arXiv subjects

Lei Tang

Publications and source records attributed to Lei Tang.

At least 19 recordsLinked to original sources

arXiv2026

Mint-Agent: Introducing Finance-Native Agentic Foundation Models

Financial agents must do more than recall domain knowledge: they must be both reliable, executing precise operations over grounded evidence, and executive, sustaining long-horizon research whose conclusions remain auditable. We present Mint-Agent, a family of finance-native agentic models designed around these two scales of financial intelligence. Mint-Agent is built upon three pillars: data, harness, and algorithm. Our data engine constructs clean, specialized tasks for atomic financial capabilities and long-horizon agentic execution from real-world financial sources. MintHarness enables stable interaction with open-ended environments and maintains auditable evidence trails across extended research trajectories. Our training recipe combines SFT, critical-step OPD, and RLVR to develop separate financial reasoning and agentic execution experts, which are then unified through model merging and multi-teacher on-policy distillation into compact, general-purpose financial agents. This pipeline yields two flagship models, Mint-Cu (9B) and Mint-Ag (27B). Across professional financial benchmarks, our models demonstrate two defining strengths: (1) Reliability: Mint-Ag achieves 98.33% on RFC-Bench, surpassing GPT-5.6-Sol and Claude-Opus-4.8 by 3.66 and 3.00 points; and (2) Executability: Mint-Cu reaches 69.86% on FinSearchComp T2, outperforming Agents-A1-35B and Nex-N2-mini by 22.83 and 12.78 points, while Mint-Ag achieves 76.00% and 60.49% on FinanceAgentBench v1.1 and v2, respectively. These results establish a path toward trustworthy financial intelligence in which domain expertise, long-horizon execution, and auditable evidence are jointly engineered as a unified foundation for frontier agentic models.

cs.CL↗

arXiv2026

ISPCloak: Weaponizing ISP for Optimization-Free Physical Camouflage against Deepfake Detectors

The rapid advancement of generative models has spurred the critical need to evaluate the worst-case robustness of deepfake detectors. In this paper, we reveal a fundamental blind spot in current forensic paradigms: while existing detectors excel at capturing digital synthesis artifacts, their effectiveness drops drastically when AI-generated content is cloaked in authentic physical imaging characteristics. We posit that genuine photographs inherently possess hardware-intrinsic statistical signatures, which are imperceptible footprints imprinted by optical sensors and Image Signal Processing (ISP) pipelines, and are fundamentally absent in purely data-driven generative models. Driven by this insight, we propose ISPCloak, a novel optimization-free adversarial attack framework that explicitly weaponizes the ISP pipeline to mislead the judgment of deepfake detectors. Rather than relying on computationally expensive gradient perturbations, our method first employs an Invertible ISP network to project images into the RAW domain. Then, we seamlessly imprint the complex statistical priors of real cameras onto AI-generated images by injecting realistic Poisson-Gaussian sensor noise and conducting forward ISP reconstruction. Synergized with generative artifact suppression and adaptive masking, this streamlined physical simulation enables ultra-fast generation of adversarial examples. Extensive experiments show that embedding authentic physical perturbations fundamentally disrupts a broad range of current detection mechanisms, yielding universally evasive adversarial examples with imperceptible visual alterations.

cs.CV↗

arXiv2026

MagicSelector: Joint Optimization for Agent Tool Selection via Counterfactual Decomposition and Progressive Reranking

We present MagicSelector, a joint optimization framework integrating Counterfactual task decomposition, Progressive reranking, and Dynamic Top-K, designed to address the fundamental challenges of tool retrieval in agents. MagicSelector is a specialized framework capable of translating ambiguous user instructions into executable atomic subtasks and guiding high-precision tool retrieval, effectively mitigating redundant noise and severe context distraction in out-of-domain (OOD) scenarios. We empower MagicSelector with these capabilities through three key contributions: (1) a preference-guided counterfactual task decomposition mechanism that utilizes a counterfactual reward to quantify the marginal causal gain of decomposition on retrieval ranking, effectively imposing fine-grained structural supervision on logical coherence; (2) a progressive tool reranking method driven by self-distillation hard negative mining, which optimizes both point-wise and list-wise relevance to enhance fine-grained discrimination among highly similar tools; and (3) a dual semantic boundary-aware dynamic Top-K strategy that adaptively monitors reranking score cliffs and inter-tool semantic shifts to dynamically truncate the candidate list, maximizing relevant tool recall while filtering long-tail noise. Evaluated on MTDTool, the first task decomposition benchmark we constructed tailored for mobile multi-turn interactions with process-level annotations, MagicSelector yields promising performance. Extensive experiments demonstrate that MagicSelector significantly outperforms state-of-the-art methods in terms of tool retrieval accuracy, OOD generalization capability, and overall token efficiency, thereby demonstrating the effectiveness of our proposed framework.

