Searcharxiv⌕ Search

arXiv subjects

Xinyuan Zhang

Publications and source records attributed to Xinyuan Zhang.

At least 37 records · Page 2Linked to original sources
arXiv2026

Neural Certificate Pricing for Combinatorial Optimization Problems

Combinatorial optimization (CO) problems are difficult because certifiable discrete structure induces exponential search. One needs to search over the set exponentially many candidates to certify optimality, however, the structural feasibility of a path, packing, or cover can be verified in polynomial time once supplied. In this study, we introduce Neural Certificate Pricing (NCP) that exploits this asymmetry under an unsupervised learning framework. A neural network is trained to predict certificate-level dual prices, while a structured recovery layer constructs the induced primal marginal. NCP can be viewed as amortized separation: instead of enumerating violated inequalities, it learns the residual prices through which their aggregate effect enters recovery. When the certificate-consistency condition holds, the recovered marginal is globally feasible, and a local theory shows that first-order errors in the predicted price induce only second-order loss in objective value. Across three classes of CO problems, NCP either outperforms state-of-the-art neural baselines by large margins or matches them at a fraction of the computation time, and shows stronger out-of-distribution generalization.

cs.LG↗
arXiv2026

Half-Moon Cookie: Private, Similarity-Based Blocklisting with TOCTOU-Attack Resilience

Blocklisting is a common technique for preventing the use of known malicious content. However, conventional blocklisting infrastructures require either the blocklist to be public or clients to reveal their queries to the blocklist server. We introduce a private blocklisting framework, Half-Moon Cookie, by which a client can check an item against a proprietary blocklist held by a server, to determine whether the item is close to any blocklist element in a metric space. Critically, our design separates the embedding step from the blocklist check, enabling independent choice of methods to compute them privately and efficiently. Still, this computation might be too costly to perform on the critical path of using the item, and so our design also supports a very efficient check that an item previously passed the blocklist check. In doing so, we support applications where one client can perform the blocklist check on the item before sending it, and recipients can more efficiently confirm the previous result before using the item, thereby avoiding TOCTOU attacks. We show how Half-Moon Cookie can be instantiated for similarity-based malware detection, to block malicious executables without revealing client inputs or disclosing the blocklist.

cs.CR↗
arXiv2026

Simulating Gaussian boson sampling on graphs in polynomial time

We show that a distribution related to Gaussian Boson Sampling (GBS) on graphs can be sampled classically in polynomial time. Graphical applications of GBS typically sample from this distribution, and thus quantum algorithms do not provide exponential speedup for these applications. We also show that another distribution related to Boson sampling can be sampled classically in polynomial time.

quant-ph↗
arXiv2026

Rethinking Sales Lead Scoring with LLM-based Hierarchical Preference Ranking

Sales lead conversion in high-stakes domains (e.g., automotive, real estate) differs fundamentally from e-commerce recommendation due to prolonged decision cycles and multi-stage funnels. Traditional lead scoring methods rule-based scorecards, machine learning, or pointwise CTR models face severe challenges: sparse supervision, a semantic gap in unstructured CRM logs, and inability to capture relative lead priority. While Large Language Models(LLMs) offer superior semantic understanding of customer interactions, general-purpose LLMs are ill-suited for lead ranking: they generate text rather than comparable scores, and lack alignment with the hierarchical priorities of sales funnels. We introduce an LLM-based discriminative framework for sales lead scoring, which supports joint modeling of structured CRM features and unstructured customer interactions. On top of this framework, we propose HPRO (Hierarchical Preference Ranking Optimization), which augments sales lead scoring with a hierarchical preference ranking objective. HPRO employs a margin-aware Bradley-Terry formulation to transform sparse binary labels into dense, funnel-aware preference pairs, enabling lead scoring to leverage both pointwise and pairwise supervision. Experiments on large-scale data from a leading NEV brand demonstrate state-of-the-art classification (AUC 0.8161) and ranking performance (+39.7% precision among top-ranked leads). A 132-day online A/B test validates 9.5% sales volume uplift, confirming real-world commercial impact.

