Search

Inverse synthetic aperture radar (ISAR) super-resolution imaging technology is widely applied in space target imaging. However, the performance limits of super-resolution imaging algorithms remain a rarely explored issue. This paper investigates these limits by analyzing the boundaries of super-resolution algorithms for space targets and examines the relationships between key contributing factors. In particular, drawing on the established mathematical theory of computational resolution limits (CRL) for line spectrum reconstruction, we derive mathematical expressions for the upper and lower bounds of cross-range super-resolution imaging, based on ISAR imaging model transformations. Leveraging the explicit expressions, we first explore influencing factors of these bounds, such as the traditional Rayleigh limit, the number of scatterers, and the peak signal-to-noise ratio (PSNR) of scatterers. Then we elucidate the minimum resource requirements in ISAR imaging imposed by the CRL theory to meet the desired cross-range resolution, without which studying super-resolution algorithms becomes unnecessary in practice. Furthermore, the tradeoffs between the cumulative rotation angle, the radar transmit energy, and other contributing factors in optimizing the resolution are analyzed. Simulations are conducted to demonstrate these tradeoffs across various ISAR imaging scenarios, revealing their high dependence on specific imaging targets.

In this paper, we investigate the behavior of Igusa-Todorov distances, extension and Rouquier's dimensions under cleft extensions of abelian categories. Applications are given to Morita context rings, trivial extension rings and tensor rings., Comment: 18 pages

We present a novel approach for minimally invasive flexible needle manipulations by pairing a real-time finite element simulator with the cross-entropy method. Additionally, we demonstrate how a kinematic-driven bang-bang controller can complement the control framework for better tracking performance. We show how electromagnetic (EM) tracking can be readily incorporated into the framework to provide controller feedback. Tissue phantom experiment with EM tracking shows the average targeting error is $0.16 \pm 0.29mm$., Comment: Submitted to IEEE International Conference on Robotics and Automation 2025

Homomorphic vector permutation is fundamental to privacy-preserving computations based on batch-encoded homomorphic encryption, underpinning nearly all homomorphic matrix operation algorithms and predominantly influencing their complexity. A potential approach to optimize this critical component lies in permutation decomposition, a technique we consider as not yet fully explored. In this paper, we enhance the efficiency of homomorphic permutations through novel decomposition techniques, thus advancing privacy-preserving computations. We start by estimating the ideal performance of decompositions on permutations and proposing an algorithm that searches depth-1 ideal decomposition solutions. This enables us to ascertain the full-depth ideal decomposability of specific permutations in homomorphic matrix transposition (SIGSAC 18) and multiplication (CCSW 22), allowing these privacy-preserving computations to achieve asymptotic improvement in speed and rotation key reduction. We further devise a new method for computing arbitrary homomorphic permutations, aiming to approximate the performance of ideal decomposition, as permutations with weak structures are unlikely to be ideally factorized. Our design deviates from the conventional scope of permutation decomposition. It outperforms state-of-the-art techniques (EUROCRYPT 12, CRYPTO 14) with a speed-up of up to $\times2.27$ under the minimum requirement of rotation keys., Comment: Submission on 11/11/2024, context and authors' info updated

We introduced the notion of Orlov spectra of Abelian categories, and study its some properties. In particular, we give precise result of Orlov spectra of algebras with type $\mathbb{A}_{n}$., Comment: 28 pages; We provide examples for the definition of [U]_n, Definition 2.4, and Lemma 2.8