cs.IR↗

arXiv2026

PCTD: Preference-Guided Counterfactual Task Decomposition for Agent Tool Retrieval

Task decomposition aims to transform ambiguous instructions into executable atomic subtasks, thereby guiding high-precision tool retrieval. However, our analysis reveals that directly adopting tool retrieval metrics, i.e., Recall or NDCG, as rewards for task decomposition can easily induce reward hacking in reinforcement learning-based methods. Specifically, models tend to maximize retrieval matching through strategies such as repetitive decomposition. This spurious correlation between the shallow features of decomposition results and retrieval metric impairs generalization in Out-of-Domain (OOD) scenarios involving unseen tools. To address this issue, we propose PCTD, a Preference-guided Counterfactual Task Decomposition framework. PCTD quantifies the marginal causal gain of decomposition on retrieval ranking through a counterfactual reward, thereby cutting off spurious correlations at their source. Meanwhile, it introduces a preference reward to impose fine-grained structural supervision on logical coherence and atomicity, encouraging the model to generate high-quality decompositions. In addition, we construct MTDTool, the task decomposition benchmark specifically designed for mobile multi-turn interactions. Extensive experiments demonstrate that PCTD alleviates repetitive decomposition and surpasses SOTA methods in retrieval, decomposition quality, and OOD generalization.

cs.IR↗

arXiv2026

MalSkillBench: A Runtime-Verified Benchmark of Malicious Agent Skills

AI coding agents such as Claude Code and Gemini CLI increasingly extend themselves with third-party skills: markdown packages bundling natural-language instructions, executable scripts, and tool permissions. Because a skill is at once code and agent-facing instruction, it introduces a supply chain dependency whose risk is neither pure code nor pure prompt. Detection tools have never been measured against verified ground truth spanning this hybrid space, leaving their effectiveness unknown and wild-only evaluations biased. We present MalSkillBench, the first runtime-verified benchmark of malicious agent skills: 3,944 malicious skills labeled along a three-dimensional taxonomy of 108 cells. Of these, 3,214 come from a closed-loop Generate-Verify-Feedback pipeline admitting only samples whose malicious behavior fires inside a Docker sandbox under system-call monitoring and an LLM judge; we add 703 in-the-wild and 4,000 matched benign skills. Our measurements are consistent: code injection reaches 94.5% verification yield but prompt injection only 75.8%, the same fragility that later makes it hard to detect; the wild sample is narrow, dominated by one cryptocurrency-theft campaign (86.6% one behavior, 81% from two accounts) with a small but architecturally new tail attacking the agent control plane; the strongest skill-specific detector reaches 98.4% recall on code injection yet collapses on prompt-injection and agent-control attacks, and wild-only scoring swings the ranking by up to 66 recall points; supply-chain scanners and prompt-injection defenses each see only half of a skill, and no combination recovers the code-instruction relationship. Detecting malicious skills therefore requires reasoning jointly over task intent, code, and instructions. We release the dataset, pipeline, baselines, and results.

cs.CR↗

arXiv2025

Exploring Generative Process Reward Modeling for Semi-Structured Data: A Case Study of Table Question Answering

