Searcharxiv⌕ Search

arXiv subjects

Chen Jiang

Publications and source records attributed to Chen Jiang.

At least 55 records · Page 3Linked to original sources
arXiv2025

Exploiting Layer Normalization Fine-tuning in Visual Transformer Foundation Models for Classification

LayerNorm is pivotal in Vision Transformers (ViTs), yet its fine-tuning dynamics under data scarcity and domain shifts remain underexplored. This paper shows that shifts in LayerNorm parameters after fine-tuning (LayerNorm shifts) are indicative of the transitions between source and target domains; its efficacy is contingent upon the degree to which the target training samples accurately represent the target domain, as quantified by our proposed Fine-tuning Shift Ratio ($FSR$). Building on this, we propose a simple yet effective rescaling mechanism using a scalar $λ$ that is negatively correlated to $FSR$ to align learned LayerNorm shifts with those ideal shifts achieved under fully representative data, combined with a cyclic framework that further enhances the LayerNorm fine-tuning. Extensive experiments across natural and pathological images, in both in-distribution (ID) and out-of-distribution (OOD) settings, and various target training sample regimes validate our framework. Notably, OOD tasks tend to yield lower $FSR$ and higher $λ$ in comparison to ID cases, especially with scarce data, indicating under-represented target training samples. Moreover, ViTFs fine-tuned on pathological data behave more like ID settings, favoring conservative LayerNorm updates. Our findings illuminate the underexplored dynamics of LayerNorm in transfer learning and provide practical strategies for LayerNorm fine-tuning.

cs.CV↗
arXiv2025

Towards a Universal 3D Medical Multi-modality Generalization via Learning Personalized Invariant Representation

Variations in medical imaging modalities and individual anatomical differences pose challenges to cross-modality generalization in multi-modal tasks. Existing methods often concentrate exclusively on common anatomical patterns, thereby neglecting individual differences and consequently limiting their generalization performance. This paper emphasizes the critical role of learning individual-level invariance, i.e., personalized representation $\mathbb{X}_h$, to enhance multi-modality generalization under both homogeneous and heterogeneous settings. It reveals that mappings from individual biological profile to different medical modalities remain static across the population, which is implied in the personalization process. We propose a two-stage approach: pre-training with invariant representation $\mathbb{X}_h$ for personalization, then fine-tuning for diverse downstream tasks. We provide both theoretical and empirical evidence demonstrating the feasibility and advantages of personalization, showing that our approach yields greater generalizability and transferability across diverse multi-modal medical tasks compared to methods lacking personalization. Extensive experiments further validate that our approach significantly enhances performance in various generalization scenarios.

cs.CV↗
arXiv2025

The Noether inequality for threefolds and three moduli spaces with minimal volumes

We establish the Noether inequality \[\textrm{Vol}(X)\geq \frac{4}{3}p_g(X)-\frac{10}{3}\] for all projective $3$-folds $X$ of general type with geometric genus $5\leq p_g(X)\leq 10$ where $\textrm{Vol}(X)$ is the canonical volume. This result resolves all remaining cases of the Noether inequality for $3$-folds. We further investigate the moduli spaces of canonical $3$-folds with small genera and minimal volumes. For a $3$-fold of general type with geometric genus $2$ and with minimal canonical volume $\frac{1}{3}$, we prove that its canonical model is a hypersurface of degree $16$ in $\mathbb{P}(1,1,2,3,8)$, which gives an explicit description of its canonical ring. This implies that the coarse moduli space $\mathcal{M}_{\frac{1}{3}, 2}$, parametrizing all canonical $3$-folds with canonical volume $\frac{1}{3}$ and geometric genus $2$, is an irreducible unirational variety of dimension $189$. Parallel studies show that $\mathcal{M}_{1, 3}$ is irreducible, unirational, and $236$-dimensional, and that $\mathcal{M}_{2, 4}$ is irreducible, unirational, and $270$-dimensional. As being conceived, every member in these 3 families is simply-connected.

math.AG↗
arXiv2025

Structure-aware Semantic Discrepancy and Consistency for 3D Medical Image Self-supervised Learning