There is a large literature on cover-free families of finite sets, because of their many applications in combinatorial group testing, cryptographic and communications. This work studies the generalization of cover-free families from sets to finite vector spaces. Let $V$ be an $n$-dimensional vector space over the finite field $\mathbb{F}_{q}$ and let $\left[V\atop k\right]_q$ denote the family of all $k$-dimensional subspaces of $V$. A family $\mathcal{F}\subseteq \left[V\atop k\right]_q$ is called cover-free if there are no three distinct subspaces $F_{0}, F_{1}, F_{2}\in \mathcal{F}$ such that $F_{0}\leq (F_{0}\cap F_{1})+(F_{0}\cap F_{2})$. A family $\mathcal{H}\subseteq \left[V\atop k\right]_q$ is called a $q$-Steiner system $S_{q}(t, k, n)$ if for every $T\in \left[V\atop t\right]_q$, there is exactly one $H\in \mathcal{H}$ such that $T\leq H$. In this paper we investigate cover-free families in the vector space $V$. Firstly, we determine the maximum size of a cover-free family in $\left[V\atop k\right]_q$. Secondly, we characterize the structures of all maximum cover-free families which are closely related to $q$-Steiner systems., Comment: 25 pages

There is a well-known class of algebras called Igusa-Todorov algebras which were introduced in relation to finitistic dimension conjecture. As a generalization of Igusa-Todorov algebras, the new notion of $(m,n)$-Igusa-Todorov algebras provides a wider framework for studying derived dimensions. In this paper, we give methods for constructing $(m,n)$-Igusa-Todorov algebras. As an application, we present for general artin algebras a relationship between the derived dimension and the representation distance. Moreover, we end this paper to show that the main result can be used to give a better upper bound for the derived dimension for some classes of algebras.

This paper presents a novel approach for predicting molecular properties with high accuracy without the need for extensive pre-training. Employing ensemble learning and supervised fine-tuning of BERT, RoBERTa, and XLNet, our method demonstrates significant effectiveness compared to existing advanced models. Crucially, it addresses the issue of limited computational resources faced by experimental groups, enabling them to accurately predict molecular properties. This innovation provides a cost-effective and resource-efficient solution, potentially advancing further research in the molecular domain., Comment: 16 pages,7 figures

The Igusa-Todorov distances of an Artin algebra give a reasonable way of measuring how far an algebra is from being Igusa-Todorov algebras. We show that the Igusa-Todorov distance is an invariant under the singular equivalence of Morita type with level, and establish a series of bounds for the Igusa-Todorov distances of the algebras involved in a recollement of derived module categories., Comment: 29pages. arXiv admin note: text overlap with arXiv:2103.10393 by other authors

In this paper, our primary focus is on investigating the extension dimensions of syzygy module categories associated with Artin algebras, particularly under various equivalences. We demonstrate that, for sufficiently large $i$, the $i$-th syzygy module categories of derived equivalent algebras exhibit identical extension dimensions. Furthermore, we establish that the extension dimension of the $i$-th syzygy module category is an invariant under both stable equivalence and separable equivalence for each nonnegative integer $i$.

Educational chatbots are a promising tool for assisting student learning. However, the development of effective chatbots in education has been challenging, as high-quality data is seldom available in this domain. In this paper, we propose a framework for generating synthetic teacher-student interactions grounded in a set of textbooks. Our approaches capture one aspect of learning interactions where curious students with partial knowledge interactively ask a teacher questions about the material in the textbook. We highlight various quality criteria that such dialogues should fulfill and compare several approaches relying on either prompting or fine-tuning large language models. We use synthetic dialogues to train educational chatbots and show benefits of further fine-tuning in different educational domains. However, human evaluation shows that our best data synthesis method still suffers from hallucinations and tends to reiterate information from previous conversations. Our findings offer insights for future efforts in synthesizing conversational data that strikes a balance between size and quality. We will open-source our data and code., Comment: 24 pages, 19 tables, 2 figures

During Percutaneous Nephrolithotomy (PCNL) operations, the surgeon is required to define the incision point on the patient's back, align the needle to a pre-planned path, and perform puncture operations afterward. The procedure is currently performed manually using ultrasound or fluoroscopy imaging for needle orientation, which, however, implies limited accuracy and low reproducibility. This work incorporates Augmented Reality (AR) visualization with an optical see-through head-mounted display (OST-HMD) and Human-Robot Collaboration (HRC) framework to empower the surgeon's task completion performance. In detail, Eye-to-Hand calibration, system registration, and hologram model registration are performed to realize visual guidance. A Cartesian impedance controller is used to guide the operator during the needle puncture task execution. Experiments are conducted to verify the system performance compared with conventional manual puncture procedures and a 2D monitor-based visualisation interface. The results showed that the proposed framework achieves the lowest median and standard deviation error across all the experimental groups, respectively. Furthermore, the NASA-TLX user evaluation results indicate that the proposed framework requires the lowest workload score for task completion compared to other experimental setups. The proposed framework exhibits significant potential for clinical application in the PCNL task, as it enhances the surgeon's perception capability, facilitates collision-free needle insertion path planning, and minimises errors in task completion., Comment: Accepted by IEEE Robotics and Automation Magazine