Process reward models (PRMs) enhance complex reasoning in large language models (LLMs) by evaluating candidate solutions step-by-step and selecting answers based on aggregated step scores. While effective in domains such as mathematics, their applicability to tasks involving semi-structured data, like table question answering (TQA), remains unexplored. TQA poses unique challenges for PRMs, including abundant irrelevant information, loosely connected reasoning steps, and domain-specific reasoning. This work presents the first systematic study of PRMs for TQA. We evaluate state-of-the-art generative PRMs on TQA from both answer and step perspectives. Results show that PRMs that combine textual and code verification can aid solution selection but struggle to generalize to out-of-domain data. Analysis reveals a weak correlation between performance in step-level verification and answer accuracy, possibly stemming from weak step dependencies and loose causal links. Our findings highlight limitations of current PRMs on TQA and offer valuable insights for building more robust, process-aware verifiers.

cs.CL↗

arXiv2025

Near-field radiative heat transfer in the dual nanoscale regime between polaritonic membranes

The enhancement and attenuation of near-field radiative heat transfer between polaritonic SiC, SiN and SiO2 subwavelength membranes is analyzed. Fluctuational electrodynamics simulations combined with a modal analysis show that all membranes support corner and edge modes, which can induce a large 5.1-fold enhancement for SiC and a 2.1-fold attenuation for SiO2 of the heat transfer coefficient with respect to that between infinite surfaces. The enhancement or attenuation is directly related to material losses which reduce the density of available electromagnetic states between the membranes.

cond-mat.mes-hall↗

arXiv2025

Proving Cypher Query Equivalence

Graph database systems store graph data as nodes and relationships, and utilize graph query languages (e.g., Cypher) for efficiently querying graph data. Proving the equivalence of graph queries is an important foundation for optimizing graph query performance, ensuring graph query reliability, etc. Although researchers have proposed many SQL query equivalence provers for relational database systems, these provers cannot be directly applied to prove the equivalence of graph queries. The difficulty lies in the fact that graph query languages (e.g., Cypher) adopt significantly different data models (property graph model vs. relational model) and query patterns (graph pattern matching vs. tabular tuple calculus) from SQL. In this paper, we propose GraphQE, an automated prover to determine whether two Cypher queries are semantically equivalent. We design a U-semiring based Cypher algebraic representation to model the semantics of Cypher queries. Our Cypher algebraic representation is built on the algebraic structure of unbounded semirings, and can sufficiently express nodes and relationships in property graphs and complex Cypher queries. Then, determining the equivalence of two Cypher queries is transformed into determining the equivalence of the corresponding Cypher algebraic representations, which can be verified by SMT solvers. To evaluate the effectiveness of GraphQE, we construct a dataset consisting of 148 pairs of equivalent Cypher queries. Among them, we have successfully proven 138 pairs of equivalent Cypher queries, demonstrating the effectiveness of GraphQE.

cs.DB↗

arXiv2025

Dark-Matter Induced Scalarization of Black Holes in Extended Scalar-Tensor-Gauss-Bonnet Theories

In the extended scalar-tensor-Gauss-Bonnet theory, spontaneous scalarization in the GB$^-$ regime typically occurs only in rotating black holes, while it is absent in spherically symmetric black holes, a phenomenon known as spin-induced scalarization. However, we find that when the spacetime is permeated by perfect fluid dark matter, spontaneous scalarization can also be induced by dark matter in the GB$^-$ regime. Analytical calculations reveal that this scalarization occurs when the dark matter parameter $b/M$ exceeds a critical value $(b/M)_\text{crit}\simeq1.86287$, a threshold determined by the lower boundary of the unstable region for scalar perturbations as the coupling constant approaches negative infinity. Additionally, we verified these findings through numerical analysis of the time evolution of scalar perturbations, identifying the unstable parameter region. The results show that when coupling constant $-\lambda/M^2$ is small, spontaneous scalarization only occurs near the extremal black hole limit. As $-\lambda/M^2$ increases, the scalarization region expands; however, its lower boundary remains above $b/M \simeq 1.86287$, consistent with theoretical predictions.

gr-qc↗

arXiv2025

AI-Driven Automated Tool for Abdominal CT Body Composition Analysis in Gastrointestinal Cancer Management