cs.IR↗
arXiv2026

From Profiles to Steering Vectors: Global Sparse Priors and Local Semantic Calibration for Personalized Text Generation

Personalized text generation requires models to capture user-specific writing styles from historical data. Existing approaches based on retrieval, parameter-efficient fine-tuning, or activation steering either introduce inference and storage overhead or struggle to separate stylistic signals from semantic content. We propose GLASS, a training-free framework for personalized generation via Global-Local Activation Steering with Sparse priors. GLASS uses sparse autoencoders to extract a global user-style prior from historical responses and constructs local contrastive style vectors over clustered interaction scenarios. During inference, it jointly injects global and local vectors into different model layers, enabling context-aware personalization without retrieval or parameter updates. Experiments on LaMP and LongLaMP show that GLASS outperforms retrieval-, fine-tuning-, and steering-based baselines across ROUGE metrics and LLM-as-judge evaluations. Further analyses show that SAE-based representations are more robust to topic and length shifts, suggesting better disentanglement of stylistic information from semantic residue.

cs.AI↗
arXiv2026

Edge-Tilting Field Dynamics: Rapid Mixing at the Uniqueness Threshold and Optimal Mixing for Swendsen-Wang Dynamics

We prove two results on the mixing times of Markov chains for two-spin systems. First, we show that the Glauber dynamics mixes in polynomial time for the Gibbs distributions of antiferromagnetic two-spin systems at the critical threshold of the uniqueness phase transition of the Gibbs measure on infinite regular trees. This completes the computational phase transition picture for antiferromagnetic two-spin systems, which includes near-linear-time optimal mixing in the uniqueness regime [Chen--Liu--Vigoda, STOC '21; Chen--Feng--Yin--Zhang, FOCS '22], NP-hardness of approximate sampling in the non-uniqueness regime [Sly--Sun, FOCS '12], and polynomial-time mixing at criticality (this work). Second, we prove an optimal $O(\log n)$ mixing time bound as well as an optimal $Ω(1)$ spectral gap for the Swendsen--Wang dynamics for the ferromagnetic Ising model with an external field on bounded-degree graphs. To the best of our knowledge, these are the first sharp bounds on the mixing rate of this classical global Markov chain beyond mean-field or strong spatial mixing (SSM) regimes, and resolve a conjecture of [Feng--Guo--Wang, IANDC '23]. A key ingredient in both proofs is a new family of localization schemes that extends the field dynamics of [Chen--Feng--Yin--Zhang, FOCS '21] by tilting general edge (or hyperedge) weights rather than vertex fields. This framework, which subsumes the classical Swendsen--Wang dynamics as a special case, extends the localization framework of [Chen--Eldan, FOCS '22] beyond stochastic and field localizations, and enables controlled tilting of interaction strengths while preserving external fields.

cs.DS↗
arXiv2026

Subquadratic Counting via Perfect Marginal Sampling

We study the computational complexity of approximately computing the partition function of a spin system. Techniques based on standard counting-to-sampling reductions yield $\tilde{O}(n^2)$-time algorithms, where $n$ is the size of the input graph. We present new counting algorithms that break the quadratic-time barrier in a wide range of settings. For example, for the hardcore model of $λ$-weighted independent sets in graphs of maximum degree $Δ$, we obtain a $\tilde{O}(n^{2-δ})$-time approximate counting algorithm, for some constant $δ> 0$, when the fugacity $λ< \frac{1}{Δ-1}$, improving over the previous regime of $λ= o(Δ^{-3/2})$ by Anand, Feng, Freifeld, Guo, and Wang (2025). Our results apply broadly to many other spin systems, such as the Ising model, hypergraph independent sets, and vertex colorings. Interestingly, our work reveals a deep connection between $\textit{subquadratic}$ counting and $\textit{perfect}$ marginal sampling. For two-spin systems such as the hardcore and Ising models, we show that the existence of perfect marginal samplers directly yields subquadratic counting algorithms in a $\textit{black-box}$ fashion. For general spin systems, we show that almost all existing perfect marginal samplers can be adapted to produce a sufficiently low-variance marginal estimator in sublinear time, leading to subquadratic counting algorithms.

