Federated learning (FL) provides a decentralized machine learning paradigm where a server collaborates with a group of clients to learn a global model without accessing the clients' data. User heterogeneity is a significant challenge for FL, which together with the class-distribution imbalance further enhances the difficulty of FL. Great progress has been made in large vision-language models, such as Contrastive Language-Image Pre-training (CLIP), which paves a new way for image classification and object recognition. Inspired by the success of CLIP on few-shot and zero-shot learning, we use CLIP to optimize the federated learning between server and client models under its vision-language supervision. It is promising to mitigate the user heterogeneity and class-distribution balance due to the powerful cross-modality representation and rich open-vocabulary prior knowledge. In this paper, we propose the CLIP-guided FL (CLIP2FL) method on heterogeneous and long-tailed data. In CLIP2FL, the knowledge of the off-the-shelf CLIP model is transferred to the client-server models, and a bridge is built between the client and server. Specifically, for client-side learning, knowledge distillation is conducted between client models and CLIP to improve the ability of client-side feature representation. For server-side learning, in order to mitigate the heterogeneity and class-distribution imbalance, we generate federated features to retrain the server model. A prototype contrastive learning with the supervision of the text encoder of CLIP is introduced to generate federated features depending on the client-side gradients, and they are used to retrain a balanced server classifier. Extensive experimental results on several benchmarks demonstrate that CLIP2FL achieves impressive performance and effectively deals with data heterogeneity and long-tail distribution. The code is available at https://github.com/shijiangming1/CLIP2FL.

Diffusion auction design is a new trend in mechanism design for which the main goal is to incentivize existing buyers to invite new buyers, who are their neighbors on a social network, to join an auction even though they are competitors. With more buyers, a diffusion auction will be able to give a more efficient allocation and receive higher revenue. Existing studies have proposed many interesting diffusion auctions to attract more buyers, but the seller’s revenue is not optimized. Hence, in this study, we investigate what optimal revenue the seller can achieve by attracting more buyers. Different from the traditional setting, the revenue that can be achieved in a diffusion auction highly relies on the structure of the network. Hence, we focus on optimal auctions with given classes of underlying networks. We propose a class of mechanisms, where for any given structure, an optimal diffusion mechanism can be found. We point out that it implies an idea of “reserve structure”. Moreover, we show that an optimal mechanism that handles all structures does not exist. Therefore, we also propose mechanisms that have bounded approximations of the optimal revenue in all structures.

In this paper, the distributed optimization problem is investigated for a class of general nonlinear model-free multi-agent systems. The dynamical model of each agent is unknown and only the input/output data are available. A model-free adaptive control method is employed, by which the original unknown nonlinear system is equivalently converted into a dynamic linearized model. An event-triggered consensus scheme is developed to guarantee that the consensus error of the outputs of all agents is convergent. Then, by means of the distributed gradient descent method, a novel event-triggered model-free adaptive distributed optimization algorithm is put forward. Sufficient conditions are established to ensure the consensus and optimality of the addressed system. Finally, simulation results are provided to validate the effectiveness of the proposed approach.

Despite profound impacts on human health and nature, accurately predicting microbe-metabolite interactions remains challenging due to inherent data noise. This study applies non-negative matrix factorization (NMF) and multi-view NMF to reduce noise and exploit associations across data perspectives. NMF obtains low-dimensional microbial and metabolic representations, effectively reducing noise. The dimension-reduced spectral matrices were input into the generative network model to derive conditional probabilities of individual microbe-associated metabolites and microbe-metabolite co-occurrence probabilities, the latter enabling prediction of microbe-metabolite interactions. Moreover, multi-view NMF integrates microbial and metabolic data by mapping them into a shared subspace, thereby enhancing prediction performance and validating cross-perspective correlation modeling. This study demonstrates NMF's efficacy in noise reduction through dimensionality reduction, and multiview NMF's ability to leverage cross-view associations. Both approaches demonstrate enhanced microbe-metabolite interaction prediction utilizing NMF-based and multi-view NMF-based generative network models.

Image classification for long-tailed scenarios has attracted more attention because its distribution is more similar to real-world image data. From the perspective of solving imbalance at the sample level, we propose a simple but effective method, named Sample-aware Knowledge Distillation (SAKD), which includes Selective Knowledge Distillation module and Stable Feature Center Learning module. The former conducts knowledge distillation at the sample-level by selecting samples, in which whether the sample needs to be distilled and to what extent is determined by evaluating the teacher network’s predictions for this sample. The latter is used to obtaining the stable feature center and making the feature center free from perturbation by hard samples, then further improving the classification boundary. We conduct extensive experiments on several long-tailed benchmark datasets and these results demonstrate that SAKD is effective. In addition, our SFCL module can be combined with other methods and also improve their performance.

In this paper, we study the financing decision for an emission-dependent supply chain with one supplier and one manufacturer, both of which are financially constrained and in need of short-term financing for emission abatement. Three kinds of financing strategies are considered: (1) bank financing separately (BFS), (2) the manufacturer obtains partial loans from the bank, and borrows from the supplier (MPB), (3) the supplier obtains partial loans from the bank, and borrows from the manufacturer (SPB). The BFS is viewed as a noncollaborative financing strategy, and the MPB and SPB are collaborative financing strategies. The financing strategies are obtained at the point where wholesale price, production quantity, and emission reduction are endogenously determined. We found that the collaborative financing strategy is beneficial to the supplier and the whole supply chain compared with the noncollaborative financing strategy, the SPB strategy may outperform the MPB strategy for the supplier and the whole supply chain, depending on sharing proportion. The advantage of the MPB and SPB strategies relative to the BFS strategy mainly depends on sharing proportion from the manufacturer's perspective. We demonstrated that the collaborative financing strategy is the equilibrium strategy of the emission-dependent supply chain under certain conditions.