The incidence of gastrointestinal cancers remains significantly high, particularly in China, emphasizing the importance of accurate prognostic assessments and effective treatment strategies. Research shows a strong correlation between abdominal muscle and fat tissue composition and patient outcomes. However, existing manual methods for analyzing abdominal tissue composition are time-consuming and costly, limiting clinical research scalability. To address these challenges, we developed an AI-driven tool for automated analysis of abdominal CT scans to effectively identify and segment muscle, subcutaneous fat, and visceral fat. Our tool integrates a multi-view localization model and a high-precision 2D nnUNet-based segmentation model, demonstrating a localization accuracy of 90% and a Dice Score Coefficient of 0.967 for segmentation. Furthermore, it features an interactive interface that allows clinicians to refine the segmentation results, ensuring high-quality outcomes effectively. Our tool offers a standardized method for effectively extracting critical abdominal tissues, potentially enhancing the management and treatment for gastrointestinal cancers. The code is available at https://github.com/NanXinyu/AI-Tool4Abdominal-Seg.git}{https://github.com/NanXinyu/AI-Tool4Abdominal-Seg.git.

eess.IV↗

arXiv2025

Adaptive Few-shot Prompting for Machine Translation with Pre-trained Language Models

Recently, Large language models (LLMs) with in-context learning have demonstrated remarkable potential in handling neural machine translation. However, existing evidence shows that LLMs are prompt-sensitive and it is sub-optimal to apply the fixed prompt to any input for downstream machine translation tasks. To address this issue, we propose an adaptive few-shot prompting (AFSP) framework to automatically select suitable translation demonstrations for various source input sentences to further elicit the translation capability of an LLM for better machine translation. First, we build a translation demonstration retrieval module based on LLM's embedding to retrieve top-k semantic-similar translation demonstrations from aligned parallel translation corpus. Rather than using other embedding models for semantic demonstration retrieval, we build a hybrid demonstration retrieval module based on the embedding layer of the deployed LLM to build better input representation for retrieving more semantic-related translation demonstrations. Then, to ensure better semantic consistency between source inputs and target outputs, we force the deployed LLM itself to generate multiple output candidates in the target language with the help of translation demonstrations and rerank these candidates. Besides, to better evaluate the effectiveness of our AFSP framework on the latest language and extend the research boundary of neural machine translation, we construct a high-quality diplomatic Chinese-English parallel dataset that consists of 5,528 parallel Chinese-English sentences. Finally, extensive experiments on the proposed diplomatic Chinese-English parallel dataset and the United Nations Parallel Corpus (Chinese-English part) show the effectiveness and superiority of our proposed AFSP.

cs.CL↗

arXiv2024

Phase-selective tripartite entanglement and asymmetric Einstein-Podolsky-Rosen steering in squeezed optomechanics

The generation and manipulation of multipartite entanglement and EPR steering in macroscopic systems not only play a fundamental role in exploring the nature of quantum mechanics, but are also at the core of current developments of various nascent quantum technologies. Here we report a theoretical method using squeezing-phase-controlled quantum noise flows to selectively generate and manipulate quantum entanglement and asymmetric EPR steering in a nonlinear $\chi^{(2)}$ whispering-gallery-mode (WGM) optomechanical resonator. We show that by pumping the $\chi^{(2)}$ nonlinear medium with two-photon optical fields and broadband squeezed lights, a pair of counterpropagating squeezed optical modes could be introduced to the WGM resonator, each coupled with an independent squeezed vacuum reservoir. This configuration could enable squeezing-phase-controlled light-reservoir interaction for each squeezed optical mode, providing a flexible tool for tailoring asymmetric optical noise flows in the counterpropagating modes. Based on this unique feature, it is found that with the injection of asymmetric noise flows, the generation of various types of bipartite and tripartite entanglement become phase-dependent and thus they can be produced in an asymmetric way. More excitingly, it is also found that by further properly adjusting the squeezing parameters, the overall asymmetry of EPR steering can also be stepwise driven from no-way regime, one-way regime to two-way regime. These findings, holding promise for preparing rich types of entangled quantum resources with asymmetric features, may have potential applications in the area of secure quantum information processing such as quantum secure direct communication and one-way quantum computing.

quant-ph↗

arXiv2024

PAM: A Propagation-Based Model for Segmenting Any 3D Objects across Multi-Modal Medical Images