cs.DS↗
arXiv2026

AgentRaft: Automated Detection of Data Over-Exposure in LLM Agents

The rapid integration of Large Language Model (LLM) agents into autonomous task execution has introduced significant privacy concerns within cross-tool data flows. In this paper, we systematically investigate and define a novel risk termed Data Over-Exposure (DOE) in LLM Agent, where an Agent inadvertently transmits sensitive data beyond the scope of user intent and functional necessity. We identify that DOE is primarily driven by the broad data paradigms in tool design and the coarse-grained data processing inherent in LLMs. In this paper, we present AgentRaft, the first automated framework for detecting DOE risks in LLM agents. AgentRaft combines program analysis with semantic reasoning through three synergistic modules: (1) it constructs a Cross-Tool Function Call Graph (FCG) to model the interaction landscape of heterogeneous tools; (2) it traverses the FCG to synthesize high-quality testing user prompts that act as deterministic triggers for deep-layer tool execution; and (3) it performs runtime taint tracking and employs a multi-LLM voting committee grounded in global privacy regulations (e.g., GDPR, CCPA, PIPL) to accurately identify privacy violations. We evaluate AgentRaft on a testing environment of 6,675 real-world agent tools. Our findings reveal that DOE is indeed a systemic risk, prevalent in 57.07% of potential tool interaction paths. AgentRaft achieves a high detection accuracy and effectiveness, outperforming baselines by 87.24%. Furthermore, AgentRaft reaches near-total DOE coverage (99%) within only 150 prompts while reducing per-chain verification costs by 88.6%. Our work provides a practical foundation for building auditable and privacy-compliant LLM agent systems.

cs.SE↗
arXiv2026

MiniCPM-SALA: Hybridizing Sparse and Linear Attention for Efficient Long-Context Modeling

The evolution of large language models (LLMs) towards applications with ultra-long contexts faces challenges posed by the high computational and memory costs of the Transformer architecture. While existing sparse and linear attention mechanisms attempt to mitigate these issues, they typically involve a trade-off between memory efficiency and model performance. This paper introduces MiniCPM-SALA, a 9B-parameter hybrid architecture that integrates the high-fidelity long-context modeling of sparse attention (InfLLM-V2) with the global efficiency of linear attention (Lightning Attention). By employing a layer selection algorithm to integrate these mechanisms in a 1:3 ratio and utilizing a hybrid positional encoding (HyPE), the model maintains efficiency and performance for long-context tasks. Furthermore, we introduce a cost-effective continual training framework that transforms pre-trained Transformer-based models into hybrid models, which reduces training costs by approximately 75% compared to training from scratch. Extensive experiments show that MiniCPM-SALA maintains general capabilities comparable to full-attention models while offering improved efficiency. On a single NVIDIA A6000D GPU, the model achieves up to 3.5x the inference speed of the full-attention model at the sequence length of 256K tokens and supports context lengths of up to 1M tokens, a scale where traditional full-attention 8B models fail because of memory constraints.

cs.CL↗
arXiv2026

Unsupervised Denoising of Diffusion-Weighted Images with Bias and Variance Corrected Noise Modeling

Diffusion magnetic resonance imaging (dMRI) plays a vital role in both clinical diagnostics and neuroscience research. However, its inherently low signal-to-noise ratio (SNR), especially under high diffusion weighting, significantly degrades image quality and impairs downstream analysis. Recent self-supervised and unsupervised denoising methods offer a practical solution by enhancing image quality without requiring clean references. However, most of these methods do not explicitly account for the non-Gaussian noise characteristics commonly present in dMRI magnitude data during the supervised learning process, potentially leading to systematic bias and heteroscedastic variance, particularly under low-SNR conditions. To overcome this limitation, we introduce noise-corrected training objectives that explicitly model Rician statistics. Specifically, we propose two alternative loss functions: one derived from the first-order moment to remove mean bias, and another from the second-order moment to correct squared-signal bias. Both losses include adaptive weighting to account for variance heterogeneity and can be used without changing the network architecture. These objectives are instantiated in an image-specific, unsupervised Deep Image Prior (DIP) framework. Comprehensive experiments on simulated and in-vivo dMRI show that the proposed losses effectively reduce Rician bias and suppress noise fluctuations, yielding higher image quality and more reliable diffusion metrics than state-of-the-art denoising baselines. These results underscore the importance of bias- and variance-aware noise modeling for robust dMRI analysis under low-SNR conditions.

