Large language models (LLMs) have demonstrated impressive performance across various domains. However, for clinical diagnosis, higher expectations are required for LLM’s reliability and sensitivity: thinking like physicians and remaining sensitive to key medical information that affects diagnostic reasoning, as subtle variations can lead to different diagnosis results. Yet, existing works focus mainly on investigating the sensitivity of LLMs to irrelevant context and overlook the importance of key information. In this paper, we investigate the sensitivity of LLMs, i.e. GPT-3.5, GPT-4, Gemini, Claude3 and LLaMA2-7b, to key medical information by introducing different perturbation strategies. The evaluation results highlight the limitations of current LLMs in remaining sensitive to key medical information for diagnostic decision-making. The evolution of LLMs must focus on improving their reliability, enhancing their ability to be sensitive to key information, and effectively utilizing this information. These improvements will enhance human trust in LLMs and facilitate their practical application in real-world scenarios. Our code and dataset are available at https://github.com/chenwei23333/DiagnosisQA.

Current parameterized image representations embed visual information along the semantic boundaries and struggle to express the internal detailed texture structures of image components, leading to a lack of content consistency after image editing and driving. To address these challenges, this work proposes a novel parameterized representation based on hierarchical image proxy geometry, utilizing multi-layer hierarchically interrelated proxy geometric control points to embed multi-scale long-range structures and fine-grained texture details. The proposed representation enables smoother and more continuous interpolation during image rendering and ensures high-quality consistency within image components during image editing. Additionally, under the layer-wise representation strategy based on semantic-aware image layer decomposition, we enable decoupled image shape/texture editing of the targets of interest within the image. Extensive experimental results on image vectorization and editing tasks demonstrate that our proposed method achieves high rendering accuracy of general images, including natural images, with a significantly higher image parameter compression ratio, facilitating user-friendly editing of image semantic components.

Robust and efficient local feature matching plays a crucial role in applications such as SLAM and visual localization for robotics. Despite great progress, it is still very challenging to extract robust and discriminative visual features in scenarios with drastic lighting changes, low texture areas, or repetitive patterns. In this paper, we propose a new lightweight network called LiftFeat, which lifts the robustness of raw descriptor by aggregating 3D geometric feature. Specifically, we first adopt a pre-trained monocular depth estimation model to generate pseudo surface normal label, supervising the extraction of 3D geometric feature in terms of predicted surface normal. We then design a 3D geometry-aware feature lifting module to fuse surface normal feature with raw 2D descriptor feature. Integrating such 3D geometric feature enhances the discriminative ability of 2D feature description in extreme conditions. Extensive experimental results on relative pose estimation, homography estimation, and visual localization tasks, demonstrate that our LiftFeat outperforms some lightweight state-of-the-art methods. Code will be released at: https://github.com/lyp-deeplearning/LiftFeat.

This paper investigates the problem of optimal consensus of high-order nonlinear multi-agent systems under privacy constraints. To address the potential privacy leakage during inter-agent communication, the Paillier homomorphic encryption scheme is incorporated into the consensus control protocol. Two privacy-preserving mechanisms are proposed, based on direct state difference and weighted state difference, respectively. These mechanisms enable effective encrypted information exchange without disclosing the agents’ actual states. Theoretical analysis is conducted to evaluate the system’s privacy properties, proving that neither curious-but-honest agents within the system nor external eavesdroppers can infer the initial states of other agents. Simulation results verify the effectiveness and advantages of the proposed method.

Cervical optical coherence tomography (OCT) imaging serves as an effective diagnostic tool, and the development of deep learning classification models for OCT has the potential to enhance diagnosis. However, the complex imaging patterns of OCT data, significant noise, and the substantial domain gap from multi-center data result in high uncertainty and low accuracy in classification networks. To address these challenges, we propose a Multi-scale Prototype-Guided Diffusion learning method (MPGD), which is constructed with theMulti-scale Feature Condition (MFC),Diffusion-based Classification Calibrator (DCC), andMulti-scale Prototype Bank (MPB)modules. Specifically, MFC provides initial classification based on multi-scale features, DCC calibrates MFC's classification results through a diffusion model, and MPB refines DCC's visual guidance using prototypes obtained from clustering. Extensive experiments demonstrate that MPGD outperforms widely-used competitors for cervical OCT image classification, showing excellent generalization performance.

