arXiv subjects

Wenlong Yu

Publications and source records attributed to Wenlong Yu.

At least 19 recordsLinked to original sources

arXiv2026

Histopathological Spectrum-Guided Prostate Stratification via Segmentation-Assisted Diagnostic Transformer

Prostate cancer diagnosis with multiparametric MRI (mpMRI) is commonly based on PI-RADS assessment or binary classification, which suffer from subjectivity and fail to capture clinically relevant pathological heterogeneity. To address this limitation, we construct a Prostate Cancer Histopathology Spectrum Dataset (PCa-HSD) and formulate a clinically meaningful four-class classification task, addressing the underrepresentation of benign lesions that are easily confounded with prostate cancer in existing datasets. We propose Language-guided Segmentation-assisted Diagnostic Transformer model (LSDT), which leverages zero-shot segmentation to provide anatomical priors and performs effective multi-modal slice fusion for classification. Our proposed method consistently improves accuracy across backbones, achieving the best average accuracy of 0.633 and JointRecall of 0.768 in five-fold cross-validation on a cohort of 344 patients. These results demonstrate that integrating pathology supervision and anatomical priors significantly enhances fine-grained prostate MRI classification and provides a more clinically relevant paradigm for risk stratification. Code will be made publicly available in a future revision.

cs.CV↗

arXiv2026

Teaching the Way, Not the Answer: Privileged Tutoring Distillation for Multimodal Policy Optimization

Recent post-training methods, particularly Reinforcement Learning with Verifiable Rewards (RLVR), have significantly enhanced the reasoning ability of Large Vision-Language Models (LVLMs). However, the sparse nature of verifiable rewards provides little token-level supervision for failed rollouts, often leading to inefficient exploration in complex multimodal reasoning tasks. Although policy distillation can offer dense guidance, external teacher based methods introduce substantial computational overhead, while answer conditioned tuning methods may expose answer-level information and induce shortcut-like generation behavior. To address these limitations, we propose PTD-PO, a Privileged Tutoring Distillation Policy Optimization framework for RLVR that provides dense guidance without exposing the answer to the student policy. Specifically, PTD-PO constructs structured privileged hints from spatial attention guidance and intermediate textual reasoning steps, and uses them through in-context learning to produce step-wise token-distribution supervision. The student is still optimized under the original answer-free context, and its failed rollouts are aligned with the hint-augmented reference model at the token-distribution level. To further stabilize distillation under the distribution shift between guided and unguided contexts, we introduce a Top-K Jensen-Shannon divergence objective that focuses alignment on informative token probabilities while reducing memory overhead. Experiments on LVLMs ranging from 2B to 8B parameters show that PTD-PO consistently outperforms RLVR and distillation baselines, mitigates entropy collapse, and improves complex multimodal reasoning performance.

cs.AI↗

arXiv2026

SFCoT: Safer Chain-of-Thought via Active Safety Evaluation and Calibration

Large language models (LLMs) have demonstrated remarkable capabilities in complex reasoning tasks. However, they remain highly susceptible to jailbreak attacks that undermine their safety alignment. Existing defense mechanisms typically rely on post hoc filtering applied only to the final output, leaving intermediate reasoning steps unmonitored and vulnerable to adversarial manipulation. To address this gap, this paper proposes a SaFer Chain-of-Thought (SFCoT) framework, which proactively evaluates and calibrates potentially unsafe reasoning steps in real time. SFCoT incorporates a three-tier safety scoring system alongside a multi-perspective consistency verification mechanism, designed to detect potential risks throughout the reasoning process. A dynamic intervention module subsequently performs targeted calibration to redirect reasoning trajectories toward safe outcomes. Experimental results demonstrate that SFCoT reduces the attack success rate from $58.97\%$ to $12.31\%$, demonstrating it as an effective and efficient LLM safety enhancement method without a significant decline in general performance.

cs.CR↗

arXiv2026

Fine-Grained Generalization via Structuralizing Concept and Feature Space into Commonality, Specificity and Confounding