q-bio.QM↗
arXiv2026

Trustworthy and Fair SkinGPT-R1 for Democratizing Dermatological Reasoning across Diverse Ethnicities

The clinical translation of dermatological AI is hindered by opaque reasoning and systematic performance disparities across skin tones. Here we present SkinGPT-R1, a multimodal large language model that integrates chain-of-thought diagnostic reasoning with a fairness-aware mixture-of-experts architecture for interpretable and equitable skin disease diagnosis. Through parameter-efficient adaptation of a frozen reasoning backbone, SkinGPT-R1 generates structured diagnostic reports comprising visual findings, differential reasoning, and final diagnosis. Across seven external datasets spanning diverse pathologies and imaging conditions, SkinGPT-R1 achieves state-of-the-art accuracy on six benchmarks, including 82.50\% on a challenging 40-class long-tail classification task (+19.30\% over leading baselines). Blinded evaluation by five board-certified dermatologists on 1,000 phenotypically balanced cases yields a mean score of 3.6 out of 5, with the highest ratings in safety (3.8) and reasoning coherence (3.6), indicating that the generated rationales are clinically safe, logically grounded, and suitable for supporting diagnostic decision-making. Critically, SkinGPT-R1 mitigates algorithmic bias across the full Fitzpatrick spectrum, achieving a robust worst-group performance of 41.40\% on the Fitz17k benchmark and a five-fold relative improvement in lower-bound accuracy on the DDI dataset compared to standard multimodal baselines. These results establish a framework for trustworthy, fair, and explainable AI-assisted dermatological diagnosis.

cs.CV↗
arXiv2026

Unified Multimodal and Multilingual Retrieval via Multi-Task Learning with NLU Integration

Multimodal retrieval systems typically employ Vision Language Models (VLMs) that encode images and text independently into vectors within a shared embedding space. Despite incorporating text encoders, VLMs consistently underperform specialized text models on text-only retrieval tasks. Moreover, introducing additional text encoders increases storage, inference overhead, and exacerbates retrieval inefficiencies, especially in multilingual settings. To address these limitations, we propose a multi-task learning framework that unifies the feature representation across images, long and short texts, and intent-rich queries. To our knowledge, this is the first work to jointly optimize multilingual image retrieval, text retrieval, and natural language understanding (NLU) tasks within a single framework. Our approach integrates image and text retrieval with a shared text encoder that is enhanced by NLU features for intent understanding and retrieval accuracy.

cs.IR↗
arXiv2026

Honesty-Aware Multi-Agent Framework for High-Fidelity Synthetic Data Generation in Digital Psychiatric Intake Doctor-Patient Interactions

Data scarcity and unreliable self-reporting -- such as concealment or exaggeration -- pose fundamental challenges to psychiatric intake and assessment. We propose a multi-agent synthesis framework that explicitly models patient deception to generate high-fidelity, publicly releasable synthetic psychiatric intake records. Starting from DAIC-WOZ interviews, we construct enriched patient profiles and simulate a four-role workflow: a \emph{Patient} completes self-rated scales and participates in a semi-structured interview under a topic-dependent honesty state; an \emph{Assessor} selects instruments based on demographics and chief complaints; an \emph{Evaluator} conducts the interview grounded in rater-administered scales, tracks suspicion, and completes ratings; and a \emph{Diagnostician} integrates all evidence into a diagnostic summary. Each case links the patient profile, self-rated and rater-administered responses, interview transcript, diagnostic summary, and honesty state. We validate the framework through four complementary evaluations: diagnostic consistency and severity grading, chain-of-thought ablations, human evaluation of clinical realism and dishonesty modeling, and LLM-based comparative evaluation. The resulting corpus spans multiple disorders and severity levels, enabling controlled study of dishonesty-aware psychiatric assessment and the training and evaluation of adaptive dialogue agents.