3D medical image self-supervised learning (mSSL) holds great promise for medical analysis. Effectively supporting broader applications requires considering anatomical structure variations in location, scale, and morphology, which are crucial for capturing meaningful distinctions. However, previous mSSL methods partition images with fixed-size patches, often ignoring the structure variations. In this work, we introduce a novel perspective on 3D medical images with the goal of learning structure-aware representations. We assume that patches within the same structure share the same semantics (semantic consistency) while those from different structures exhibit distinct semantics (semantic discrepancy). Based on this assumption, we propose an mSSL framework named $S^2DC$, achieving Structure-aware Semantic Discrepancy and Consistency in two steps. First, $S^2DC$ enforces distinct representations for different patches to increase semantic discrepancy by leveraging an optimal transport strategy. Second, $S^2DC$ advances semantic consistency at the structural level based on neighborhood similarity distribution. By bridging patch-level and structure-level representations, $S^2DC$ achieves structure-aware representations. Thoroughly evaluated across 10 datasets, 4 tasks, and 3 modalities, our proposed method consistently outperforms the state-of-the-art methods in mSSL.

cs.CV↗
arXiv2025

Numerical invariants of hyper-Kähler manifolds

We study various constraints on the Beauville quadratic form and the Huybrechts-Riemann-Roch polynomial for hyper-Kähler manifolds, mostly in dimension 6 and in the presence of an isotropic class. In an appendix, Chen Jiang proves that in general, the Huybrechts-Riemann-Roch polynomial can always be written as a linear combination with nonnegative coefficients of certain explicit polynomials with positive coefficients. This implies that the Huybrechts-Riemann-Roch polynomial satisfies a curious symmetry property

math.AG↗
arXiv2025

Stability of syzygy bundles on varieties of Picard number one

We give a criterion for slope-stability of the syzygy bundle of a globally generated ample line bundle on a smooth projective variety of Picard number $1$ in terms of Hilbert polynomial. As applications, we prove the stability of syzygy bundles on many varieties, such as smooth Fano or Calabi--Yau complete intersections, hyperkähler varieties of Picard number 1, abelian varieties of Picard number $1$, rational homogeneous varieties of Picard number 1, weak Calabi--Yau varieties of Picard number $1$ of dimension $\leq 4$, and Fano varieties of Picard number $1$ of dimension $\leq5$. Also we prove the stability of syzygy bundles on all hyperkähler varieties.

math.AG↗
arXiv2025

SD-MAD: Sign-Driven Few-shot Multi-Anomaly Detection in Medical Images

Medical anomaly detection (AD) is crucial for early clinical intervention, yet it faces challenges due to limited access to high-quality medical imaging data, caused by privacy concerns and data silos. Few-shot learning has emerged as a promising approach to alleviate these limitations by leveraging the large-scale prior knowledge embedded in vision-language models (VLMs). Recent advancements in few-shot medical AD have treated normal and abnormal cases as a one-class classification problem, often overlooking the distinction among multiple anomaly categories. Thus, in this paper, we propose a framework tailored for few-shot medical anomaly detection in the scenario where the identification of multiple anomaly categories is required. To capture the detailed radiological signs of medical anomaly categories, our framework incorporates diverse textual descriptions for each category generated by a Large-Language model, under the assumption that different anomalies in medical images may share common radiological signs in each category. Specifically, we introduce SD-MAD, a two-stage Sign-Driven few-shot Multi-Anomaly Detection framework: (i) Radiological signs are aligned with anomaly categories by amplifying inter-anomaly discrepancy; (ii) Aligned signs are selected further to mitigate the effect of the under-fitting and uncertain-sample issue caused by limited medical data, employing an automatic sign selection strategy at inference. Moreover, we propose three protocols to comprehensively quantify the performance of multi-anomaly detection. Extensive experiments illustrate the effectiveness of our method.

cs.CV↗
arXiv2025

Pushing the boundaries of asteroseismic individual frequency modelling: unveiling two evolved very-low metallicity red giants