Volumetric segmentation is important in medical imaging, but current methods face challenges like requiring lots of manual annotations and being tailored to specific tasks, which limits their versatility. General segmentation models used for natural images don't perform well with the unique features of medical images. There's a strong need for an adaptable approach that can effectively handle different 3D medical structures and imaging modalities. In this study, we present PAM (Propagating Anything Model), a segmentation approach that uses a 2D prompt, like a bounding box or sketch, to create a complete 3D segmentation of medical image volumes. PAM works by modeling relationships between slices, maintaining information flow across the 3D structure. It combines a CNN-based UNet for processing within slices and a Transformer-based attention module for propagating information between slices, leading to better generalizability across various imaging modalities. PAM significantly outperformed existing models like MedSAM and SegVol, with an average improvement of over 18.1% in dice similarity coefficient (DSC) across 44 medical datasets and various object types. It also showed stable performance despite prompt deviations and different propagation setups, and faster inference speeds compared to other models. PAM's one-view prompt design made it more efficient, reducing interaction time by about 63.6% compared to two-view prompts. Thanks to its focus on structural relationships, PAM handled unseen and complex objects well, showing a unique ability to generalize to new situations. PAM represents an advancement in medical image segmentation, effectively reducing the need for extensive manual work and specialized training. Its adaptability makes it a promising tool for more automated and reliable analysis in clinical settings.

cs.CV↗

arXiv2024

Quantum Cross Nonlinearity for Photon-Number-Resolving Nondestructive Detection

We present an unconventional mechanism for quantum nonlinearity in a system comprising of a V-type quantum emitter (QE) and two Fabry-Perot cavities. The two transitions of the V-type QE are effectively coupled with two independent cavity modes. The system exhibits a strong quantum nonlinear control in the transmission even at the single-photon level, which we refer to as quantum cross nonlinearity. The underlying physics can be understood as quantum competition between the two transitions of the QE sharing a common ground state. By leveraging this quantum cross nonlinearity, we further show photon-number-resolving quantum nondestructive detection. Owing to the widespread nature of this V-type configuration, our approach can be readily extended to diverse cavity quantum electrodynamic systems beyond the realm of optics, encompassing, e.g., microwave photons and acoustic wave phonons. This versatility may facilitate numerous unique applications for quantum information processing.

quant-ph↗

arXiv2023

Large enhancement of near-field radiative heat transfer in the dual nanoscale regime enabled by electromagnetic corner and edge modes

It is well established that near-field radiative heat transfer (NFRHT) can exceed Planck's blackbody limit1 by orders of magnitude owing to the tunneling of evanescent electromagnetic frustrated and surface modes2-4, as has been demonstrated experimentally for NFRHT between two large parallel surfaces5-7 and between two subwavelength membranes8,9. However, while nanostructures can also sustain a much richer variety of localized electromagnetic modes at their corners and edges,10,11 the contributions of such additional modes to further enhancing NFRHT remain unexplored. Here, for the first time, we demonstrate both theoretically and experimentally a new physical mechanism of NFRHT mediated by these corner and edge modes, and show it can dominate the NFRHT in the "dual nanoscale regime" in which both the thickness of the emitter and receiver, and their gap spacing, are much smaller than the thermal photon wavelengths. For two coplanar 20 nm thick SiC membranes separated by a 100 nm vacuum gap, the NFRHT coefficient at room temperature is both predicted and measured to be 830 W/m2K, which is 5.5 times larger than that for two infinite SiC surfaces separated by the same gap, and 1400 times larger than the corresponding blackbody limit accounting for the geometric view factor between the emitter and receiver. This enhancement is dominated by the electromagnetic corner and edge modes which account for 81% of the NFRHT between these SiC membranes. These findings are important for future NFRHT applications in thermal management and energy conversion.