cs.DB↗
arXiv2026

Rapid Mixing at the Uniqueness Threshold

Over the past decades, a fascinating computational phase transition has been identified in sampling from Gibbs distributions. Though, the computational complexity at the critical point remains poorly understood, as previous algorithmic and hardness results all required a constant slack from this threshold. In this paper, we resolve this open question at the critical phase transition threshold, thus completing the picture of the computational phase transition. We show that for the hardcore model on graphs with maximum degree $Δ\ge 3$ at the uniqueness threshold $λ= λ_c(Δ)$, the mixing time of Glauber dynamics is upper bounded by a polynomial in $n$, but is not nearly linear in the worst case. For the Ising model (either antiferromagnetic or ferromagnetic), we establish similar results. For the Ising model on graphs with maximum degree $Δ\ge 3$ at the critical temperature $β$ where $|β| = β_c(Δ)$, with the tree-uniqueness threshold $β_c(Δ)$, we show that the mixing time of Glauber dynamics is upper bounded by $\tilde{O}\left(n^{3 + O(1/Δ)}\right)$ and lower bounded by $Ω\left(n^{3/2}\right)$ in the worst case. For the Ising model specified by a critical interaction matrix $J$ with $\left \lVert J \right \rVert_2=1$, we obtain an upper bound $\tilde{O}(n^{3/2})$ for the mixing time, matching the lower bound $Ω\left(n^{3/2}\right)$ on the complete graph up to a logarithmic factor. Our mixing time upper bounds are derived from a new interpretation and analysis of the localization scheme method introduced by Chen and Eldan (2022), applied to the field dynamics for the hardcore model and the proximal sampler for the Ising model. As key steps in both our upper and lower bounds, we establish sub-linear upper and lower bounds for spectral independence at the critical point for worst-case instances.

cs.DS↗
arXiv2025

Rapid Mixing on Random Regular Graphs beyond Uniqueness

The hardcore model is a fundamental probabilistic model extensively studied in statistical physics, probability theory, and computer science. For graphs of maximum degree $Δ$, a well-known computational phase transition occurs at the tree-uniqueness threshold $λ_c(Δ) = \frac{(Δ-1)^{Δ-1}}{(Δ-2)^Δ}$, where the mixing behavior of the Glauber dynamics (a simple Markov chain) undergoes a sharp transition. It is conjectured that random regular graphs exhibit different mixing behavior, with the slowdown occurring far beyond the uniqueness threshold. We confirm this conjecture by showing that, for the hardcore model on random $Δ$-regular graphs, the Glauber dynamics mixes rapidly with high probability when $λ= O(1/\sqrtΔ)$, which is significantly beyond the uniqueness threshold $λ_c(Δ) \approx e/Δ$. Our result establishes a sharp distinction between the hardcore model on worst-case and beyond-worst-case instances, showing that the worst-case and average-case complexities of sampling and counting are fundamentally different. This result of rapid mixing on random instances follows from a new criterion we establish for rapid mixing of Glauber dynamics for any distribution supported on a downward closed set family. Our criterion is simple, general, and easy to check. In addition to proving new mixing conditions for the hardcore model, we also establish improved mixing time bounds for sampling uniform matchings or $b$ matchings on graphs, the random cluster model on matroids with $q \in [0,1)$, and the determinantal point process. Our proof of this new criterion for rapid mixing combines and generalizes several recent tools in a novel way, including a trickle down theorem for field dynamics, spectral/entropic stability, and a new comparison result between field dynamics and Glauber dynamics.

