Network automation is critical for Intelligent Internet of Things (IIoT) systems, such as industrial networks and smart cities, where managing heterogeneous devices at scale poses significant challenges. Intent-Based Networking (IBN) powered by Large Language Models (LLMs) enables natural language intent translation into executable policies, yet struggles with accuracy in few-shot In-Context Learning (ICL) due to inefficient example retrieval for multi-intent scenarios. To address this issue, we propose SubmodICL, a submodular optimization-based method that retrieves contextual examples to maximize query-guided mutual information, thereby ensuring both coverage and representativeness. Integrated with LLMs, SubmodICL significantly improves intent translation performance. We validate our approach through a use case translating flow-table operation intents into API calls, and evaluate LLMs across different parameter scales. Experiments demonstrate that SubmodICL outperforms baselines by 9.121%–12.377% in accuracy for resource-constrained settings (small LLMs, limited examples, or multi-intent tasks). The proposed method provides a generalizable framework for enhancing intent translation in network automation, with potential applicability to IIoT systems.

Sentiment analysis is an important task in social computing and behavior analysis, and is a typical indicator of social health. It is a challenging mission to predict the sentiment of people in socially implemented Internet of Medical Things (IoMT) systems. The existing methods have several defects, and a typical defect is that most methods ignore the fact that there is much noise in IoMT systems and it is far not enough only to develop classification models for one type of data. In socially implemented IoMT systems, many methods treat the review text as plain text but ignore the potential knowledge structure. To solve those problems, in this article, we propose a novel solution, which is composed of adversarial learning and a hierarchical attention mechanism. We construct a hierarchical attention mechanism to learn the knowledge structure of a text. We propose to apply the attention mechanism both at the word level and sentence level, enabling us to learn the knowledge from each word and each sentence. We propose to use adversarial learning to learn new knowledge as non-random perturbations, which promotes the model’s robustness. We evaluate our method on several large-scale real-world datasets, covering a wide range of cases of sentiment analysis. Experimental results demonstrate that our method achieves superior performance compared to state-of-the-art methods.

Using the increasingly wireless communication capacity of 5G/6G technology, edge intelligence (EI) enables modern vehicles to leverage the powerful computing resources of edge servers scattered aside the roads to implement intelligent transportation applications (ITA). The running of these intelligent applications is always accompanied by the calculating and transmitting of massive amounts of road situation data. For the protection of drivers and passengers, these complex processes should be handled in an ultra-reliable and low latent manner. Intelligent processing function offloading from terminals to edge servers is a promising approach to address these issues. However, the runtime environment specification of each intelligent processing function and the interdependence between two consecutive functions pose a challenge for the assignment of the offloaded functions among edge servers. In this paper, we propose a dependence-aware edge intelligent function offloading scheme for 6G-based Internet of Vehicle (IoV). All traditional ITAs are split into different chains of standard intelligent functions. Each edge server can provide some specific intelligent functional services. These services can receive data from cars and serve as different intelligent functions. Then, an intelligent application offloading scheme is changed into an embedding scheme of a service chain. An NP-hard objective function is constructed using a multi-winner committee selection model for this offloading service chain embedding problem. We design two algorithms to get the optimal assignment of intelligent functions using a greedy strategy and dynamic programming strategy separately. Finally, experiments show that when the proportion of vehicles meeting the constraint conditions is not in [9%, 10%], our algorithms are fast.

Multidocument summarization problem deals with extracting main information and ideas from a set of related documents. Solution to this problem is to find an extraction strategy that aims at finding a small subset of sentences that is able to cover the most important information about the whole document set. Although a large number of machine-learning-based methods have shown great promise, the lack of high-quality training data poses an inherent obstacle to them. Furthermore, because of the proliferation of low-quality documents on the Internet, the existing summarization strategies, which are merely based on statistical features, get poor performance. In this article, we propose a new two-phase multidocument summarization strategy using content attention-based subtopic detection. First, inspired by distance dynamics-based community detection mechanism, we extract subtopics from the set of documents by having insight into their own content attention and also underlying semantic relations. Instead of complicated neural attention mechanisms, we propose a simple iteration-based content attention method to complete the subtopic detection task. Second, we formulate summarization from different subtopics as a combinatorial optimization problem of minimizing sentence distance and maximizing topic diversity. We prove the submodularity of the above optimization problem, which allows us to propose a new multidocument summarization algorithm based on the greedy mechanism. Finally, we experimentally validate our new algorithms on BBC news summary and wikiHow data. The results show our new algorithms outperform the state-of-the-art methods.

