arXiv subjects

Yuan Sun

Publications and source records attributed to Yuan Sun.

At least 37 records · Page 2Linked to original sources

arXiv2026

Sensing Low-Frequency Field with Rydberg Atoms via Quantum Weak Measurement

Recently, Rydberg atom has emerged as an attractive choice to realize quantum sensing of low-frequency electric field. The progress so far has mostly utilized the intensity and phase changes in probe laser and the corresponding detection mechanism still remains classical. Nevertheless, external field acting on the Rydberg state can induce the polarization variation of probe laser in the Rydberg electromagnetically induced transparency (EIT) system embedded in realistic multi-state atoms. We experimentally observe this phenomenon and realize signal extraction by appropriately utilizing the polarization degrees of freedom. Based on such a mechanism, we further design and implement a quantum weak measurement scheme, which clearly suppresses the technical noise and leads to considerable improvement of performance. Evaluation of the sensitivities across different post-selection angles demonstrates that the weak measurement results agree well with the theoretical model predictions. The advantages of our method are analyzed from multiple aspects, including characterizing the responses over different frequencies and comparing the responses of the weak measurement scheme and the traditional transmission-based method. After accounting for the screening effect of a measured ratio 17\% where the $^\text{87}$Rb atoms experience a substantially reduced field inside the glass cell, the performance reaches 33 $μ\text{V}~\text{cm}^\text{-1}~\text{Hz}^\text{-1/2}$ in sensitivity and 1.0 $μ\text{V/cm}$ in minimal detectable field for an integration time of 1000 s, as perceived by the atoms.

physics.atom-ph↗

arXiv2026

Be Friendly, Not Friends: How LLM Sycophancy Shapes User Trust

LLM-powered conversational agents are increasingly influencing our decision-making, raising concerns about "sycophancy" - the tendency for LLMs to excessively agree with users even at the expense of truthfulness. While prior work has primarily examined LLM sycophancy as a model behavior, our understanding of how users perceive this phenomenon and its impact on user trust remains significantly lacking. In this work, we conceptualize LLM sycophancy along two key constructs: conversational demeanor (complimentary vs. neutral) and stance adaptation (adaptive vs. consistent). A 2 x 2 between-subjects experiment (N = 224) revealed complex dynamics: complimentary LLMs that adapted their stance reduced perceived authenticity and trust, while neutral LLMs that adapted enhanced both, suggesting a pathway for manipulating users into over-trusting LLMs beyond their actual capabilities. Our findings advance user-centric understanding of LLM sycophancy and provide profound implications for developing more ethical and trustworthy LLM systems.

cs.HC↗

arXiv2026

HeadLighter: Disentangling Illumination in Generative 3D Gaussian Heads via Lightstage Captures

Recent 3D-aware head generative models based on 3D Gaussian Splatting achieve real-time, photorealistic and view-consistent head synthesis. However, a fundamental limitation persists: the deep entanglement of illumination and intrinsic appearance prevents controllable relighting. Existing disentanglement methods rely on strong assumptions to enable weakly supervised learning, which restricts their capacity for complex illumination. To address this challenge, we introduce HeadLighter, a novel supervised framework that learns a physically plausible decomposition of appearance and illumination in head generative models. Specifically, we design a dual-branch architecture that separately models lighting-invariant head attributes and physically grounded rendering components. A progressive disentanglement training is employed to gradually inject head appearance priors into the generative architecture, supervised by multi-view images captured under controlled light conditions with a light stage setup. We further introduce a distillation strategy to generate high-quality normals for realistic rendering. Experiments demonstrate that our method preserves high-quality generation and real-time rendering, while simultaneously supporting explicit lighting and viewpoint editing. We will publicly release our code and dataset.

cs.CV↗

arXiv2026

Semantic-Consistent Bidirectional Contrastive Hashing for Noisy Multi-Label Cross-Modal Retrieval

Cross-modal hashing (CMH) facilitates efficient retrieval across different modalities (e.g., image and text) by encoding data into compact binary representations. While recent methods have achieved remarkable performance, they often rely heavily on fully annotated datasets, which are costly and labor-intensive to obtain. In real-world scenarios, particularly in multi-label datasets, label noise is prevalent and severely degrades retrieval performance. Moreover, existing CMH approaches typically overlook the partial semantic overlaps inherent in multi-label data, limiting their robustness and generalization. To tackle these challenges, we propose a novel framework named Semantic-Consistent Bidirectional Contrastive Hashing (SCBCH). The framework comprises two complementary modules: (1) Cross-modal Semantic-Consistent Classification (CSCC), which leverages cross-modal semantic consistency to estimate sample reliability and reduce the impact of noisy labels; (2) Bidirectional Soft Contrastive Hashing (BSCH), which dynamically generates soft contrastive sample pairs based on multi-label semantic overlap, enabling adaptive contrastive learning between semantically similar and dissimilar samples across modalities. Extensive experiments on four widely-used cross-modal retrieval benchmarks validate the effectiveness and robustness of our method, consistently outperforming state-of-the-art approaches under noisy multi-label conditions.