cs.DS↗
arXiv2025

A Plug-and-Play Framework for Volumetric Light-Sheet Image Reconstruction

Cardiac contraction is a rapid, coordinated process that unfolds across three-dimensional tissue on millisecond timescales. Traditional optical imaging is often inadequate for capturing dynamic cellular structure in the beating heart because of a fundamental trade-off between spatial and temporal resolution. To overcome these limitations, we propose a high-performance computational imaging framework that integrates Compressive Sensing (CS) with Light-Sheet Microscopy (LSM) for efficient, low-phototoxic cardiac imaging. The system performs compressed acquisition of fluorescence signals via random binary mask coding using a Digital Micromirror Device (DMD). We propose a Plug-and-Play (PnP) framework, solved using the alternating direction method of multipliers (ADMM), which flexibly incorporates advanced denoisers, including Tikhonov, Total Variation (TV), and BM3D. To preserve structural continuity in dynamic imaging, we further introduce temporal regularization enforcing smoothness between adjacent z-slices. Experimental results on zebrafish heart imaging under high compression ratios demonstrate that the proposed method successfully reconstructs cellular structures with excellent denoising performance and image clarity, validating the effectiveness and robustness of our algorithm in real-world high-speed, low-light biological imaging scenarios.

cs.CV↗
arXiv2025

MapComp: A Secure View-based Collaborative Analytics Framework for Join-Group-Aggregation

Join-group-aggregation (JGA) queries are fundamental to data analytics, yet executing them collaboratively across different parties poses significant privacy risks. Secure multi-party computation (MPC) offers a cryptographic solution. However, existing MPC-based JGA approaches consider only a one-time query paradigm and suffer from significant performance bottlenecks. It executes expensive join operations from scratch across multiple queries and employs inefficient group-aggregation (GA) protocols, both of which hinder their practical use for scalable, real-time analysis. This paper introduces MapComp, a novel view-based framework to facilitate JGA queries for secure collaborative analytics. Through specially crafted materialized views for join and novel design of GA protocols, MapComp removes duplicate join workload and expedites subsequent GA, improving the efficiency of JGA query execution. To address the challenge of continuous data updates, our materialized view offers payload-independence feature and provides significant efficiency improvements in view refreshing with free MPC overhead. This feature, on the other hand, also allows further acceleration for GA, where we devise multiple novel protocols that outperform prior works. Notably, our work represents the first endeavor to expedite secure collaborative JGA queries using materialized views. Our rigorous experiments demonstrate a significant advantage of MapComp, achieving up to a 308.9x improvement in efficiency over the baseline in real-world query simulations. Moreover, our optimized GA protocols achieve up to a 1140.5x improvement compared to prior oblivious sorting-based solutions.

cs.CR↗
arXiv2025

dInfer: An Efficient Inference Framework for Diffusion Language Models

Diffusion-based large language models (dLLMs) have emerged as a promising alternative to autoregressive (AR) LLMs, leveraging denoising-based generation to enable inherent parallelism. Even more and more open-sourced dLLM models emerge, yet their widespread adoption remains constrained by the lack of a standardized and efficient inference framework. We present dInfer, an efficient and extensible framework for dLLM inference. dInfer decomposes the inference pipeline into four modular components--model, diffusion iteration manager, decoding strategy, and KV-cache manager--and integrates novel algorithms for each component alongside system-level optimizations. Through this combination of algorithmic innovations and system enhancements, dInfer achieves substantial efficiency gains without compromising output quality on LLaDA-MoE. At batch size 1, it surpasses 1,100 tokens per second on HumanEval and averages over 800 tokens per second across six benchmarks on $8\times$ H800 GPUs. Compared to prior systems, dInfer delivers a $10\times$ speedup over Fast-dLLM while maintaining similar model performance. Even compared to the AR model (with a comparable number of activation parameters and performance) QWen2.5-3B, which is highly optimized with the latest vLLM inference engine, dInfer still delivers a $2$-$3\times$ speedup. The implementation of dInfer is open-sourced at https://github.com/inclusionAI/dInfer.

cs.CL↗