Mobile edge computing system provides cloud computing capabilities at the edge of wireless mobile networks, ensuring low latency, highly efficient computing, and improved user experience. At the same time, computationally intensive components are offloaded from mobile devices to edge servers and distributed among the servers. Due to the special constraints (mobile devices' battery capacities, limited computing resources of one single edge server, inevitable edge server failure, etc.), there emerges a following problem. 1) How to guarantee the reliability of the offloaded computing? This problem brings in the following two other problems. 2) How to find the appropriate offloading point in the mobile program such that the computing tasks offloaded to cloud can be maximized, while the transmission energy consumption is minimized? 3) What is the achievable minimum latency tasks allocation strategy among multiple users' mobile devices and multiple edge servers? In this paper, we try to address the aforementioned problems. First, for the appropriate offloading point problem, we consider the offloading valuable basic constraint and propose a task merging strategy based on mobile program component call graphs to minimize the computational complexity of the program partition. Second, we formulate the second problem as a combinatorial optimization problem and transform it into an n-fold integer programming problem by mapping the remaining computing resources to a virtual component. Third, we design a reliable shadow component scheme between multilevel severs for the reliability problem. Finally, we develop a fast algorithm for the mix problem and analyze its performance and conduct experiments to prove the accuracy of our theoretical results.

As vehicles are becoming more and more intelligent, mobile data traffic in vehicular ad hoc network (VANET) has been increasing dramatically. This makes the communication capacity of VANET systems and the computing resources of vehicles insufficient. In the meantime, location-aware large-scale distributed services with very low latency and high reliability are demanded by most of the novel functions, such as accident alarming, and congestion warning, in the intelligent transportation system. To meet these claimed characteristics of VANET, we first present a novel architecture that integrates vehicular fog computing and vehicle-to-vehicle (V2V) communication technologies. Lower latency and higher quality services can be supplied to vehicles by nearby fog servers, which are virtualized from vehicles that locate close enough and communicate using the V2V link. However, like all collaborative systems, computing reliability is vital to collaborative VANET. In this article, we design a novel energy-efficient proactive replication mechanism. Follower vehicles calculate with a lazy rate act as backups of host vehicles to ensure the reliability of the system. Considering the time sensitivity of computing requirements in VANET, the upper bound on the total number of failures is proposed through theoretical analysis. Then, the lower bound on the lazy calculating rate of followers is derived by balancing the tradeoffs between delay and energy. A fast algorithm for searching this lower bound based on the discrete Newton method is also proposed. Results of numerical experiments show that our new mechanism is effective in energy saving and reliability enhancing.

New age smartphones are equipped with high processing power and internet connectivity. Hence, smartphones are capable of executing applications, which were only possible by desktops or laptops until recently. Some examples of such applications are email, banking, flight booking etc. People prefer to use mobile devices for these applications due to the usability and portability of mobile devices. However, because of hardware limitations, mobiles have limited resources such as battery life, power and capacity. Researchers are constantly looking for ways to maximize the usage of these resources. The execution of any application on mobile, needs storage capacity of mobile to store, battery life of mobile to keep running and processing capacity of mobile to process. Thus, more resources are needed to run more applications on these devices. To reduce the load of applications on mobile devices and use the resources efficiently, it is necessary to move some load of applications to remote cloud server in such a way that the applications will run seamlessly. Computation offloading for mobile-edge computing MEC) is a mechanism to utilize mobile resources well by moving resource-intensive applications to cloud server at network edge. In the case of multiple users, the total computing capacity of the server needs to be taken into consideration for allocating resources to multiple users. The key to efficient computation offloading is allocating applications to mobile and remote server in such a way that minimizes transmission energy. In this paper, we formulate the computation offloading problem as graph cut problem and propose a solution based on spectral clustering computation. First, for the applications on mobile a corresponding network flow graph model is defined. Then, label propagation theory is applied on the network graph and the network graph is simplified by compressing and combining. Finally, the optimal solution is obtained by computation using spectral clustering algorithm. Experiments show that the algorithm is effective in handling programs with loosely coupled as well as highly coupled functions.