Optical Coherence Tomography (OCT) cervical images offer a sensitive and non-invasive solution for screening cervical cancer. Currently, utilizing deep learning techniques to interpret these images has emerged as a promising approach for computer-aided diagnosis. However, challenges in computational efficiency and privacy protection persist in this field. To address these issues, we propose an uncertainty-aware filtering before distillation (UFD) framework to compress the OCT cervical dataset in a trustworthy and efficient manner. Our approach involves pre-training an uncertainty proxy to evaluate the importance of each sample, allowing us to extract a valuable subset (i.e., coreset) from the original dataset. We then apply importance-weighted distribution matching for dataset distillation. The distribution matching loss is calculated as the distance between the class centers of the uncertainty-weighted coreset and the synthetic distilled dataset. Extensive experiments validate the feasibility and effectiveness of our proposed UFD framework. It is worth noting that after selecting the most representative samples from the original OCT dataset through coreset selection, the performance of the trained network is comparable to, or sometimes even slightly higher than, that of the network trained with all the original data. This indicates that automated dataset compression and cleaning techniques are crucial for improving image classification problems in medical imaging datasets, such as OCT and ultrasound, which are rich in noise and erroneous samples.

Denoising diffusion probabilistic models (DDPMs) exhibit significant potential in the realm of medical image segmentation. Nevertheless, current DDPM implementations rely on original image features as conditional information, thus lacking the ability to specifically emphasize edge information, a critical aspect in addressing the primary challenge of segmentation. Furthermore, the necessary semantic features for conditioning the diffusion process lack effective alignment with the noise embedding. To address the above issues, we propose a novel edge-and-mask integration-driven diffusion model (EMidDiff). Specifically, 1) an edge-and-mask condition strategy is proposed for the segmentation diffusion model to effectively leverage rich semantic features, particularly the edge feature. 2) A novel co-attention guidance block is designed to align the segmentation map and condition features. The experimental results on brain tumor segmentation and optic-cup segmentation underscore the effectiveness of our approach, surpassing the performance of some state-of-the-art segmentation diffusion models.

The utilization of optical coherence tomography (OCT) holds significant promise in the realm of cervical cancer screening. This study considers the characteristics of OCT, such as sample scarcity and noise, and introduces a noise adaptive diffusion model (NADM) for data augmentation. The NADM comprises two components: the Adaptive Blind De-noising Module (ABD) and the Class-Guided Diffusion Model (CDM). The CDM is responsible for generating high-quality cervical samples, while the ABD is designed to adaptively suppress noise. Extensive experiments are done to establish the performance and generalizability of NADM, based on data collected from three distinct hospitals. To the best of our knowledge, we are the first to employ a data augmentation technique utilizing generative models for the purpose of diagnosing cervical OCT images.

Pulmonary nodule classification from computerized tomography(CT) Scans is a vital task for the early screening of Lung cancers. The algorithm is aiming at distinguishing malignant pulmonary nodules, benign nodules and the ones with their subtypes. In this paper, we defined a detailed pulmonary nodule classification task considering 5 semantic labels. We are facing with a series of non-trival problems dealing with such a task. First, the available medical image data for training is quite limited. We enlarged the training dataset by cropping out three-dimension(3D) volume of each pulmonary nodule and generating 15 planes with different orientations from these volumes. Secondly, the global modeling ability of the existing convolutional neural network(CNN) based architectures can not meet the need of medical image analysis well. To learn discriminative abstract information, we down-sample feature maps between successive stages and adopt the BotNet-50 backbone which is a combination of ResNet backbone and self-attention modules. Such an architecture can extract local and non-local information in low-level and high-level layers, respectively. Last but not the least, the data distribution of training data and testing data don't share similar distribution in real-world multi-center medical image classification scenes. We assigned the samples with modified wights while calculating the loss value for optimization. The proposed method can eliminate the spurious correlation between features and labels. Experiments demonstrate the effectiveness of each component.

Web search is an essential way for humans to obtain information, but it's still a great challenge for machines to understand the contents of web pages.In this paper, we introduce the task of structural reading comprehension (SRC) on web.Given a web page and a question about it, the task is to find the answer from the web page.This task requires a system not only to understand the semantics of texts but also the structure of the web page.Moreover, we proposed Web-SRC, a novel Web-based Structural Reading Comprehension dataset.WebSRC consists of 400K question-answer pairs, which are collected from 6.4K web pages.Along with the QA pairs, corresponding HTML source code, screenshots, and metadata are also provided in our dataset.Each question in WebSRC requires a certain structural understanding of a web page to answer, and the answer is either a text span on the web page or yes/no.We evaluate various baselines on our dataset to show the difficulty of our task.We also investigate the usefulness of structural information and visual features.Our dataset and baselines have been publicly available 1 .