Fine-Grained Domain Generalization (FGDG) presents greater challenges than conventional domain generalization due to the subtle inter-class differences and relatively pronounced intra-class variations inherent in fine-grained recognition tasks. Under domain shifts, the model becomes overly sensitive to fine-grained cues, leading to the suppression of critical features and a significant drop in performance. Cognitive studies suggest that humans classify objects by leveraging both common and specific attributes, enabling accurate differentiation between fine-grained categories. However, current deep learning models have yet to incorporate this mechanism effectively. Inspired by this mechanism, we propose Concept-Feature Structuralized Generalization (CFSG). This model explicitly disentangles both the concept and feature spaces into three structured components: common, specific, and confounding segments. To mitigate the adverse effects of varying degrees of distribution shift, we introduce an adaptive mechanism that dynamically adjusts the proportions of common, specific, and confounding components. In the final prediction, explicit weights are assigned to each pair of components. Extensive experiments on three single-source benchmark datasets demonstrate that CFSG achieves an average performance improvement of 9.87% over baseline models and outperforms existing state-of-the-art methods by an average of 3.08%. Additionally, explainability analysis validates that CFSG effectively integrates multi-granularity structured knowledge and confirms that feature structuralization facilitates the emergence of concept structuralization.

cs.CV↗

arXiv2025

Scalable architecture for dark photon searches: Superconducting-qubit proof of principle

The dark photon is a well-motivated candidate of dark matter due to its potential to open the window of new physics beyond the Standard Model. A fundamental mass-range-sensitivity dilemma is always haunting the dark photon searching experiments: The resonant haloscopes have excellent sensitivity but are narrowband, and vice versa for the non-resonant ones. A scalable architecture integrating numerous resonant haloscopes will be a desirable solution to this dilemma. However, even the concept of scalable searching remains rarely explored, due to the size limitation of conventional haloscopes imposed by the dark photon wavelength. Here we propose and demonstrate a novel architecture using superconducting qubits as sub-wavelength haloscope units. By virtue of the scalability of superconducting qubits, it is possible to integrate multiple qubits with different frequencies on a chip-scale device. Furthermore, the frequencies of the qubits can be tuned to extend the searching mass range. Thus, our architectures allow for searching for dark photons in a broad mass range with high sensitivity. As a proof-of-principle experiment, we designed and fabricated a three-qubit chip and successfully demonstrated a scalable dark-photon searching. Our work established constraints on dark photons in the mass range of 15.632 $\mu$eV$\sim$15.638 $\mu$eV, 15.838 $\mu$eV$\sim$15.845 $\mu$eV, and 16.463 $\mu$eV$\sim$16.468 $\mu$eV, simultaneously, and the constraints are much more stringent than the cosmology constraints. Our work can be scaled up in the future to boost the scrutiny of new physics and extended to search for more dark matter candidates, including dark photons, axions and axion-like particles.

hep-ph↗

arXiv2025

CoE: Chain-of-Explanation via Automatic Visual Concept Circuit Description and Polysemanticity Quantification

Explainability is a critical factor influencing the wide deployment of deep vision models (DVMs). Concept-based post-hoc explanation methods can provide both global and local insights into model decisions. However, current methods in this field face challenges in that they are inflexible to automatically construct accurate and sufficient linguistic explanations for global concepts and local circuits. Particularly, the intrinsic polysemanticity in semantic Visual Concepts (VCs) impedes the interpretability of concepts and DVMs, which is underestimated severely. In this paper, we propose a Chain-of-Explanation (CoE) approach to address these issues. Specifically, CoE automates the decoding and description of VCs to construct global concept explanation datasets. Further, to alleviate the effect of polysemanticity on model explainability, we design a concept polysemanticity disentanglement and filtering mechanism to distinguish the most contextually relevant concept atoms. Besides, a Concept Polysemanticity Entropy (CPE), as a measure of model interpretability, is formulated to quantify the degree of concept uncertainty. The modeling of deterministic concepts is upgraded to uncertain concept atom distributions. Finally, CoE automatically enables linguistic local explanations of the decision-making process of DVMs by tracing the concept circuit. GPT-4o and human-based experiments demonstrate the effectiveness of CPE and the superiority of CoE, achieving an average absolute improvement of 36% in terms of explainability scores.

cs.CV↗

arXiv2024

Buried Dirac points in quantum spin Hall insulators: Implications for Majorana Kramers pair-based quantum computing