Quantum entanglement plays a pivotal role in various quantum information processing tasks. However, there still lacks a universal and effective way to detecting entanglement structures, especially for high-dimensional and multipartite quantum systems. Noticing the mathematical similarities between the common representations of many-body quantum states and the data structures of images, we are inspired to employ advanced computer vision technologies for data analysis. In this work, we propose a hybrid CNN-Transformer model for both the classification of GHZ and W states and the detection of various entanglement structures. By leveraging the feature extraction capabilities of CNNs and the powerful modeling abilities of Transformers, we can not only effectively reduce the time and computational resources required for the training process but also obtain high detection accuracies. Through numerical simulation and physical verification, it is confirmed that our hybrid model is more effective than traditional techniques and thus offers a powerful tool for independent detection of multipartite entanglement.

Large language models (LLMs), such as ChatGPT, have received substantial attention due to their capabilities for understanding and generating human language. While there has been a burgeoning trend in research focusing on the employment of LLMs in supporting different medical tasks (e.g., enhancing clinical diagnostics and providing medical education), a review of these efforts, particularly their development, practical applications, and outcomes in medicine, remains scarce. Therefore, this review aims to provide a detailed overview of the development and deployment of LLMs in medicine, including the challenges and opportunities they face. In terms of development, we provide a detailed introduction to the principles of existing medical LLMs, including their basic model structures, number of parameters, and sources and scales of data used for model development. It serves as a guide for practitioners in developing medical LLMs tailored to their specific needs. In terms of deployment, we offer a comparison of the performance of different LLMs across various medical tasks, and further compare them with state-of-the-art lightweight models, aiming to provide an understanding of the advantages and limitations of LLMs in medicine. Overall, in this review, we address the following questions: 1) What are the practices for developing medical LLMs 2) How to measure the medical task performance of LLMs in a medical setting? 3) How have medical LLMs been employed in real-world practice? 4) What challenges arise from the use of medical LLMs? and 5) How to more effectively develop and deploy medical LLMs? By answering these questions, this review aims to provide insights into the opportunities for LLMs in medicine and serve as a practical resource. We also maintain a regularly updated list of practical guides on medical LLMs at https://github.com/AI-in-Health/MedLLMsPracticalGuide, Comment: Preprint. Version 6. Update Figures 1-5; Tables 2-3; 31 pages

Large Multimodal Models (LMMs) have demonstrated impressive performance across various vision and language tasks, yet their potential applications in recommendation tasks with visual assistance remain unexplored. To bridge this gap, we present a preliminary case study investigating the recommendation capabilities of GPT-4V(ison), a recently released LMM by OpenAI. We construct a series of qualitative test samples spanning multiple domains and employ these samples to assess the quality of GPT-4V's responses within recommendation scenarios. Evaluation results on these test samples prove that GPT-4V has remarkable zero-shot recommendation abilities across diverse domains, thanks to its robust visual-text comprehension capabilities and extensive general knowledge. However, we have also identified some limitations in using GPT-4V for recommendations, including a tendency to provide similar responses when given similar inputs. This report concludes with an in-depth discussion of the challenges and research opportunities associated with utilizing GPT-4V in recommendation scenarios. Our objective is to explore the potential of extending LMMs from vision and language tasks to recommendation tasks. We hope to inspire further research into next-generation multimodal generative recommendation models, which can enhance user experiences by offering greater diversity and interactivity. All images and prompts used in this report will be accessible at https://github.com/PALIN2018/Evaluate_GPT-4V_Rec., Comment: In Progress