physics.optics↗

arXiv2023

Progressive Dual Priori Network for Generalized Breast Tumor Segmentation

To promote the generalization ability of breast tumor segmentation models, as well as to improve the segmentation performance for breast tumors with smaller size, low-contrast and irregular shape, we propose a progressive dual priori network (PDPNet) to segment breast tumors from dynamic enhanced magnetic resonance images (DCE-MRI) acquired at different centers. The PDPNet first cropped tumor regions with a coarse-segmentation based localization module, then the breast tumor mask was progressively refined by using the weak semantic priori and cross-scale correlation prior knowledge. To validate the effectiveness of PDPNet, we compared it with several state-of-the-art methods on multi-center datasets. The results showed that, comparing against the suboptimal method, the DSC and HD95 of PDPNet were improved at least by 5.13% and 7.58% respectively on multi-center test sets. In addition, through ablations, we demonstrated that the proposed localization module can decrease the influence of normal tissues and therefore improve the generalization ability of the model. The weak semantic priors allow focusing on tumor regions to avoid missing small tumors and low-contrast tumors. The cross-scale correlation priors are beneficial for promoting the shape-aware ability for irregular tumors. Thus integrating them in a unified framework improved the multi-center breast tumor segmentation performance. The source code and open data can be accessed at https://github.com/wangli100209/PDPNet.

eess.IV↗

arXiv2023

Effects of thermal annealing on thermal conductivity of LPCVD silicon carbide thin films

The thermal conductivity (k) of polycrystalline silicon carbide thin films is relevant for thermal management in emerging silicon carbide applications like MEMS and optoelectronic devices. In such films k can be substantially reduced by microstructure features including grain boundaries, thin film surfaces, and porosity, while these microstructural effects can also be manipulated through thermal annealing. Here, we investigate these effects by using microfabricated suspended devices to measure the thermal conductivities of nine LPCVD silicon carbide films of varying thickness (from 120 - 300 nm) and annealing conditions (as-grown and annealed at 950 degrees Celsius and 1100 degrees Celsius for 2 hours, and in one case 17 hours). Fourier-transform infrared spectroscopy (FTIR) and X-ray diffraction (XRD) spectra and density measurements are also used to characterize the effects of the annealing on the microstructure of selected samples. Compared to as-deposited films, annealing at 1100 degrees Celsius typically increases the estimated grain size from 5.5 nm to 6.6 nm while decreasing the porosity from around 6.5% to practically fully dense. This corresponds to a 34% increase in the measured thin film thermal conductivity near room temperature, from 5.8 W/m-K to 7.8 W/m-K. These thermal conductivity measurements show good agreement of better than 3% with fits using a simple theoretical model based on kinetic theory combined with a Maxwell-Garnett porosity correction. Grain boundary scattering plays the dominant role in reducing the thermal conductivity of these films compared to bulk single-crystal values, while both grain size increase and porosity decrease play important roles in the partial k recovery of the films upon annealing. This work demonstrates the effects of modifying the microstructure and thus the thermal conductivity of silicon carbide thin films by thermal annealing.

cond-mat.mtrl-sci↗

arXiv2023

propnet: Propagating 2D Annotation to 3D Segmentation for Gastric Tumors on CT Scans

**Background:** Accurate 3D CT scan segmentation of gastric tumors is pivotal for diagnosis and treatment. The challenges lie in the irregular shapes, blurred boundaries of tumors, and the inefficiency of existing methods. **Purpose:** We conducted a study to introduce a model, utilizing human-guided knowledge and unique modules, to address the challenges of 3D tumor segmentation. **Methods:** We developed the PropNet framework, propagating radiologists' knowledge from 2D annotations to the entire 3D space. This model consists of a proposing stage for coarse segmentation and a refining stage for improved segmentation, using two-way branches for enhanced performance and an up-down strategy for efficiency. **Results:** With 98 patient scans for training and 30 for validation, our method achieves a significant agreement with manual annotation (Dice of 0.803) and improves efficiency. The performance is comparable in different scenarios and with various radiologists' annotations (Dice between 0.785 and 0.803). Moreover, the model shows improved prognostic prediction performance (C-index of 0.620 vs. 0.576) on an independent validation set of 42 patients with advanced gastric cancer. **Conclusions:** Our model generates accurate tumor segmentation efficiently and stably, improving prognostic performance and reducing high-throughput image reading workload. This model can accelerate the quantitative analysis of gastric tumors and enhance downstream task performance.

eess.IV↗