For heterostructures formed by a quantum spin Hall insulator (QSHI) placed in proximity to a superconductor (SC), no external magnetic field is necessary to drive the system into a phase supporting topological superconductivity with Majorana zero energy states, making them very attractive for the realization of non-Abelian states and fault-tolerant qubits. Despite considerable work investigating QSHI edge states, there is still an open question about their resilience to large magnetic fields and the implication of such resilience for the formation of a quasi-1D topological superconducting state. In this work, we investigate the transport properties of helical edge states in a QSHI-SC junction formed by a InAs/GaSb (15nm/5nm) double quantum well and a superconducting tantalum (Ta) constriction. We observe a robust conductance plateau up to 2 T, signaling resilient edge state transport. Such resilience is consistent with the Dirac point for the edge states being buried in the bulk valence band. Using a modified Landauer-Buttiker analysis, we find that the conductance is consistent with 98% Andreev reflection probability owing to the high transparency of the InAs/GaSb-Ta interface. We further theoretically show that a buried Dirac point does not affect the robustness of the quasi-1D topological superconducting phase, and favors the hybridization of Majorana Kramer pairs and fermionic modes in the QSHI resulting in extended MKP states, highlighting the subtle role of buried Dirac points in probing MKPs.

cond-mat.supr-con↗

arXiv2024

Fine-Grained Domain Generalization with Feature Structuralization

Fine-grained domain generalization (FGDG) is a more challenging task than traditional DG tasks due to its small inter-class variations and relatively large intra-class disparities. When domain distribution changes, the vulnerability of subtle features leads to a severe deterioration in model performance. Nevertheless, humans inherently demonstrate the capacity for generalizing to out-of-distribution data, leveraging structured multi-granularity knowledge that emerges from discerning the commonality and specificity within categories. Likewise, we propose a Feature Structuralized Domain Generalization (FSDG) model, wherein features experience structuralization into common, specific, and confounding segments, harmoniously aligned with their relevant semantic concepts, to elevate performance in FGDG. Specifically, feature structuralization (FS) is accomplished through joint optimization of five constraints: a decorrelation function applied to disentangled segments, three constraints ensuring common feature consistency and specific feature distinctiveness, and a prediction calibration term. By imposing these stipulations, FSDG is prompted to disentangle and align features based on multi-granularity knowledge, facilitating robust subtle distinctions among categories. Extensive experimentation on three benchmarks consistently validates the superiority of FSDG over state-of-the-art counterparts, with an average improvement of 6.2% in FGDG performance. Beyond that, the explainability analysis on explicit concept matching intensity between the shared concepts among categories and the model channels, along with experiments on various mainstream model architectures, substantiates the validity of FS.

cs.CV↗

arXiv2022

Leggett Modes in Dirac Semimetals

In recent years experimentalists have been able to clearly show that several materials, such as MgB2, iron-based superconductors3, monolayer NbSe2, are multiband superconductors. Superconducting pairing in multiple bands can give rise to novel and very interesting phenomena. Leggett modes are exemplary of the unusual effects that can be present in multiband superconductors. A Leggett mode describes the collective periodic oscillation of the relative phase between the phases of the superconducting condensates formed by electrons in different bands. It can be thought of as the mode arising from an inter-band Josephson effect. The experimental observation of Leggett modes is challenging for several reasons: (i) Multiband superconductors are rare; (ii) they describe charge fluctuations between bands and therefore are hard to probe directly; (iii) their mass gap is often larger than the superconducting gaps and therefore are strongly overdamped via relaxation processes into the quasiparticle continuum. In this work we show that Leggett modes, and their frequency, can be detected unambigously in a.c. driven superconducting quantum interference devices (SQUIDs). We then use the results to analyze the measurements of a SQUID based on Cd3As2, an exemplar Dirac semimetal, in which superconductivity is induced by proximity to superconducting Al. The experimental results show the theoretically predicted unique signatures of Leggett modes and therefore allow us to conclude that a Leggett mode is present in the two-band superconducting state of Dirac semimetal (DSM) Cd3As2.

cond-mat.supr-con↗

arXiv2022

Titanium Nitride Film on Sapphire Substrate with Low Dielectric Loss for Superconducting Qubits