Recent years have witnessed a surge of interests of using neural topic models for automatic topic extraction from text, since they avoid the complicated mathematical derivations for model inference as in traditional topic models such as Latent Dirichlet Allocation (LDA).However, these models either typically assume improper prior (e.g.Gaussian or Logistic Normal) over latent topic space or could not infer topic distribution for a given document.To address these limitations, we propose a neural topic modeling approach, called Bidirectional Adversarial Topic (BAT) model, which represents the first attempt of applying bidirectional adversarial training for neural topic modeling.The proposed BAT builds a twoway projection between the document-topic distribution and the document-word distribution.It uses a generator to capture the semantic patterns from texts and an encoder for topic inference.Furthermore, to incorporate word relatedness information, the Bidirectional Adversarial Topic model with Gaussian (Gaussian-BAT) is extended from BAT.To verify the effectiveness of BAT and Gaussian-BAT, three benchmark corpora are used in our experiments.The experimental results show that BAT and Gaussian-BAT obtain more coherent topics, outperforming several competitive baselines.Moreover, when performing text clustering based on the extracted topics, our models outperform all the baselines, with more significant improvements achieved by Gaussian-BAT where an increase of near 6% is observed in accuracy.

Advances on deep generative models have attracted significant research interest in neural topic modeling.The recently proposed Adversarial-neural Topic Model models topics with an adversarially trained generator network and employs Dirichlet prior to capture the semantic patterns in latent topics.It is effective in discovering coherent topics but unable to infer topic distributions for given documents or utilize available document labels.To overcome such limitations, we propose Topic Modeling with Cycle-consistent Adversarial Training (ToMCAT) and its supervised version sToMCAT.ToMCAT employs a generator network to interpret topics and an encoder network to infer document topics.Adversarial training and cycle-consistent constraints are used to encourage the generator and the encoder to produce realistic samples that coordinate with each other.sToMCAT extends ToM-CAT by incorporating document labels into the topic modeling process to help discover more coherent topics.The effectiveness of the proposed models is evaluated on unsupervised/supervised topic modeling and text classification.The experimental results show that our models can produce both coherent and informative topics, outperforming a number of competitive baselines.

Social tie prediction is an important issue in social network analysis. Transfer learning is often used for social tie prediction to address the problem of insufficient labeled training data, since few users manually annotate their social relationships. In this paper, we propose TranGAN, a novel generative adversarial network (GAN) based transfer learning framework for social tie prediction, which leverages social theories as the common knowledge to bridge the source network and the target network. GAN helps augment the original data set by generating data samples that have a similar probability distribution to that of the original data, and the training of TranGAN converges faster compared to existing transfer learning models. We evaluate the performance of TranGAN with extensive experiments, and show that TranGAN outperforms traditional learning algorithms and existing transfer learning algorithm on several metrics, and is efficient for large-scale social networks.

To enable dynamic spectrum access, service providers with spare spectrum (sellers) trade with those who are in need of additional spectrum (buyers). In a spectrum market, the transaction result is essentially a matching between sellers and buyers. Though it is tempting to optimize the matching over certain utility functions, a stable matching is more desirable, since no participants have incentives to deviate from the matching result. Existing spectrum matching algorithms only consider the maximum number of channels a buyer can purchase, but ignore minimum spectrum requirement that is essential to support proper operation of wireless communications. In this paper, we present a new framework of spectrum matching with both maximum quota and minimum requirements. Different from conventional matching problems, the spectrum market poses distinctive challenges due to spectrum reusability. To tackle this problem, we design two novel algorithms that satisfy different stability criterion: Extended Deferred Acceptance (EDA) algorithm that is fair but wasteful and the Multistage Deferred Acceptance (MDA) algorithm that is non-wasteful but weakly fair. Both algorithms converge to an interference-free matching and guarantees the minimum spectrum requirement. The simulation results show that the two proposed algorithms can raise buyer happiness and the channel utilization.

To enable dynamic spectrum access, service providers with spare spectrum (sellers) trade with those who are in need of additional spectrum (buyers). In a spectrum market, the transaction result is essentially a match between sellers and buyers. Though it is tempting to optimize the matching over certain utility functions, a stable matching is more desirable, since it takes into account a diverse set of preferences of buyers and sellers, and produces a matching result which no participants have incentives to deviate from. While existing works on spectrum matching only consider the maximum number of channels a buyer can purchase, in real-world scenarios, the minimum spectrum requirement should be satisfied to support the proper operation of wireless communication. To address this issue, in this paper, we present a new framework of spectrum matching with both maximum and minimum requirements. Different from conventional matching problems, the spectrum market poses distinctive challenges due to spectrum reusability. Instead of being sold exclusively to just one buyer, the same channel can be reused by multiple buyers who are not interfering with each other. To tackle this problem, we design a novel algorithm, called Extended Deferred Acceptance (EDA), that converges to an interference-free matching and guarantees the minimum spectrum requirement. We theoretically prove the stability of the matching result. Our simulation results show that EDA can achieve a 100% coverage on the minimum requirements, while alternative benchmark algorithms fail to do so, and buyers are more satisfied with the matching result of EDA than that of alternative algorithms.