Large Language Models (LLMs) have introduced a new era of proficiency in comprehending complex healthcare and biomedical topics. However, there is a noticeable lack of models in languages other than English and models that can interpret multi-modal input, which is crucial for global healthcare accessibility. In response, this study introduces Qilin-Med-VL, the first Chinese large vision-language model designed to integrate the analysis of textual and visual data. Qilin-Med-VL combines a pre-trained Vision Transformer (ViT) with a foundational LLM. It undergoes a thorough two-stage curriculum training process that includes feature alignment and instruction tuning. This method enhances the model's ability to generate medical captions and answer complex medical queries. We also release ChiMed-VL, a dataset consisting of more than 1M image-text pairs. This dataset has been carefully curated to enable detailed and comprehensive interpretation of medical data using various types of images.

Integrating large language models (LLMs) into healthcare holds great potential but faces challenges. Pre-training LLMs from scratch for domains like medicine is resource-heavy and often unfeasible. On the other hand, sole reliance on Supervised Fine-tuning (SFT) can result in overconfident predictions and may not tap into domain-specific insights. In response, we present a multi-stage training method combining Domain-specific Continued Pre-training (DCPT), SFT, and Direct Preference Optimization (DPO). In addition, we publish a 3Gb Chinese Medicine (ChiMed) dataset, encompassing medical question answering, plain texts, knowledge graphs, and dialogues, segmented into three training stages. The medical LLM trained with our pipeline, Qilin-Med, shows substantial performance improvement. In the CPT and SFT phases, Qilin-Med achieved 38.4% and 40.0% accuracy on the CMExam test set, respectively. It outperformed the basemodel Baichuan-7B (accuracy: 33.5%), by 7.5%. In the DPO phase, it scored 16.66 in BLEU-1 and 27.44 in ROUGE-1 on the Huatuo-26M test set, bringing further improvement to the SFT phase (12.69 in BLEU-1 and 24.21 in ROUGE-1). Additionally, we have further enhanced the model's performance through the Retrieval Augmented Generation (RAG) approach. Experiments demonstrate that Qilin-Med-RAG achieves an accuracy rate of 42.8% on CMExam. These results highlight the contribution of our novel training approach in building LLMs for medical applications.

Renal biopsies are the gold standard for diagnosis of kidney diseases. Lesion scores made by renal pathologists are semi-quantitative and exhibit high inter-observer variability. Automating lesion classification within segmented anatomical structures can provide decision support in quantification analysis and reduce the inter-observer variability. Nevertheless, classifying lesions in regions-of-interest (ROIs) is clinically challenging due to (a) a large amount of densely packed anatomical objects (up to 1000), (b) class imbalance across different compartments (at least 3), (c) significant variation in object scales (i.e. sizes and shapes), and (d) the presence of multi-label lesions per anatomical structure. Existing models lack the capacity to address these complexities efficiently and generically. This paper presents \textbf{a generalized technical solution} for large-scale, multi-source datasets with diverse lesions. Our approach utilizes two sub-networks: dense instance segmentation and lesion classification. We introduce \textbf{DiffRegFormer}, an end-to-end dense instance segmentation model designed for multi-class, multi-scale objects within ROIs. Combining diffusion models, transformers, and RCNNs, DiffRegFormer efficiently recognizes over 500 objects across three anatomical classes (glomeruli, tubuli, arteries) within ROIs on a single NVIDIA GeForce RTX 3090 GPU. On a dataset of 303 ROIs (from 148 Jones' silver-stained renal WSIs), it outperforms state of art models, achieving AP of 52.1\% (detection) and 46.8\% (segmentation). Our lesion classification sub-network achieves 89.2\% precision and 64.6\% recall on 21889 object patches (from the 303 ROIs). Importantly, the model demonstrates direct domain transfer to PAS-stained WSIs without fine-tuning., Comment: 16 pages, 15 figures, 6 tables, Journal