cs.CV↗

arXiv2026

RelayGR: Scaling Long-Sequence Generative Recommendation via Cross-Stage Relay-Race Inference

Real-time recommender systems execute multi-stage cascades (retrieval, pre-processing, fine-grained ranking) under strict tail-latency SLOs, leaving only tens of milliseconds for ranking. Generative recommendation (GR) models can improve quality by consuming long user-behavior sequences, but in production their online sequence length is tightly capped by the ranking-stage P99 budget. We observe that the majority of GR tokens encode user behaviors that are independent of the item candidates, suggesting an opportunity to pre-infer a user-behavior prefix once and reuse it during ranking rather than recomputing it on the critical path. Realizing this idea at industrial scale is non-trivial: the prefix cache must survive across multiple pipeline stages before the final ranking instance is determined, the user population implies cache footprints far beyond a single device, and indiscriminate pre-inference would overload shared resources under high QPS. We present RelayGR, a production system that enables in-HBM relay-race inference for GR. RelayGR selectively pre-infers long-term user prefixes, keeps their KV caches resident in HBM over the request lifecycle, and ensures the subsequent ranking can consume them without remote fetches. RelayGR combines three techniques: 1) a sequence-aware trigger that admits only at-risk requests under a bounded cache footprint and pre-inference load, 2) an affinity-aware router that co-locates cache production and consumption by routing both the auxiliary pre-infer signal and the ranking request to the same instance, and 3) a memory-aware expander that uses server-local DRAM to capture short-term cross-request reuse while avoiding redundant reloads. We implement RelayGR on Huawei Ascend NPUs and evaluate it with real queries. Under a fixed P99 SLO, RelayGR supports up to 1.5$\times$ longer sequences and improves SLO-compliant throughput by up to 3.6$\times$.

cs.DC↗

arXiv2025

Neighbor-aware Instance Refining with Noisy Labels for Cross-Modal Retrieval

In recent years, Cross-Modal Retrieval (CMR) has made significant progress in the field of multi-modal analysis. However, since it is time-consuming and labor-intensive to collect large-scale and well-annotated data, the annotation of multi-modal data inevitably contains some noise. This will degrade the retrieval performance of the model. To tackle the problem, numerous robust CMR methods have been developed, including robust learning paradigms, label calibration strategies, and instance selection mechanisms. Unfortunately, they often fail to simultaneously satisfy model performance ceilings, calibration reliability, and data utilization rate. To overcome the limitations, we propose a novel robust cross-modal learning framework, namely Neighbor-aware Instance Refining with Noisy Labels (NIRNL). Specifically, we first propose Cross-modal Margin Preserving (CMP) to adjust the relative distance between positive and negative pairs, thereby enhancing the discrimination between sample pairs. Then, we propose Neighbor-aware Instance Refining (NIR) to identify pure subset, hard subset, and noisy subset through cross-modal neighborhood consensus. Afterward, we construct different tailored optimization strategies for this fine-grained partitioning, thereby maximizing the utilization of all available data while mitigating error propagation. Extensive experiments on three benchmark datasets demonstrate that NIRNL achieves state-of-the-art performance, exhibiting remarkable robustness, especially under high noise rates.

cs.CV↗

arXiv2025

Sub-Ensemble Correlations as a Covariance Geometry

Conventional practice of spatially resolved detection in diffusion-coupled thermal atomic vapors implicitly treat localized responses as mutually independent. However, in this study, it is shown that observable correlations are governed by the intrinsic spatiotemporal covariance of a global spin-fluctuation field, such that spatial separation specifies only overlapping statistical projections rather than independent physical components. A unified field-theoretic description is established in which sub-ensembles are defined as measurement-induced statistical projections of a single stochastic field. Within this formulation, sub-ensemble correlations are determined by the covariance operator, inducing a natural geometry in which statistical independence corresponds to orthogonality of the measurement functionals. For collective spin fluctuations described by a diffusion-relaxation Ornstein-Uhlenbeck stochastic field, the covariance spectrum admits only a finite set of fluctuation modes in a bounded domain, imposing an intrinsic, field-level limit on the number of statistically distinguishable sub-ensembles. The loss of sub-ensemble independence is formalized through the notion of spatial sampling overlap, which quantifies the unavoidable statistical coupling arising from shared access to common low-order fluctuation modes. While multi-channel atomic magnetometry provides a concrete physical setting in which these constraints become explicit, the framework applies generically to diffusion-coupled stochastic fields.