Dielectric loss is one of the major decoherence sources of superconducting qubits. Contemporary high-coherence superconducting qubits are formed by material systems mostly consisting of superconducting films on substrate with low dielectric loss, where the loss mainly originates from the surfaces and interfaces. Among the multiple candidates for material systems, a combination of titanium nitride (TiN) film and sapphire substrate has good potential because of its chemical stability against oxidization, and high quality at interfaces. In this work, we report a TiN film deposited onto sapphire substrate achieving low dielectric loss at the material interface. Through the systematic characterizations of a series of transmon qubits fabricated with identical batches of TiN base layers, but different geometries of qubit shunting capacitors with various participation ratios of the material interface, we quantitatively extract the loss tangent value at the substrate-metal interface smaller than $8.9 \times 10^{-4}$ in 1-nm disordered layer. By optimizing the interface participation ratio of the transmon qubit, we reproducibly achieve qubit lifetimes of up to 300 $\mu$s and quality factors approaching 8 million. We demonstrate that TiN film on sapphire substrate is an ideal material system for high-coherence superconducting qubits. Our analyses further suggest that the interface dielectric loss around the Josephson junction part of the circuit could be the dominant limitation of lifetimes for state-of-the-art transmon qubits.

quant-ph↗

arXiv2021

Fluxonium: an alternative qubit platform for high-fidelity operations

Superconducting qubits provide a promising path toward building large-scale quantum computers. The simple and robust transmon qubit has been the leading platform, achieving multiple milestones. However, fault-tolerant quantum computing calls for qubit operations at error rates significantly lower than those exhibited in the state of the art. Consequently, alternative superconducting qubits with better error protection have attracted increasing interest. Among them, fluxonium is a particularly promising candidate, featuring large anharmonicity and long coherence times. Here, we engineer a fluxonium-based quantum processor that integrates high qubit-coherence, fast frequency-tunability, and individual-qubit addressability for reset, readout, and gates. With simple and fast gate schemes, we achieve an average single-qubit gate fidelity of 99.97% and a two-qubit gate fidelity of up to 99.72%. This performance is comparable to the highest values reported in the literature of superconducting circuits. Thus our work, for the first time within the realm of superconducting qubits, reveals an approach toward fault-tolerant quantum computing that is alternative and competitive to the transmon system.

quant-ph↗

arXiv2021

Ultrahigh-inductance materials from spinodal decomposition

Disordered superconducting nitrides with kinetic inductance have long been considered a leading material candidate for high-inductance quantum-circuit applications. Despite continuing efforts in reducing material dimensions to increase the kinetic inductance and the corresponding circuit impedance, it becomes a fundamental challenge to improve further without compromising material qualities. To this end, we propose a method to drastically increase the kinetic inductance of superconducting materials via spinodal decomposition while keeping a low microwave loss. We use epitaxial Ti\textsubscript{0.48}Al\textsubscript{0.52}N as a model system, and for the first time demonstrate the utilization of spinodal decomposition to trigger the insulator-to-superconductor transition with a drastically enhanced material disorder. The measured kinetic inductance has increased by 2-3 orders of magnitude compared with all the best reported disordered superconducting nitrides. Our work paves the way for substantially enhancing and deterministically controlling the inductance for advanced superconducting quantum circuits.

quant-ph↗

arXiv2021

Epitaxial titanium nitride microwave resonators: Structural, chemical, electrical, and microwave properties

Titanium nitride is an attractive material for a range of superconducting quantum-circuit applications owing to its low microwave losses, high surface inductance, and chemical stability. The physical properties and device performance, nevertheless, depend strongly on the quality of the materials. Here we focus on the highly crystalline and epitaxial titanium nitride thin films deposited on sapphire substrates using magnetron sputtering at an intermediate temperature (300$^{\circ}$C). We perform a set of systematic and comprehensive material characterization to thoroughly understand the structural, chemical, and transport properties. Microwave losses at low temperatures are studied using patterned microwave resonators, where the best internal quality factor in the single-photon regime is measured to be $3.3\times 10^6$, and $> 1.0\times 10^7$ in the high-power regime. Adjusted with the material filling factor of the resonators, the microwave loss-tangent here compares well with the previously reported best values for superconducting resonators. This work lays the foundation of using epitaxial titanium nitride for low-loss superconducting quantum circuits.

