arXiv subjects

Lei Su

Publications and source records attributed to Lei Su.

At least 19 recordsLinked to original sources

arXiv2026

Exact fermionic dual of the Bose-Hubbard model

Recent developments have established exact bosonization and fermionization with a $\mathbb{Z}_2$ symmetry as dualities through gauging. In this work, we apply fermionic gauging, which realizes generalized Jordan-Wigner transformations, to the Bose-Hubbard (BH) model with a global $U(1)$ symmetry and derive an exact dual description in terms of fermionic composites, built from bosons and fermions. In 1D, this duality generalizes the exact mapping between the extended hard-core BH model and the spinless Fermi-Hubbard model to now cover soft-core interactions. At low energies, the mapping reduces to the well-known equivalence between the sine-Gordon model and the Thirring model. The oscillation wave vector of the fermionic composite correlation function in the gapless phase is fixed by their density, providing a novel manifestation of Luttinger's theorem. We verify the exact duality using density matrix renormalization group (DMRG) calculations and demonstrate that the gapless phase and the phase transition are governed by the compact boson conformal field theory. Our construction naturally extends to generic bosonic systems and higher dimensions, opening new avenues for studying Bose-Fermi mixtures in optical lattices and other strongly correlated quantum systems.

cond-mat.str-el↗

arXiv2026

FTU-Seek: Foundation Model-Guided Hard-Negative Learning for Sparse Functional Tissue Unit Segmentation

Functional tissue units (FTUs), including tertiary lymphoid structures (TLSs), blood vessels, and glands, encode localized immune, vascular, and epithelial organization in histopathology. Accurate quantification of these structures is important for studying tissue architecture and disease-associated tissue organization. However, FTUs are frequently sparse, heterogeneous, and surrounded by large amounts of morphologically similar background tissue, making automated segmentation in whole-slide images (WSIs) challenging. We therefore developed FTU-Seek, a pathology foundation model-guided framework that treats morphology-aware negative-patch selection as a key component of sparse FTU segmentation. FTU-Seek uses frozen multi-depth features from the UNI pathology foundation model to train a patch-level classifier that distinguishes FTU-containing from FTU-absent tissue. Target-absent patches are subsequently ranked according to their predicted target-containing probabilities, and the highest-scoring hard negatives are selected through a static Top$K$ strategy to construct compact segmentation training sets. The framework was evaluated using five-fold cross-validation and internal test cohorts across TLS, blood-vessel, and gland segmentation tasks, with an additional independent 30-WSI held-out cohort for TLS. Positive-only, all-tissue, random-negative, and matched random Top$K$ sampling strategies served as comparators. Segmentation-derived phenotypes were further explored in external TCGA cohorts.

cs.CV↗

arXiv2026

Global regularity for 2D gravity water waves with two retreating point vortices

We prove global well-posedness for the two-dimensional infinite-depth gravity water wave system coupled to a pair of point vortices of opposite strengths. The vortex dipole is initially placed deep below the free surface and oriented so that its leading motion is away from the interface. We show that the vortices retreat almost linearly, and that the velocity induced on the free boundary is therefore integrable in time. Combining this decay with Wu's cubic formulation, together with a transition-of-derivatives and localization argument, we obtain global bounds for small localized perturbations of the flat surface.

math.AP↗

arXiv2026

Computational Oncology of Chemotaxis-Driven Tumour--Immune Spatial Patterning and Stability

We develop a reaction--diffusion--chemotaxis model for spatial tumour--immune--chemokine dynamics that couples logistic tumour growth, immune-mediated killing, chemokine-dependent immune recruitment, chemotactic migration, and signal production. For the nondimensional system, we establish local classical solvability, nonnegativity, a uniform tumour-density bound, and global mass estimates for the immune and chemokine components. The tumour-free equilibrium is stable precisely when the baseline immune-control index satisfies $\sigma_0/\delta>1$, whereas positive homogeneous coexistence is characterized by a scalar nonlinear equation. Linearization in the Neumann Laplacian eigenbasis yields a mode-dependent cubic dispersion relation, showing that chemotaxis does not alter the tumour-invasion threshold but can destabilize homogeneous coexistence through a finite-wavelength oscillatory instability above a critical sensitivity $\xi_c$. A conservative finite-volume discretization with upwind chemotactic fluxes and implicit backward differentiation formula time integration is used to test these predictions. Numerical experiments recover the analytical equilibria and growth rates, identify the dominant unstable mode, reproduce the transition to spatial heterogeneity, and quantify the effects of immune recruitment, decay, and diffusion on the stability boundary. Grid-refinement, mass-balance, residual, and nonnegativity diagnostics support the computational reliability of the results.