physics.atom-ph↗

arXiv2025

3-path-connectivity of Cayley graphs generated by wheel graphs

Let $G = (V(G), E(G))$ be a simple connected graph and $Ω$ a subset of $ V(G)$ with $|Ω|\geq2$. An $Ω$-path in $G$ is a path that connects all vertices of $Ω$. Two $Ω$-paths $P_i$ and $P_j$ are said to be internally disjoint if $V(P_i)\cap V(P_j)=Ω$ and $E(P_i)\cap E(P_j)=\emptyset$. Denote $π_G(Ω)$ by the maximum number of internally disjoint $Ω$-paths in $G$. For an integer $k\geq2$, the $k$-path-connectivity $π_k(G)$ of $G$ is defined as $\min\{π_G(Ω)\midΩ\subseteq V(G)$ and $|Ω|=k\}$. Let $CW_n$ denote the Cayley graph generated by the $n$-vertex wheel graph. In this paper, we investigate the $3$-path-connectivity of $CW_n$ and prove that $π_3(CW_n)=\lfloor\frac{6n-9}4\rfloor$ for all $n\geq4$.

math.CO↗

arXiv2025

Updatable Balanced Index for Fast On-device Search with Auto-selection Model

Diverse types of edge data, such as 2D geo-locations and 3D point clouds, are collected by sensors like lidar and GPS receivers on edge devices. On-device searches, such as k-nearest neighbor (kNN) search and radius search, are commonly used to enable fast analytics and learning technologies, such as k-means dataset simplification using kNN. To maintain high search efficiency, a representative approach is to utilize a balanced multi-way KD-tree (BMKD-tree). However, the index has shown limited gains, mainly due to substantial construction overhead, inflexibility to real-time insertion, and inconsistent query performance. In this paper, we propose UnIS to address the above limitations. We first accelerate the construction process of the BMKD-tree by utilizing the dataset distribution to predict the splitting hyperplanes. To make the continuously generated data searchable, we propose a selective sub-tree rebuilding scheme to accelerate rebalancing during insertion by reducing the number of data points involved. We then propose an auto-selection model to improve query performance by automatically selecting the optimal search strategy among multiple strategies for an arbitrary query task. Experimental results show that UnIS achieves average speedups of 17.96x in index construction, 1.60x in insertion, 7.15x in kNN search, and 1.09x in radius search compared to the BMKD-tree. We further verify its effectiveness in accelerating dataset simplification on edge devices, achieving a speedup of 217x over Lloyd's algorithm.

cs.DB↗

arXiv2025

PORTAL: Controllable Landscape Generator for Continuous Optimization-Part I: Framework

Benchmarking is central to optimization research, yet existing test suites for continuous optimization remain limited: classical collections are fixed and rigid, while previous generators cover only narrow families of landscapes with restricted variability and control over details. This paper introduces PORTAL (Platform for Optimization Research, Testing, Analysis, and Learning), a general benchmark generator that provides fine-grained, independent control over basin curvature, conditioning, variable interactions, and surface ruggedness. PORTAL's layered design spans from individual components to block-wise compositions of multi-component landscapes with controllable partial separability and imbalanced block contributions. It offers precise control over the shape of each component in every dimension and direction, and supports diverse transformation patterns through both element-wise and coupling operators with compositional sequencing. All transformations preserve component centers and local quadratic structure, ensuring stability and interpretability. A principled neutralization mechanism prevents unintended component domination caused by exponent or scale disparities, which addresses a key limitation of prior landscape generators. On this foundation, transformations introduce complex landscape characteristics, such as multimodality, asymmetry, and heterogeneous ruggedness, in a controlled and systematic way. PORTAL enables systematic algorithm analysis by supporting both isolation of specific challenges and progressive difficulty scaling. It also facilitates the creation of diverse datasets for meta-algorithmic research, tailored benchmark suite design, and interactive educational use. The complete Python and MATLAB source code for PORTAL is publicly available at [https://github.com/EvoMindLab/PORTAL].

