie in the same near horizon region and shift outward with increasing local Hawking temperature, coinciding with the peak region of the local Hawking temperature and indicating that different aspects of tripartite quantum information are most sensitive to the local Hawking effect in the same atmospheric region. In contrast to genuine multipartite entanglement, tripartite nonlocality is more fragile and is suppressed under stronger local Hawking effects. These results provide a unified characterization of density matrix restructuring, entanglement redistribution, and nonlocality in tripartite mixed states affected by the black hole quantum atmosphere.

quant-ph↗

arXiv2026

On Computing Total Variation Distance Between Mixtures of Product Distributions

We study the problem of approximating the total variation distance between two mixtures of product distributions over an $n$-dimensional discrete domain. Given two mixtures $\mathbb{P}$ and $\mathbb{Q}$ with $k_1$ and $k_2$ product distributions over $[q]^n$, respectively, we give a randomized algorithm that approximates $d_{\mathrm{TV}}\left({\mathbb{P}},{\mathbb{Q}}\right)$ within a multiplicative error of $(1\pm \varepsilon)$ in time $\mathrm{poly}((nq)^{k_1+k_2},1/\varepsilon)$. We also study the special case of mixtures of Boolean subcubes over $\{0,1\}^n$. For this class, we give a deterministic algorithm that exactly computes the total variation distance in time $\mathrm{poly}(n,2^{O(k_1+k_2)})$, and show that exact computation is $\#\mathsf{P}$-hard when $k_1+k_2=\Theta(n)$.

cs.DS↗

arXiv2026

Hierarchical Progressive Optimization for Multi-Qubit Pauli Noise Modeling

Quantum Noise Characterization (QNC) is indispensable for benchmarking and mitigating errors in Noisy Intermediate-Scale Quantum (NISQ) devices. However, traditional Quantum Process Tomography (QPT) suffers from an exponential parameter explosion, severely hindering its scalability. In this paper, we propose a Hierarchical Progressive Optimization (HPO) framework to efficiently extract high-order spatial crosstalk in multi-qubit systems. The complexity analysis shows that the combinatorial projection mask reduces the required number of Pauli transfer matrix (PTM) elements from O($16^N$) to O($N^2 3^N$). Numerical simulations on a 10-qubit HHL circuit achieve a fidelity of 0.9381 with the HPO method, compared to 0.7431 obtained using global depolarizing-noise mitigation.

quant-ph↗

arXiv2026

PdrQC: Pauli-space Discriminative Representations based Quantum Classifier

Quantum classification faces two key challenges. First, the difficulty of distinguishing between different classes varies: some class pairs are easy to separate, while others are more challenging. Second, practical execution is affected by noise, finite sampling, and measurement overhead. To address these issues, we propose the Pauli-Space Discriminative-Representation based Quantum Classifier (PdrQC), a framework for task-adaptive multiclass quantum classification. The method evaluates candidate upload circuits using low-weight Pauli features and formulates upload design as a structured model selection problem based on discriminative representations. By progressively selecting upload structures and compact Pauli readout features for the target multiclass task, the framework achieves a better balance between classification accuracy and resource efficiency. Numerical simulations were conducted on the MNIST and Fashion-MNIST datasets with $K\in\{2,3,5,7,10\}$. The results demonstrate that PdrQC, through its task-adaptive Pauli representation, achieves an effective balance among multiclass classification accuracy, quantum-circuit complexity, and measurement overhead, making it suitable for multiclass quantum classification under limited hardware resources.

quant-ph↗

arXiv2026

Rethinking MLLM Itself as a Segmenter with a Single Segmentation Token

Recent segmentation methods leveraging Multi-modal Large Language Models (MLLMs) have shown reliable object-level segmentation and enhanced spatial perception. However, almost all previous methods predominantly rely on specialist mask decoders to interpret masks from generated segmentation-related embeddings and visual features, or incorporate multiple additional tokens to assist. This paper aims to investigate whether and how we can unlock segmentation from MLLM itSELF with 1 segmentation Embedding (SELF1E) while achieving competitive results, which eliminates the need for external decoders. To this end, our approach targets the fundamental limitation of resolution reduction in pixel-shuffled image features from MLLMs. First, we retain image features at their original uncompressed resolution, and refill them with residual features extracted from MLLM-processed compressed features, thereby improving feature precision. Subsequently, we integrate pixel-unshuffle operations on image features with and without LLM processing, respectively, to unleash the details of compressed features and amplify the residual features under uncompressed resolution, which further enhances the resolution of refilled features. Moreover, we redesign the attention mask with dual perception pathways, i.e., image-to-image and image-to-segmentation, enabling rich feature interaction between pixels and the segmentation token. Comprehensive experiments across multiple segmentation tasks validate that SELF1E achieves performance competitive with specialist mask decoder-based methods, demonstrating the feasibility of decoder-free segmentation in MLLMs. Project page: https://github.com/ANDYZAQ/SELF1E.