math.AP↗

arXiv2026

Tacmap: Bridging the Tactile Sim-to-Real Gap via Geometry-Consistent Penetration Depth Map

Vision-Based Tactile Sensors (VBTS) are essential for achieving dexterous robotic manipulation, yet the tactile sim-to-real gap remains a fundamental bottleneck. Current tactile simulations suffer from a persistent dilemma: simplified geometric projections lack physical authenticity, while high-fidelity Finite Element Methods (FEM) are too computationally prohibitive for large-scale reinforcement learning. In this work, we present Tacmap, a high-fidelity, computationally efficient tactile simulation framework anchored in volumetric penetration depth. Our key insight is to bridge the tactile sim-to-real gap by unifying both domains through a shared deform map representation. Specifically, we compute 3D intersection volumes as depth maps in simulation, while in the real world, we employ an automated data-collection rig to learn a robust mapping from raw tactile images to ground-truth depth maps. By aligning simulation and real-world in this unified geometric space, Tacmap minimizes domain shift while maintaining physical consistency. Quantitative evaluations across diverse contact scenarios demonstrate that Tacmap's deform maps closely mirror real-world measurements. Moreover, we validate the utility of Tacmap through an in-hand rotation task, where a policy trained exclusively in simulation achieves zero-shot transfer to a physical robot.

cs.RO↗

arXiv2025

Position-Prior-Guided Network for System Matrix Super-Resolution in Magnetic Particle Imaging

Magnetic Particle Imaging (MPI) is a novel medical imaging modality. One of the established methods for MPI reconstruction is based on the System Matrix (SM). However, the calibration of the SM is often time-consuming and requires repeated measurements whenever the system parameters change. Current methodologies utilize deep learning-based super-resolution (SR) techniques to expedite SM calibration; nevertheless, these strategies do not fully exploit physical prior knowledge associated with the SM, such as symmetric positional priors. Consequently, we integrated positional priors into existing frameworks for SM calibration. Underpinned by theoretical justification, we empirically validated the efficacy of incorporating positional priors through experiments involving both 2D and 3D SM SR methods.

cs.CV↗

arXiv2025

$\mathbb{Z}_2$ lattice gauge theories: fermionic gauging, transmutation, and Kramers-Wannier dualities

We generalize the gauging of $\mathbb{Z}_2$ symmetries by inserting Majorana fermions, establishing parallel duality correspondences for bosonic and fermionic lattice systems. Using this fermionic gauging, we construct fermionic analogs of $\mathbb{Z}_2$ gauge theories dual to the transverse-field Ising model, interpretable as Majorana stabilizer codes. We demonstrate a unitary equivalence between the $\mathbb{Z}_2$ gauge theory obtained by gauging the fermion parity of a free fermionic system and the conventional $\mathbb{Z}_2$ gauge theory with potentially nonlocal terms on the square lattice with toroidal geometry. This equivalence is implemented by a linear-depth local unitary circuit, connecting the bosonic and fermionic toric codes through a direction-dependent anyonic transmutation. The gauge theory obtained by gauging fermion parity is further shown to be equivalent to a folded Ising chain obtained via the Jordan--Wigner transformation. We clarify the distinction between the recently proposed Kramers--Wannier dualities and those obtained by gauging the $\mathbb{Z}_2$ symmetry along a space-covering path. Our results extend naturally to higher-dimensional $\mathbb{Z}_2$ lattice gauge theories, providing a unified framework for bosonic and fermionic dualities and offering new insights for quantum computation and simulation.