cs.NE↗

arXiv2025

Human-in-the-Loop Generation of Adversarial Texts: A Case Study on Tibetan Script

DNN-based language models excel across various NLP tasks but remain highly vulnerable to textual adversarial attacks. While adversarial text generation is crucial for NLP security, explainability, evaluation, and data augmentation, related work remains overwhelmingly English-centric, leaving the problem of constructing high-quality and sustainable adversarial robustness benchmarks for lower-resourced languages both difficult and understudied. First, method customization for lower-resourced languages is complicated due to linguistic differences and limited resources. Second, automated attacks are prone to generating invalid or ambiguous adversarial texts. Last but not least, language models continuously evolve and may be immune to parts of previously generated adversarial texts. To address these challenges, we introduce HITL-GAT, an interactive system based on a general approach to human-in-the-loop generation of adversarial texts. Additionally, we demonstrate the utility of HITL-GAT through a case study on Tibetan script, employing three customized adversarial text generation methods and establishing its first adversarial robustness benchmark, providing a valuable reference for other lower-resourced languages.

cs.CL↗

arXiv2025

A Unified Approach for Multi-Granularity Search over Spatial Datasets

There has been increased interest in data search as a means to find relevant datasets or data points in data lakes and repositories. Although approaches have been proposed to support spatial dataset search and data point search, they consider the two types of searches independently. To enable search operations ranging from the coarse-grained dataset level to the fine-grained data point level, we provide an integrated one that supports diverse query types and distance metrics. In this paper, we focus on designing a multi-granularity spatial data search system, called Spadas, that supports both dataset and data point search operations. To address the challenges of the high cost of indexing and susceptibility to outliers, we propose a unified index that can drastically improve query efficiency in various scenarios by organizing data reasonably and removing outliers in datasets. Moreover, to accelerate all data search operations, we propose a set of pruning mechanisms based on the unified index, including fast bound estimation, approximation technique with error bound, and pruning in batch techniques, to effectively filter out non-relevant datasets and points. Finally, we report the results of a detailed experimental evaluation using six spatial data repositories, achieving orders of magnitude faster than the state-of-the-art algorithms and demonstrating the effectiveness by case study. An online spatial data search system of Spadas is also implemented and made accessible to users.

cs.DB↗

arXiv2025

Refine coherent control of atomic qubits via wave-function approach conditioned on no-decay

As a fundamental phenomenon in quantum systems, spontaneous emission constitutes an inevitable source of error, which ultimately degrades the fidelity of quantum logic gates. A successful quantum logic gate needs to operate on the condition that no decay event, such as spontaneous emission, occurs. Such successes can be ensured by post-selection based on syndrome extraction according to the theory of quantum error correction or quantum error mitigation. In this case, the wave function of qubits remains a pure state but is subject to additional influences from spontaneous emission, even without actual decay events. Therefore, such a process must be appropriately described by a modified version of Schrödinger equation for the dynamics conditioned on no-decay. Calculations reveal that this effect must be seriously taken into consideration for the design of high-fidelity quantum logic gates. With respect to realistic experimental conditions, even if the coherence is well preserved, improving the fidelity of manipulating physical qubits requires careful consideration of the subtle influences of decay processes such as spontaneous emission. Specifically, the gate and readout processes in the atomic qubit platform are discussed.

quant-ph↗

arXiv2025

PG-CE: A Progressive Generation Dataset with Constraint Enhancement for Controllable Text Generation

With the rapid development of Large Language Models (LLMs), Controllable Text Generation (CTG) has become a critical technology for enhancing system reliability and user experience. Addressing the limitations of traditional methods, this paper proposes the PG-CE (Progressive Generation with Constraint Enhancement) approach, which decomposes CTG tasks into three steps: type prediction, constraint construction, and guided generation. This method employs constraint generation models to dynamically build multi-dimensional constraints including tone, expression style, and thematic focus to guide output. Experiments demonstrate that PG-CE significantly improves generation quality across multiple scenarios while maintaining text controllability, thematic relevance, and response practicality. The research developed a dataset containing 90,000 constraint-text pairs (with an 8:2 ratio between daily and other topics), effectively reflecting real-world application requirements.

cs.CL↗

arXiv2025

Enhancing Cross-Lingual Transfer through Reversible Transliteration: A Huffman-Based Approach for Low-Resource Languages