physics.app-ph↗

arXiv2021

Microwave response in a topological superconducting quantum interference device

Photon detection at microwave frequency is of great interest due to its application in quantum computation information science and technology. Herein are results from studying microwave response in a topological superconducting quantum interference device (SQUID) realized in Dirac semimetal Cd3As2. The temperature dependence and microwave power dependence of the SQUID junction resistance are studied, from which we obtain an effective temperature at each microwave power level. It is observed the effective temperature increases with the microwave power. This observation of microwave response may pave the way for single photon detection at the microwave frequency in topological quantum materials.

cond-mat.mtrl-sci↗

arXiv2018

{\pi} and 4{\pi} Josephson Effects Mediated by a Dirac Semimetal

Cd3As2 is a three-dimensional topological Dirac semimetal with connected Fermi-arc surface states. It has been suggested that topological superconductivity can be achieved in the nontrivial surface states of topological materials by utilizing the superconductor proximity effect. Here we report observations of both {\pi} and 4{\pi} periodic supercurrents in aluminum-Cd3As2-aluminum Josephson junctions. The {\pi} period is manifested by both the magnetic-field dependence of the critical supercurrent and the appearance of half-integer Shapiro steps in the a.c. Josephson effect. Our macroscopic theory suggests that the {\pi} period arises from interference between the induced bulk superconductivity and the induced Fermi arc surface superconductivity. The 4{\pi} period is manifested by the missing first Shapiro steps and is expected for topological superconductivity.

cond-mat.supr-con↗

arXiv2017

Strong Photothermoelectric Response and Contact Reactivity of the Dirac Semimetal ZrTe5

The family of three-dimensional topological insulators opens new avenues to discover novel photophysics and to develop novel types of photodetectors. ZrTe5 has been shown to be a Dirac semimetal possessing unique topological electronic and optical properties. Here we present spatially-resolved photocurrent measurements on devices made of nanoplatelets of ZrTe5, demonstrating the photothermoelectric origin of the photoresponse. Due to the high electrical conductivity and good Seebeck coefficient, we obtain noise-equivalent powers as low as 42 pW/Hz1/2 at room temperature for visible light illumination at zero bias. We also show that these devices suffer from significant ambient reactivity such as the formation of a Te-rich surface region driven by Zr oxidation, as well as severe reactions with the metal contacts. This reactivity results in significant stresses in the devices, leading to unusual geometries that are useful for gaining insight into the photocurrent mechanisms. Our results indicate that both the large photothermoelectric response and reactivity must be considered when designing or interpreting photocurrent measurements in these systems.

cond-mat.mes-hall↗

arXiv2014

Giant supercurrent states in a superconductor-InAs/GaSb-superconductor junction

Superconductivity in topological materials has attracted a great deal of interest in both electron physics and material sciences since the theoretical predictions that Majorana fermions can be realized in topological superconductors [1-4]. Topological superconductivity could be realized in a type II, band-inverted, InAs/GaSb quantum well if it is in proximity to a conventional superconductor. Here we report observations of the proximity effect induced giant supercurrent states in an InAs/GaSb bilayer system that is sandwiched between two superconducting tantalum electrodes to form a superconductor-InAs/GaSb-superconductor junction. Electron transport results show that the supercurrent states can be preserved in a surprisingly large temperature-magnetic field (T-H) parameter space. In addition, the evolution of differential resistance in T and H reveals an interesting superconducting gap structure.

cond-mat.supr-con↗

arXiv2014

Superconducting proximity effect in inverted InAs/GaSb quantum well structures with Ta electrodes

We present our recent electronic transport results in top-gated InAs/GaSb quantum well hybrid structures with superconducting Ta electrodes. We show that the transport across the InAs-Ta junction depends largely on the interfacial transparency, exhibiting distinct zero-bias behavior. For a relatively resistive interface a broad conductance peak is observed at zero bias. When a transparent InAs-Ta interface is achieved, a zero-bias conductance dip appears with two coherent-peak-like features forming at bias voltages corresponding to the superconducting gap of Ta. The conductance spectra of the transparent InAs-Ta junction at different gate voltages can be fit well using the standard Blonder-Tinkham-Klapwijk theory.

cond-mat.mes-hall↗