cond-mat.str-el↗

arXiv2025

Bosonization and Kramers-Wannier dualities in general dimensions

It is well known that the noninteracting Majorana chain is dual to the one-dimensional transverse-field Ising model, either through the Jordan-Wigner transformation or by gauging fermion parity. In this correspondence, the minimal translation of the Majorana chain maps to the celebrated Kramers-Wannier (KW) duality of the spin model, with the critical point mapped to the self-dual point. In this work, we generalize this mapping to two and higher dimensions by constructing a unitary equivalence between the parity-gauged fermionic system and a spin system defined on arbitrary polyhedral decompositions of space. Imposing the flatness condition on the gauge field yields a bosonization duality between the original (ungauged) fermionic system and a gauged spin system obeying a Gauss law. The dependence of the Gauss law in the spin system on the Kasteleyn orientation (and the discrete spin structure) of the fermionic system is made explicit. Applying this bosonization to one or two copies of Majorana fermions on translationally invariant lattices, we derive higher-dimensional analogs of KW (self-)dualities in spin systems arising from fermionic minimal translations. The KW (self-)dualities are non-invertible due to projections onto eigenspaces of higher-form symmetries in the associated symmetry operators. The bosonization framework we present is intuitive, general, and systematic, encompassing other known exact bosonization methods while offering a novel approach to establish new connections between fermionic and spin systems in arbitrary dimensions.

cond-mat.str-el↗

arXiv2025

Quantum Mpemba effect without global symmetries

The Mpemba effect, where a system initially farther from equilibrium relaxes faster than one closer to equilibrium, has been extensively studied in classical systems and recently explored in quantum settings. While previous studies of the quantum Mpemba effect (QME) have largely focused on isolated systems with global symmetries, we argue that the QME is ubiquitous in generic, non-integrable many-body systems lacking such symmetries, including U(1) charge conservation, spatial symmetries, and even energy conservation. Using paradigmatic models such as the quantum Ising model with transverse and longitudinal fields, we show that the QME can be understood through the energy density of initial states and their inverse participation ratio in the energy eigenbasis. Our findings provide a unified framework for the QME, linking it with classical thermal relaxation.

quant-ph↗

arXiv2025

GNCAF: A GNN-based Neighboring Context Aggregation Framework for Tertiary Lymphoid Structures Semantic Segmentation in WSI

Tertiary lymphoid structures (TLS) are organized clusters of immune cells, whose maturity and area can be quantified in whole slide image (WSI) for various prognostic tasks. Existing methods for assessing these characteristics typically rely on cell proxy tasks and require additional post-processing steps. In this work, We focus on a novel task-TLS Semantic Segmentation (TLS-SS)-which segments both the regions and maturation stages of TLS in WSI in an end-to-end manner. Due to the extensive scale of WSI and patch-based segmentation strategies, TLS-SS necessitates integrating from neighboring patches to guide target patch (target) segmentation. Previous techniques often employ on multi-resolution approaches, constraining the capacity to leverage the broader neighboring context while tend to preserve coarse-grained information. To address this, we propose a GNN-based Neighboring Context Aggregation Framework (GNCAF), which progressively aggregates multi-hop neighboring context from the target and employs a self-attention mechanism to guide the segmentation of the target. GNCAF can be integrated with various segmentation models to enhance their ability to perceive contextual information outside of the patch. We build two TLS-SS datasets, called TCGA-COAD and INHOUSE-PAAD, and make the former (comprising 225 WSIs and 5041 TLSs) publicly available. Experiments on these datasets demonstrate the superiority of GNCAF, achieving a maximum of 22.08% and 26.57% improvement in mF1 and mIoU, respectively. Additionally, we also validate the task scalability of GNCAF on segmentation of lymph node metastases.

eess.IV↗

arXiv2025

$\mathbb{Z}_2$ gauging and self-dualities of the $XX$ model and its cousins

In this work, we investigate the one-dimensional $XX$ lattice model and its cousins through the lens of momentum and winding $U(1)$ symmetries. We distinguish two closely related $\mathbb{Z}_2$ symmetries based on their relation to the $U(1)$ symmetries, and establish a web of $\mathbb{Z}_2$-gauging relations among these models, rooted in two fundamental seeds: the $ XY \pm YX$ models. These two seeds, each self-dual under gauging of the respective $\mathbb{Z}_2$-symmetries, possess manifestly symmetric conserved charges, making transparent the connection between the noninvertible symmetries and the Kramers-Wannier duality. By leveraging the self-dualities of these two seed models, we derive the self-dualities of their cousins, including the $XX$ model and the Levin-Gu model, through appropriate gauging procedures. Moreover, under these gauging schemes, the lattice T-duality matrices take the form of the identity matrix. These lattice models flow to the $c =1$ compact boson conformal field theory, with a twist that depends on the lattice size modulo four. Finally, we unify the mapping structures of local conserved charges across these models, providing a comprehensive framework for understanding their symmetries and dualities.