As large language models (LLMs) are trained on increasingly diverse and extensive multilingual corpora, they demonstrate cross-lingual transfer capabilities. However, these capabilities often fail to effectively extend to low-resource languages, particularly those utilizing non-Latin scripts. While transliterating low-resource languages into Latin script presents a natural solution, there currently lacks a comprehensive framework for integrating transliteration into LLMs training and deployment. Taking a pragmatic approach, this paper innovatively combines character transliteration with Huffman coding to design a complete transliteration framework. Our proposed framework offers the following advantages: 1) Compression: Reduces storage requirements for low-resource language content, achieving up to 50% reduction in file size and 50-80% reduction in token count. 2) Accuracy: Guarantees 100% lossless conversion from transliterated text back to the source language. 3) Efficiency: Eliminates the need for vocabulary expansion for low-resource languages, improving training and inference efficiency. 4) Scalability: The framework can be extended to other low-resource languages. We validate the effectiveness of our framework across multiple downstream tasks, including text classification, machine reading comprehension, and machine translation. Experimental results demonstrate that our method significantly enhances the model's capability to process low-resource languages while maintaining performance on high-resource languages. Our data and code are publicly available at https://github.com/CMLI-NLP/HuffmanTranslit.

cs.CL↗

arXiv2025

CUTE: A Multilingual Dataset for Enhancing Cross-Lingual Knowledge Transfer in Low-Resource Languages

Large Language Models (LLMs) demonstrate exceptional zero-shot capabilities in various NLP tasks, significantly enhancing user experience and efficiency. However, this advantage is primarily limited to resource-rich languages. For the diverse array of low-resource languages, support remains inadequate, with the scarcity of training corpora considered the primary cause. We construct and open-source CUTE Chinese, Uyghur, Tibetan,English dataset, consisting of two 25GB sets of four-language corpora (one parallel and one non-parallel), obtained through machine translation. CUTE encompasses two resource-rich languages (Chinese and English) and two low-resource languages (Uyghur and Tibetan). Prior to constructing CUTE, human assessment validates that the machine translation quality between Chinese-Uyghur and Chinese-Tibetan approaches that of Chinese-English translation. CUTE represents the largest open-source corpus for Uyghur and Tibetan languages to date, and we demonstrate its effectiveness in enhancing LLMs' ability to process low-resource languages while investigating the role of corpus parallelism in cross-lingual transfer learning. The CUTE corpus and related models are made publicly available to the research community.

cs.CL↗

arXiv2025

MambaFlow: A Mamba-Centric Architecture for End-to-End Optical Flow Estimation

Recently, the Mamba architecture has demonstrated significant successes in various computer vision tasks, such as classification and segmentation. However, its application to optical flow estimation remains unexplored. In this paper, we introduce MambaFlow, a novel framework designed to leverage the high accuracy and efficiency of the Mamba architecture for capturing locally correlated features while preserving global information in end-to-end optical flow estimation. To our knowledge, MambaFlow is the first architecture centered around the Mamba design tailored specifically for optical flow estimation. It comprises two key components: (1) PolyMamba, which enhances feature representation through a dual-Mamba architecture, incorporating a Self-Mamba module for intra-token modeling and a Cross-Mamba module for inter-modality interaction, enabling both deep contextualization and effective feature fusion; and (2) PulseMamba, which leverages an Attention Guidance Aggregator (AGA) to adaptively integrate features with dynamically learned weights in contrast to naive concatenation, and then employs the intrinsic recurrent mechanism of Mamba to perform autoregressive flow decoding, facilitating efficient flow information dissemination. Extensive experiments demonstrate that MambaFlow achieves remarkable results comparable to mainstream methods on benchmark datasets. Compared to SEA-RAFT, MambaFlow attains higher accuracy on the Sintel benchmark, demonstrating stronger potential for real-world deployment on resource-constrained devices. The source code will be made publicly available upon acceptance of the paper.

cs.CV↗

arXiv2025

Systematic approach to correlators in $T\bar{T}$ deformed CFTs

We investigate higher-order corrections to correlators in a general CFT (conformal field theory) with the double-trace $T\bar{T}$ deformation. Standard perturbation theory proves inadequate for this problem due to the intricate stress-tensor flow induced by the deformation. To tackle this challenge, we introduce a novel technique termed the conservation equation method. This method leverages the stress tensor's trace relation and conservation property to establish relationships between higher and lower-order corrections, and subsequently determine the correlators by enforcing symmetry properties. As an illustration, we compute first- and higher-order corrections to various types of correlators. Our results align with existing calculations in the literature and demonstrate the viability of our method for general CFTs.

hep-th↗