Metal-poor stars are key to understanding the first stellar generation in the Galaxy. Asteroseismic characterisation of red giants has traditionally relied on global seismic parameters, not the full spectrum of individual oscillation modes. Here, we present the first characterisation of two evolved very metal-poor stars, including the detailed mixed-mode patterns. We demonstrate that incorporating individual frequencies into grid-based modelling of red-giant stars enhances its precision, enabling detailed studies of these ancient stars and allowing us to infer the stellar properties of two $[\mathrm{Fe}/\mathrm{H}]{\sim}{-}2.5$ dex Kepler stars: KIC 4671239 and KIC 7693833. Recent developments in both observational and theoretical asteroseismology allows for detailed studies of the complex oscillation pattern of evolved giants. We employ Kepler time series and surface properties from high-resolution spectroscopic data to asteroseismically characterise the two stars using the BAyesian STellar Algorithm, BASTA. Both stars show agreement between constraints from seismic and classical observables; an overlap unrecoverable when purely considering the global seismic parameters. KIC 4671239 and KIC 7693833 were determined to have masses of $0.78^{+0.04}_{-0.03}$ and $0.83^{+0.03}_{-0.01} M_{\odot}$ with ages of $12.1^{+1.6}_{-1.5}$ and $10.3^{+0.6}_{-1.4}$ Gyr, respectively. A $\sim10$% discrepancy between observed and modelled $ν_{\mathrm{max}}$ suggests a metallicity dependence of its scaling relation, leading to overestimated masses and incorrect age inferences for metal-poor stars. Utilising the full spectrum of individual oscillation modes, we circumvent the dependence on the asteroseismic scaling relations, providing direct constraints on the stars themselves, pushing the boundaries of state-of-the-art detailed modelling of evolved stars at metallicities far different from solar.

astro-ph.SR↗
arXiv2025

JoyType: A Robust Design for Multilingual Visual Text Creation

Generating images with accurately represented text, especially in non-Latin languages, poses a significant challenge for diffusion models. Existing approaches, such as the integration of hint condition diagrams via auxiliary networks (e.g., ControlNet), have made strides towards addressing this issue. However, diffusion models often fall short in tasks requiring controlled text generation, such as specifying particular fonts or producing text in small fonts. In this paper, we introduce a novel approach for multilingual visual text creation, named JoyType, designed to maintain the font style of text during the image generation process. Our methodology begins with assembling a training dataset, JoyType-1M, comprising 1 million pairs of data. Each pair includes an image, its description, and glyph instructions corresponding to the font style within the image. We then developed a text control network, Font ControlNet, tasked with extracting font style information to steer the image generation. To further enhance our model's ability to maintain font style, notably in generating small-font text, we incorporated a multi-layer OCR-aware loss into the diffusion process. This enhancement allows JoyType to direct text rendering using low-level descriptors. Our evaluations, based on both visual and accuracy metrics, demonstrate that JoyType significantly outperforms existing state-of-the-art methods. Additionally, JoyType can function as a plugin, facilitating the creation of varied image styles in conjunction with other stable diffusion models on HuggingFace and CivitAI. Our project is open-sourced on https://jdh-algo.github.io/JoyType/.

cs.CV↗
arXiv2025

BBScore: A Brownian Bridge Based Metric for Assessing Text Coherence

Measuring the coherence of text is a vital aspect of evaluating the quality of written content. Recent advancements in neural coherence modeling have demonstrated their efficacy in capturing entity coreference and discourse relations, thereby enhancing coherence evaluation. However, many existing methods heavily depend on static embeddings or focus narrowly on nearby context, constraining their capacity to measure the overarching coherence of long texts. In this paper, we posit that coherent texts inherently manifest a sequential and cohesive interplay among sentences, effectively conveying the central theme, purpose, or standpoint. To explore this abstract relationship, we introduce the "BBScore," a novel reference-free metric grounded in Brownian bridge theory for assessing text coherence. Our findings showcase that when synergized with a simple additional classification component, this metric attains a performance level comparable to state-of-the-art techniques on standard artificial discrimination tasks. We also establish in downstream tasks that this metric effectively differentiates between human-written documents and text generated by large language models under a specific domain. Furthermore, we illustrate the efficacy of this approach in detecting written styles attributed to diverse large language models, underscoring its potential for generalizability. In summary, we present a novel Brownian bridge coherence metric capable of measuring both local and global text coherence, while circumventing the need for end-to-end model training. This flexibility allows for its application in various downstream tasks.