cond-mat.str-el↗

arXiv2025

Baichuan-M1: Pushing the Medical Capability of Large Language Models

The current generation of large language models (LLMs) is typically designed for broad, general-purpose applications, while domain-specific LLMs, especially in vertical fields like medicine, remain relatively scarce. In particular, the development of highly efficient and practical LLMs for the medical domain is challenging due to the complexity of medical knowledge and the limited availability of high-quality data. To bridge this gap, we introduce Baichuan-M1, a series of large language models specifically optimized for medical applications. Unlike traditional approaches that simply continue pretraining on existing models or apply post-training to a general base model, Baichuan-M1 is trained from scratch with a dedicated focus on enhancing medical capabilities. Our model is trained on 20 trillion tokens and incorporates a range of effective training methods that strike a balance between general capabilities and medical expertise. As a result, Baichuan-M1 not only performs strongly across general domains such as mathematics and coding but also excels in specialized medical fields. We have open-sourced Baichuan-M1-14B, a mini version of our model, which can be accessed through the following links.

cs.CL↗

arXiv2025

Memory Analysis on the Training Course of DeepSeek Models

We present a theoretical analysis of GPU memory consumption during the training of DeepSeek models such as DeepSeek-v2 and DeepSeek-v3. Our primary objective is to clarify the device-level memory requirements associated with various distributed training configurations. Specifically, we examine critical factors influencing memory usage, including micro-batch size, activation recomputation policies, 3D parallelism, and ZeRO optimizations. It is important to emphasize that the training policies discussed in this report are not representative of DeepSeek's official configurations. Instead, they are explored to provide a deeper understanding of memory dynamics in training of large-scale mixture-of-experts model.

cs.PF↗

arXiv2025

Baichuan-Omni-1.5 Technical Report

We introduce Baichuan-Omni-1.5, an omni-modal model that not only has omni-modal understanding capabilities but also provides end-to-end audio generation capabilities. To achieve fluent and high-quality interaction across modalities without compromising the capabilities of any modality, we prioritized optimizing three key aspects. First, we establish a comprehensive data cleaning and synthesis pipeline for multimodal data, obtaining about 500B high-quality data (text, audio, and vision). Second, an audio-tokenizer (Baichuan-Audio-Tokenizer) has been designed to capture both semantic and acoustic information from audio, enabling seamless integration and enhanced compatibility with MLLM. Lastly, we designed a multi-stage training strategy that progressively integrates multimodal alignment and multitask fine-tuning, ensuring effective synergy across all modalities. Baichuan-Omni-1.5 leads contemporary models (including GPT4o-mini and MiniCPM-o 2.6) in terms of comprehensive omni-modal capabilities. Notably, it achieves results comparable to leading models such as Qwen2-VL-72B across various multimodal medical benchmarks.

cs.CL↗

arXiv2024

Can We Get Rid of Handcrafted Feature Extractors? SparseViT: Nonsemantics-Centered, Parameter-Efficient Image Manipulation Localization through Spare-Coding Transformer

Non-semantic features or semantic-agnostic features, which are irrelevant to image context but sensitive to image manipulations, are recognized as evidential to Image Manipulation Localization (IML). Since manual labels are impossible, existing works rely on handcrafted methods to extract non-semantic features. Handcrafted non-semantic features jeopardize IML model's generalization ability in unseen or complex scenarios. Therefore, for IML, the elephant in the room is: How to adaptively extract non-semantic features? Non-semantic features are context-irrelevant and manipulation-sensitive. That is, within an image, they are consistent across patches unless manipulation occurs. Then, spare and discrete interactions among image patches are sufficient for extracting non-semantic features. However, image semantics vary drastically on different patches, requiring dense and continuous interactions among image patches for learning semantic representations. Hence, in this paper, we propose a Sparse Vision Transformer (SparseViT), which reformulates the dense, global self-attention in ViT into a sparse, discrete manner. Such sparse self-attention breaks image semantics and forces SparseViT to adaptively extract non-semantic features for images. Besides, compared with existing IML models, the sparse self-attention mechanism largely reduced the model size (max 80% in FLOPs), achieving stunning parameter efficiency and computation reduction. Extensive experiments demonstrate that, without any handcrafted feature extractors, SparseViT is superior in both generalization and efficiency across benchmark datasets.

