Federated Learning (FL) is a distributed deep learning framework that does not require collecting raw data from clients and performs specific tasks through distributed devices. Due to the heterogeneity caused by different data distributions between clients, clients tend to shift towards the local optimal solution during local iterative training and generate specific local models. Merging specific local models may deviate from the global optimal solution. This phenomenon will hinder the performance of the global model. Parameter regularization solves the problem of offset of the local optimal solution by restricting the update direction of the local model. Research on efficient parameter tuning and sparse federated learning has found that it is not necessary for the client to update all parameters in the global model during each round of training.Therefore, in this work, we first designed a novel sparse federated learning method: Fedl1/2, which can alleviate the degradation of FL performance by only updating some parameters in each training round. Usingl1/2regularization to control the update direction of each client while avoiding unnecessary parameter updates. To our knowledge, our study is the first to introducel1/2regularization in the field of FL to correct the local training of clients in FL. At the same time, we also presented Fedl1/2-ElasticNet, which extends the elastic network ofl1/2regularization in the FL field. In addition, considering that existing regularization coefficient adjustment plans are mostly static and there is relatively little research on dynamics, we propose a single period dynamic linear cyclic plan for regularization coefficient adjustment (SLC) and combine it with the Fedl1/2-ElasticNet algorithm to propose the SLC-Fedl1/2-ElasticNet algorithm. The superiority of the proposed method is verified by comparing it with eight state-of-the-art baselines.

Large language models (LLMs) are naturally suitable for Chinese spelling check (CSC) task in few-shot scenarios due to their powerful semantic understanding and few-shot learning capabilities. Recent CSC research has begun to use LLMs as foundational models. However, most current datasets are primarily focused on errors generated during the text generation process, with little attention given to errors occurring in the modal conversion process. Furthermore, existing LLM-based CSC methods often rely on fixed prompt samples, which limits the performance of LLMs. Therefore, we propose a framework named RagID (Retrieval-Augment Generation and Iterative Discriminator Strategy). By utilizing semantic-based similarity search and an iterative discriminator mechanism, RagID can provide well-chosen prompt samples and reduce over-correction issues in LLM-based CSC. RagID demonstrates excellent effectiveness in few-shot scenarios. We conducted comprehensive experiments, and the results show that RagID achieves the best performance on dataset that include data from multiple domains and dataset containing modal conversion spelling errors. The dataset and method are available online.

A direction-aware augmented spatial keyword top-$k$query (DAT$k\text{Q}$) returns the top-$k$objects based on a ranking function that considers spatial distance, textual similarity, query numeric attributes, and query direction. When a user initiates a DAT$k\text{Q}$, some user-desired objects (missing objects) may not appear in the query result set, and then the user wonders why they do not appear, which is called the why-not question. This paper focuses on answering why-not questions on DAT$k$Qs. We first discuss how to obtain the refined query direction by analyzing the position relationship between missing objects and original query direction in Polar coordinates. Then a DAPC index structure is designed, which can cut down irrelevant search space based on not only conventional distance pruning, keyword pruning, and attribute pruning but also query direction pruning. Particularly, by comparing the position relationship between the query direction and the sector (sector ring) region segmented by the DAPC-based method, the search space that does not meet the query direction is pruned. In addition, we discuss the applicability of our scheme for handling why-not questions on regional spatial keyword queries (SKQ), ordinary direction-aware top-$k$SKQ queries and complex scoring SKQ queries. Finally, a series of experiments are conducted on two real datasets to show the efficiency of our DAPC-based method.

With the development of mobile Internet technology, the road network in the real world is becoming larger and more complex, and the real-time response to the shortest distance query has become a high requirement for many industrial applications. Many existing approaches, such as G-tree or G*-tree, have been proposed to answer such search problems, however, these approaches are not efficient enough in dealing with very large graphs. To this end, we propose a novel index called LG-tree, which partitions the large graph into sub-graphs, and then indexes these subgraphs using a balanced tree. For each leaf node of LG-tree, a Distance Inverted File (DIF) is constructed, and these DIFs preserve all the connectivity information between the border vertices on the original graph. Inspired by the state-of-the-art label methods, we propose a novel hierarchy computing approach for each border vertex of LG-tree. Based on DIF and the hierarchy of border vertices, the border vertex list for each border is established, which is convenient for us to calculate the shortest distance and path between a query vertex$v_{q}$and a target vertex$t$. Specifically, the number of levels of a vertex hierarchy increases dramatically as the graph size increases, and calculating vertex labels on a very large graph is an NP-hard problem. In order to further improve the calculation efficiency of the shortest distance between vertices, a heuristic method is proposed to limit the level of vertex hierarchy. In addition, a stage-based dynamic programming search method is proposed to divide all search situations between$v_{q}$and$t$into three cases to guarantee the efficiency of the shortest distance search. Extensive experiments are conducted to show that LG-tree and the stage-based method have better performance than the state-of-the-art approaches.

Structured publish/subscribe (pub/sub) is a promising technique adopted on kinds of vehicle applications of Internet of industrial vehicles (IoIV), which uses Boolean expressions to capture the items with thousands of different attributes, values and spatial locations, and then processes and analyzes the vast amounts of data collected to obtain users' interests. However, existing pub/sub work with Boolean expressions either ignores spatial requirement or focuses on Euclidean space. This paper aims to fill this gap by addressing the issue of fog-based spatial-textual pub/sub problem with Boolean expressions in IoIV. A novel hybrid index called RnetBE is proposed, which exquisitely organizes traffic network structure, Boolean expressions, and spatial information of subscriptions. And RnetBE can prune huge numbers of unqualified subscriptions based on both spatial constraint and Boolean expressions, thus achieving high efficiency in indexing and matching. Moreover, range-tree deletion and orderly group processing optimization techniques are proposed to save storage space and further improve the subscription pruning efficiency. Simulation results show that RnetBE and the proposed algorithm are efficient in terms of memory consumption and matching time.