cs.CV↗

arXiv2025

Reasoning Models Know When They're Right: Probing Hidden States for Self-Verification

Reasoning models have achieved remarkable performance on tasks like math and logical reasoning thanks to their ability to search during reasoning. However, they still suffer from overthinking, often performing unnecessary reasoning steps even after reaching the correct answer. This raises the question: can models evaluate the correctness of their intermediate answers during reasoning? In this work, we study whether reasoning models encode information about answer correctness through probing the model's hidden states. The resulting probe can verify intermediate answers with high accuracy and produces highly calibrated scores. Additionally, we find models' hidden states encode correctness of future answers, enabling early prediction of the correctness before the intermediate answer is fully formulated. We then use the probe as a verifier to decide whether to exit reasoning at intermediate answers during inference, reducing the number of inference tokens by 24\% without compromising performance. These findings confirm that reasoning models do encode a notion of correctness yet fail to exploit it, revealing substantial untapped potential to enhance their efficiency.

cs.AI↗

arXiv2025

CoMBO: Conflict Mitigation via Branched Optimization for Class Incremental Segmentation

Effective Class Incremental Segmentation (CIS) requires simultaneously mitigating catastrophic forgetting and ensuring sufficient plasticity to integrate new classes. The inherent conflict above often leads to a back-and-forth, which turns the objective into finding the balance between the performance of previous~(old) and incremental~(new) classes. To address this conflict, we introduce a novel approach, Conflict Mitigation via Branched Optimization~(CoMBO). Within this approach, we present the Query Conflict Reduction module, designed to explicitly refine queries for new classes through lightweight, class-specific adapters. This module provides an additional branch for the acquisition of new classes while preserving the original queries for distillation. Moreover, we develop two strategies to further mitigate the conflict following the branched structure, \textit{i.e.}, the Half-Learning Half-Distillation~(HDHL) over classification probabilities, and the Importance-Based Knowledge Distillation~(IKD) over query features. HDHL selectively engages in learning for classification probabilities of queries that match the ground truth of new classes, while aligning unmatched ones to the corresponding old probabilities, thus ensuring retention of old knowledge while absorbing new classes via learning negative samples. Meanwhile, IKD assesses the importance of queries based on their matching degree to old classes, prioritizing the distillation of important features and allowing less critical features to evolve. Extensive experiments in Class Incremental Panoptic and Semantic Segmentation settings have demonstrated the superior performance of CoMBO. Project page: https://guangyu-ryan.github.io/CoMBO.

cs.CV↗

arXiv2024

Bridge the Points: Graph-based Few-shot Segment Anything Semantically

The recent advancements in large-scale pre-training techniques have significantly enhanced the capabilities of vision foundation models, notably the Segment Anything Model (SAM), which can generate precise masks based on point and box prompts. Recent studies extend SAM to Few-shot Semantic Segmentation (FSS), focusing on prompt generation for SAM-based automatic semantic segmentation. However, these methods struggle with selecting suitable prompts, require specific hyperparameter settings for different scenarios, and experience prolonged one-shot inference times due to the overuse of SAM, resulting in low efficiency and limited automation ability. To address these issues, we propose a simple yet effective approach based on graph analysis. In particular, a Positive-Negative Alignment module dynamically selects the point prompts for generating masks, especially uncovering the potential of the background context as the negative reference. Another subsequent Point-Mask Clustering module aligns the granularity of masks and selected points as a directed graph, based on mask coverage over points. These points are then aggregated by decomposing the weakly connected components of the directed graph in an efficient manner, constructing distinct natural clusters. Finally, the positive and overshooting gating, benefiting from graph-based granularity alignment, aggregate high-confident masks and filter out the false-positive masks for final prediction, reducing the usage of additional hyperparameters and redundant mask generation. Extensive experimental analysis across standard FSS, One-shot Part Segmentation, and Cross Domain FSS datasets validate the effectiveness and efficiency of the proposed approach, surpassing state-of-the-art generalist models with a mIoU of 58.7% on COCO-20i and 35.2% on LVIS-92i. The code is available in https://andyzaq.github.io/GF-SAM/.