cs.CV↗

arXiv2024

Mesoscopic Insights: Orchestrating Multi-scale & Hybrid Architecture for Image Manipulation Localization

The mesoscopic level serves as a bridge between the macroscopic and microscopic worlds, addressing gaps overlooked by both. Image manipulation localization (IML), a crucial technique to pursue truth from fake images, has long relied on low-level (microscopic-level) traces. However, in practice, most tampering aims to deceive the audience by altering image semantics. As a result, manipulation commonly occurs at the object level (macroscopic level), which is equally important as microscopic traces. Therefore, integrating these two levels into the mesoscopic level presents a new perspective for IML research. Inspired by this, our paper explores how to simultaneously construct mesoscopic representations of micro and macro information for IML and introduces the Mesorch architecture to orchestrate both. Specifically, this architecture i) combines Transformers and CNNs in parallel, with Transformers extracting macro information and CNNs capturing micro details, and ii) explores across different scales, assessing micro and macro information seamlessly. Additionally, based on the Mesorch architecture, the paper introduces two baseline models aimed at solving IML tasks through mesoscopic representation. Extensive experiments across four datasets have demonstrated that our models surpass the current state-of-the-art in terms of performance, computational complexity, and robustness.

cs.CV↗

arXiv2024

GCUNet: A GNN-Based Contextual Learning Network for Tertiary Lymphoid Structure Semantic Segmentation in Whole Slide Image

We focus on tertiary lymphoid structure (TLS) semantic segmentation in whole slide image (WSI). Unlike TLS binary segmentation, TLS semantic segmentation identifies boundaries and maturity, which requires integrating contextual information to discover discriminative features. Due to the extensive scale of WSI (e.g., 100,000 \times 100,000 pixels), the segmentation of TLS is usually carried out through a patch-based strategy. However, this prevents the model from accessing information outside of the patches, limiting the performance. To address this issue, we propose GCUNet, a GNN-based contextual learning network for TLS semantic segmentation. Given an image patch (target) to be segmented, GCUNet first progressively aggregates long-range and fine-grained context outside the target. Then, a Detail and Context Fusion block (DCFusion) is designed to integrate the context and detail of the target to predict the segmentation mask. We build four TLS semantic segmentation datasets, called TCGA-COAD, TCGA-LUSC, TCGA-BLCA and INHOUSE-PAAD, and make the former three datasets (comprising 826 WSIs and 15,276 TLSs) publicly available to promote the TLS semantic segmentation. Experiments on these datasets demonstrate the superiority of GCUNet, achieving at least 7.41% improvement in mF1 compared with SOTA.

cs.CV↗

arXiv2024

Comprehensive Performance Evaluation of YOLOv11, YOLOv10, YOLOv9, YOLOv8 and YOLOv5 on Object Detection of Power Equipment

With the rapid development of global industrial production, the demand for reliability in power equipment has been continuously increasing. Ensuring the stability of power system operations requires accurate methods to detect potential faults in power equipment, thereby guaranteeing the normal supply of electrical energy. In this article, the performance of YOLOv5, YOLOv8, YOLOv9, YOLOv10, and the state-of-the-art YOLOv11 methods was comprehensively evaluated for power equipment object detection. Experimental results demonstrate that the mean average precision (mAP) on a public dataset for power equipment was 54.4%, 55.5%, 43.8%, 48.0%, and 57.2%, respectively, with the YOLOv11 achieving the highest detection performance. Moreover, the YOLOv11 outperformed other methods in terms of recall rate and exhibited superior performance in reducing false detections. In conclusion, the findings indicate that the YOLOv11 model provides a reliable and effective solution for power equipment object detection, representing a promising approach to enhancing the operational reliability of power systems.

cs.CV↗