Resource-constrained Intelligent Internet of Things (IoT) environments often grapple with the challenges of security and efficiency. To this end, we present a collaborative cloud-edge IoT framework based on compressive learning (CL) and federated learning (FL), called FCL. The end sensors employ compressive sampling to simultaneously accomplish data dimensionality reduction and lightweight privacy protection. Subsequently, edge devices utilize CL algorithms for data training, and the resulting models are uploaded to the cloud server for global aggregation. Experimental results have validated the effectiveness of the proposed scheme, in which the Transformer-based FCL still achieves nearly 80% accuracy when the computation and communication overheads are reduced by 66% and 99%, respectively.

The recent advancement in the research fields of medical imaging and machine learning technologies greatly im- prove the accuracy of the content-based medical image retrie- val(Cbmir). Existing Cbmirmethods, however, often overlooks the effect of image details on retrieval accuracy, which is usually unsatisfactory. The Cbmirof high-definition CT image (CI) based on deep learning technology, combined with patient symptom information, can precisely assist physicians in disease diagnosis and treatment. To address this issue, we take liver CI as an example in this article, and propose a privacy-preservingComputer-AidedDiagnosis(Cad) method of the liver diseases based on the multi-modal clinical information (e.g., patients’ CIs and symptoms, etc). Four enabling techniques such as aWsslnmodel, a unified similarity measure of CIs, a DM-Treeindex and a lightweight privacy-preserving(LPP) scheme are pre- sented to facilitate theCadprocessing of the large liver CIs. Ex- tensive experiments demonstrate that our proposedCadme- thod outperform the state-of-the-art schemes by a large margin in terms of the retrieval accuracy and efficiency, respectively.

In this paper, we propose a generative model to automatically discover the hidden associations between topics words and opinion words. By applying those discovered hidden associations, we construct the opinion scoring models to extract statements which best express opinionists’ standpoints on certain topics. For experiments, we apply our model to the political area. First, we visualize the similarities and dissimilarities between Republican and Democratic senators with respect to various topics. Second, we compare the performance of the opinion scoring models with 14 kinds of methods to find the best ones. We find that sentences extracted by our opinion scoring models can effectively express opinionists’ standpoints.

Understanding how topics in scientific literature evolve is an interesting and important problem. Previous work simply models each paper as a bag of words and also considers the impact of authors. However, the impact of one document on another as captured by citations, one important inherent element in scientific literature, has not been considered. In this paper, we address the problem of understanding topic evolution by leveraging citations, and develop citation-aware approaches. We propose an iterative topic evolution learning framework by adapting the Latent Dirichlet Allocation model to the citation network and develop a novel inheritance topic model. We evaluate the effectiveness and efficiency of our approaches and compare with the state of the art approaches on a large collection of more than 650,000 research papers in the last 16 years and the citation network enabled by CiteSeerX. The results clearly show that citations can help to understand topic evolution better.

Modeling the behavior of bloggers is an important problem with various applications in recommender systems, targeted advertising, and event detection. In this paper, we propose three models by combining content, temporal, social dimensions: the general blogging-behavior model, the profile-based blogging-behavior model and the social- network and profile-based blogging-behavior model. The models are based on two regression techniques: Extreme Learning Machine (ELM), and Modified General Regression Neural Network (MGRNN). We choose one of the largest blogs, a political blog, DailyKos1, for our empirical evaluation. Experiments show that the social network and profile-based blogging behavior model with ELM regression techniques produce good results for the most active bloggers and can be used to predict blogging behavior.

Geometric constraint solving is a rapidly developing field, with applications in areas such as kinematics, molecular modeling, surveying, and geometric theorem proving. This paper has incorporated conic arcs into a geometric constraint solver and describes how to construct conic blending arcs from constraints using a rational parametric representation - rational quadratic Bezier which combines the separate cases of blending edges. There are some possible constraints: traverse a fixed point, or tangency to or distance from a line. Here a uniform rational Bezier representation has been developed first, then the paper presents two methods to solve tangency to or distance from a line.