cs.CV↗

arXiv2024

Optimal Visual Search with Highly Heuristic Decision Rules

Visual search is a fundamental natural task for humans and other animals. We investigated the decision processes humans use in covert (single-fixation) search with briefly presented displays having well-separated potential target locations. Performance was compared with the Bayesian-optimal decision process under the assumption that the information from the different potential target locations is statistically independent. Surprisingly, humans performed slightly better than optimal, despite humans' substantial loss of sensitivity in the fovea (foveal neglect), and the implausibility of the human brain replicating the optimal computations. We show that three factors can quantitatively explain these seemingly paradoxical results. Most importantly, simple and fixed heuristic decision rules reach near optimal search performance. Secondly, foveal neglect primarily affects only the central potential target location. Finally, spatially correlated neural noise can cause search performance to exceed that predicted for independent noise. These findings have broad implications for understanding visual search tasks and other identification tasks in humans and other animals.

q-bio.NC↗

arXiv2024

Structure-enhanced Contrastive Learning for Graph Clustering

Graph clustering is a crucial task in network analysis with widespread applications, focusing on partitioning nodes into distinct groups with stronger intra-group connections than inter-group ones. Recently, contrastive learning has achieved significant progress in graph clustering. However, most methods suffer from the following issues: 1) an over-reliance on meticulously designed data augmentation strategies, which can undermine the potential of contrastive learning. 2) overlooking cluster-oriented structural information, particularly the higher-order cluster(community) structure information, which could unveil the mesoscopic cluster structure information of the network. In this study, Structure-enhanced Contrastive Learning (SECL) is introduced to addresses these issues by leveraging inherent network structures. SECL utilizes a cross-view contrastive learning mechanism to enhance node embeddings without elaborate data augmentations, a structural contrastive learning module for ensuring structural consistency, and a modularity maximization strategy for harnessing clustering-oriented information. This comprehensive approach results in robust node representations that greatly enhance clustering performance. Extensive experiments on six datasets confirm SECL's superiority over current state-of-the-art methods, indicating a substantial improvement in the domain of graph clustering.

cs.LG↗

arXiv2024

Adaptive Pre-training Data Detection for Large Language Models via Surprising Tokens

While large language models (LLMs) are extensively used, there are raising concerns regarding privacy, security, and copyright due to their opaque training data, which brings the problem of detecting pre-training data on the table. Current solutions to this problem leverage techniques explored in machine learning privacy such as Membership Inference Attacks (MIAs), which heavily depend on LLMs' capability of verbatim memorization. However, this reliance presents challenges, especially given the vast amount of training data and the restricted number of effective training epochs. In this paper, we propose an adaptive pre-training data detection method which alleviates this reliance and effectively amplify the identification. Our method adaptively locates \textit{surprising tokens} of the input. A token is surprising to a LLM if the prediction on the token is "certain but wrong", which refers to low Shannon entropy of the probability distribution and low probability of the ground truth token at the same time. By using the prediction probability of surprising tokens to measure \textit{surprising}, the detection method is achieved based on the simple hypothesis that seeing seen data is less surprising for the model compared with seeing unseen data. The method can be applied without any access to the the pre-training data corpus or additional training like reference models. Our approach exhibits a consistent enhancement compared to existing methods in diverse experiments conducted on various benchmarks and models, achieving a maximum improvement of 29.5\%. We also introduce a new benchmark Dolma-Book developed upon a novel framework, which employs book data collected both before and after model training to provide further evaluation.

cs.CL↗

arXiv2024

Background Adaptation with Residual Modeling for Exemplar-Free Class-Incremental Semantic Segmentation

Class Incremental Semantic Segmentation~(CISS), within Incremental Learning for semantic segmentation, targets segmenting new categories while reducing the catastrophic forgetting on the old categories.Besides, background shifting, where the background category changes constantly in each step, is a special challenge for CISS. Current methods with a shared background classifier struggle to keep up with these changes, leading to decreased stability in background predictions and reduced accuracy of segmentation. For this special challenge, we designed a novel background adaptation mechanism, which explicitly models the background residual rather than the background itself in each step, and aggregates these residuals to represent the evolving background. Therefore, the background adaptation mechanism ensures the stability of previous background classifiers, while enabling the model to concentrate on the easy-learned residuals from the additional channel, which enhances background discernment for better prediction of novel categories. To precisely optimize the background adaptation mechanism, we propose Pseudo Background Binary Cross-Entropy loss and Background Adaptation losses, which amplify the adaptation effect. Group Knowledge Distillation and Background Feature Distillation strategies are designed to prevent forgetting old categories. Our approach, evaluated across various incremental scenarios on Pascal VOC 2012 and ADE20K datasets, outperforms prior exemplar-free state-of-the-art methods with mIoU of 3.0% in VOC 10-1 and 2.0% in ADE 100-5, notably enhancing the accuracy of new classes while mitigating catastrophic forgetting. Code is available in https://andyzaq.github.io/barmsite/.