cs.CL↗
arXiv2025

SHE-Net: Syntax-Hierarchy-Enhanced Text-Video Retrieval

The user base of short video apps has experienced unprecedented growth in recent years, resulting in a significant demand for video content analysis. In particular, text-video retrieval, which aims to find the top matching videos given text descriptions from a vast video corpus, is an essential function, the primary challenge of which is to bridge the modality gap. Nevertheless, most existing approaches treat texts merely as discrete tokens and neglect their syntax structures. Moreover, the abundant spatial and temporal clues in videos are often underutilized due to the lack of interaction with text. To address these issues, we argue that using texts as guidance to focus on relevant temporal frames and spatial regions within videos is beneficial. In this paper, we propose a novel Syntax-Hierarchy-Enhanced text-video retrieval method (SHE-Net) that exploits the inherent semantic and syntax hierarchy of texts to bridge the modality gap from two perspectives. First, to facilitate a more fine-grained integration of visual content, we employ the text syntax hierarchy, which reveals the grammatical structure of text descriptions, to guide the visual representations. Second, to further enhance the multi-modal interaction and alignment, we also utilize the syntax hierarchy to guide the similarity calculation. We evaluated our method on four public text-video retrieval datasets of MSR-VTT, MSVD, DiDeMo, and ActivityNet. The experimental results and ablation studies confirm the advantages of our proposed method.

cs.CV↗
arXiv2025

Aneumo: A Large-Scale Comprehensive Synthetic Dataset of Aneurysm Hemodynamics

Intracranial aneurysm (IA) is a common cerebrovascular disease that is usually asymptomatic but may cause severe subarachnoid hemorrhage (SAH) if ruptured. Although clinical practice is usually based on individual factors and morphological features of the aneurysm, its pathophysiology and hemodynamic mechanisms remain controversial. To address the limitations of current research, this study constructed a comprehensive hemodynamic dataset of intracranial aneurysms. The dataset is based on 466 real aneurysm models, and 10,000 synthetic models were generated by resection and deformation operations, including 466 aneurysm-free models and 9,534 deformed aneurysm models. The dataset also provides medical image-like segmentation mask files to support insightful analysis. In addition, the dataset contains hemodynamic data measured at eight steady-state flow rates (0.001 to 0.004 kg/s), including critical parameters such as flow velocity, pressure, and wall shear stress, providing a valuable resource for investigating aneurysm pathogenesis and clinical prediction. This dataset will help advance the understanding of the pathologic features and hemodynamic mechanisms of intracranial aneurysms and support in-depth research in related fields. Dataset hosted at https://github.com/Xigui-Li/Aneumo.

cs.CV↗
arXiv2024

BEOL Electro-Biological Interface for 1024-Channel TFT Neurostimulator with Cultured DRG Neurons

The demand for high-quality neurostimulation, driven by the development of brain-computer interfaces, has outpaced the capabilities of passive microelectrode-arrays, which are limited by channel-count and biocompatibility. This work proposes a back-end-of-line (BEOL) process for 1024-channel stimulator with bioelectrodes and waterproof encapsulation to stimulate dorsal root ganglion neurons. We introduce an active-matrix neurostimulator based on n-type low-temperature poly-silicon thin-film transistor, adding PEDOT:PSS and SU-8 as bioelectrodes and encapsulation. This enables precise stimulation of DRG neurons, addressing key challenges in neurostimulation systems.

q-bio.NC↗
arXiv2024

Robot Manipulation in Salient Vision through Referring Image Segmentation and Geometric Constraints

In this paper, we perform robot manipulation activities in real-world environments with language contexts by integrating a compact referring image segmentation model into the robot's perception module. First, we propose CLIPU$^2$Net, a lightweight referring image segmentation model designed for fine-grain boundary and structure segmentation from language expressions. Then, we deploy the model in an eye-in-hand visual servoing system to enact robot control in the real world. The key to our system is the representation of salient visual information as geometric constraints, linking the robot's visual perception to actionable commands. Experimental results on 46 real-world robot manipulation tasks demonstrate that our method outperforms traditional visual servoing methods relying on labor-intensive feature annotations, excels in fine-grain referring image segmentation with a compact decoder size of 6.6 MB, and supports robot control across diverse contexts.

cs.RO↗
arXiv2024

Effective nonvanishing for weighted complete intersections of codimension two

We show Kawamata's effective nonvanishing conjecture (also known as the Ambro--Kawamata nonvanishing conjecture) holds for quasismooth weighted complete intersections of codimension $2$. Namely, for a quasismooth weighted complete intersection $X$ of codimension $2$ and an ample Cartier divisor $H$ on $X$ such that $H-K_X$ is ample, the linear system $|H|$ is nonempty.

math.AG↗