cs.CV↗

arXiv2022

Evolutionary-based quantum architecture search

Quantum architecture search (QAS) is desired to construct a powerful and general QAS platform which can significantly accelerate quantum advantages in error-prone and depth limited quantum circuits in today Noisy Intermediate-Scale Quantum (NISQ) era. In this paper, we propose an evolutionary-based quantum architecture search (EQAS) scheme for the optimal layout to balance the higher expressive power and the trainable ability. In EQAS, each layout of quantum circuits, i.e quantum circuit architecture(QCA), is first encoded into a binary string, which is called quantum genes later. Then, an algorithm to remove the redundant parameters in QCA is performed according to the eigenvalues of the corresponding quantum Fisher information matrix (QFIM). Later, each QCA is evaluated by the normalized fitness, so that the sampling rate could be obtained to sample the parent generation by the Roulette Wheel selection strategy. Thereafter, the mutation and crossover are applied to get the next generation. EQAS is verified by the classification task in quantum machine learning for three datasets. The results show that the proposed EQAS can search for the optimal QCA with less parameterized gates, and the higher accuracies are obtained by adopting EQAS for the classification tasks over three dataset.

quant-ph↗

arXiv2022

Measuring the Effect of Training Data on Deep Learning Predictions via Randomized Experiments

We develop a new, principled algorithm for estimating the contribution of training data points to the behavior of a deep learning model, such as a specific prediction it makes. Our algorithm estimates the AME, a quantity that measures the expected (average) marginal effect of adding a data point to a subset of the training data, sampled from a given distribution. When subsets are sampled from the uniform distribution, the AME reduces to the well-known Shapley value. Our approach is inspired by causal inference and randomized experiments: we sample different subsets of the training data to train multiple submodels, and evaluate each submodel's behavior. We then use a LASSO regression to jointly estimate the AME of each data point, based on the subset compositions. Under sparsity assumptions ($k \ll N$ datapoints have large AME), our estimator requires only $O(k\log N)$ randomized submodel trainings, improving upon the best prior Shapley value estimators.

cs.LG↗

arXiv2022

Gradient-Free optimization algorithm for single-qubit quantum classifier

In the paper, a gradient-free optimization algorithm for single-qubit quantum classifier is proposed to overcome the effects of barren plateau caused by quantum devices. A rotation gate RX(ϕ) is applied on a single-qubit binary quantum classifier, and the training data and parameters are loaded into ϕ with the form of vector-multiplication. The cost function is decreased by finding the value of each parameter that yield the minimum expectation value of measuring the quantum circuit. The algorithm is performed iteratively for all parameters one by one, until the cost function satisfies the stop condition. The proposed algorithm is demonstrated for a classification task and is compared with that using Adam optimizer. Furthermore, the performance of the single-qubit quantum classifier with the proposed gradient-free optimization algorithm is discussed when the rotation gate in quantum device is under different noise. The simulation results show that the single-qubit quantum classifier with proposed gradient-free optimization algorithm can reach a high accuracy faster than that using Adam optimizer. Moreover, the proposed gradient-free optimization algorithm can quickly completes the training process of the single-qubit classifier. Additionally, the single-qubit quantum classifier with proposed gradient-free optimization algorithm has a good performance in noisy environments.

quant-ph↗

arXiv2022

Quantum algorithm for neural network enhanced multi-class parallel classification

Using the properties of quantum superposition, we propose a quantum classification algorithm to efficiently perform multi-class classification tasks, where the training data are loaded into parameterized operators which are applied to the basis of the quantum state in quantum circuit composed by \emph{sample register} and \emph{label register}, and the parameters of quantum gates are optimized by a hybrid quantum-classical method, which is composed of a trainable quantum circuit and a gradient-based classical optimizer. After several quantum-to-class repetitions, the quantum state is optimal that the state in \emph{sample register} is the same as that in \emph{label register}. %A structure of loading data many times is performed as a quantum version of neural network to improve the expression ability of quantum circuit. For a classification task of $L$-class, the analysis shows that the space and time complexity of the quantum circuit are $O(L*logL)$ and $O(logL)$, respectively. The numerical simulation results of 2-class task and 5-class task show that the proposed algorithm has a higher classification accuracy, faster convergence and higher expression ability. The classification accuracy and the speed of converging can also be improved by increasing the number times of applying multi-qubit controlled operators on the quantum circuit, especially for multiple classes classification.

quant-ph↗

arXiv2017

Suppression of quantum noises in coherent atom lithography through squeezing

The Abbe's diffraction limit restricts the resolution of an optical imaging and lithography system. Coherent Rabi oscillation is shown to be able to overcome the diffraction limit in both optical and atom lithography. In previous studies, semiclassical theory is applied where the driving field is treated as a classical light and quantum fluctuation is neglected. Here, we show that the quantum fluctuation may reduce the visibility of the superresolution pattern. However, by squeezing the photon number fluctuation we are able to significantly increase its visibility.

quant-ph↗