Large-scale point cloud semantic segmentation has wide applications. Current popular researches mainly focus on fully supervised learning which demands expensive and tedious manual point-wise annotation. Weakly supervised learning is an alternative way to avoid this exhausting an-notation. However, for large-scale point clouds with few labeled points, the network is difficult to extract discriminative features for unlabeled points, as well as the regularization of topology between labeled and unlabeled points is usually ignored, resulting in incorrect segmentation results.To address this problem, we propose a perturbed self-distillation (PSD) framework. Specifically, inspired by self-supervised learning, we construct the perturbed branch and enforce the predictive consistency among the perturbed branch and original branch. In this way, the graph topology of the whole point cloud can be effectively established by the introduced auxiliary supervision, such that the in-formation propagation between the labeled and unlabeled points will be realized. Besides point-level supervision, we present a well-integrated context-aware module to explicitly regularize the affinity correlation of labeled points. Therefore, the graph topology of the point cloud can be further refined. The experimental results evaluated on three large-scale datasets show the large gain (3.0% on average) against recent weakly supervised methods and comparable results to some fully supervised methods.

Kinship verification is an interesting and important problem in the fields of computer vision. In practice, the biggest obstacle in kinship verification is that the representation capability of extracted features may not be powerful due to the significant differences between facial images of family members. To effectively address this problem, we propose a semi-coupled synthesis and analysis dictionary pair learning (SSADL) approach, which can reduce the differences between facial images. Specifically, SSADL jointly learns two view-specific synthesis-analysis dictionary pairs as well as a mapping matrix from the training data of parent and child, with which, the heterogeneous facial images of parent and child can be transformed into coding coefficients of the same subspace, such that the kinship verification task can be conducted using the coding coefficients. Besides, we also design a hard sample based coefficient discriminant term to ensure that the obtained coefficients own favorable discriminability. Experimental results on several publicly used benchmarks show the effectiveness of our proposed approach.

Virus plays an important role in the earth's ecosystem. It maintains the balance of the ecosystem. It infects host cells, causing damage or death to the host. Understanding the relationship between virus and host is the key to preventing viral diseases. There are a large number of proven relationships between viruses and hosts in the literature. Extracting the relationships between viruses and hosts in these literature and organizing them into a virus-host knowledge base is of great significance to medical and biological research. Virus named entity recognition (VNER) is the key prerequisite step of relationship extraction. The complexity of virus naming and classification makes the identification of virus named entities challenging. In this paper, we provide a labeled corpus for the task of VNER. Furthermore, we use different pre-training models to compare its performance on downstream virus entity recognition tasks. Finally, BioBERT_ BiLSTM_ CRF got best result on the task of VNER. The precision value, recall value and F1 value are 92.18%, 91.28% and 91.85%, respectively.

Ad hoc networks rely on the mutual cooperation among individual nodes to achieve network-wide objectives. However, individual nodes may behave selfishly in order to maximize their own benefits without considering the global benefits of the network. One approach to incentivize nodes cooperation for better global benefits is to establish trust relations among nodes to guide their decision making. In this paper, we present a game theoretic analysis for the efficiency of establishing trust for improving node cooperation. The trust relations among nodes are modeled as a trust-weighted network, and we study a graphical game in this network where the nodes' payoffs are affected by their trust relations. We characterize the Nash equilibrium and the social optimum of this game and show that the game efficiency has a close relationship to the Bonacich centralities of nodes in the trust-weighted network. Furthermore, we propose to improve game efficiency by introducing heterogeneous resources to nodes according to their centralities. We provide both experimental and theoretical analysis on the improvement of the game efficiency.

The Internet offers tremendous opportunities for information sharing and content distribution. However, without proper filtering, the large amount of information may likely swarm the users rather than benefit them. Collaborative filtering is a technique for extracting useful information from the large information pool generated by interconnected online communities. In this paper, we develop a probabilistic collaborative filtering algorithm, which is based on ordered logistic regression and takes into account both similarities among the users and similarities among the items. We make inference with maximum likelihood and Bayesian frameworks, and propose a Markov Chain Monte Carlo based Expectation Maximization algorithm to optimize model parameters. The power of our proposed algorithm is its extensibility. We show that it can incorporate content and contextual information. More importantly, it can be easily extended to include the trustworthiness of users, thus being more robust to malicious data manipulation. The experimental results on a real world data set show that our proposed algorithm with the trust extension is robust under different types of attacks in recommendation systems.

Fast anomaly detection and localization is critical to ensure effective functioning of wireless sensor networks. The low bandwidth and power constraints in wireless sensor networks are the main challenges for achieving this task, especially for large scale networks. In this paper, we propose a trust-assisted framework for detecting and localizing network anomalies in a hierarchical sensor network. The proposed method makes inference based on end-to-end measurements collected by a set of measurement nodes. Network heterogeneity is exploited for better bandwidth and energy efficiency. The trustworthiness of network links is utilized to design an efficient two-phase probing strategy that can achieve a flexible tradeoff between inference accuracy and probing overhead. We performed experiments with different network settings and demonstrated the effectiveness of our proposed algorithms.

Distributed sensor networks have been widely employed to monitor and protect critical infrastructure assets. The network status can be estimated by centralized state estimation using coordinated data aggregation or by distributed state estimation, where nodes only exchange information locally to achieve enhanced scalability and adaptivity to network dynamics. One important property of state estimation is robustness against false data injection from sensors compromised by attackers. Different from most existing works in the literature that focus on centralized state estimation, we propose two novel robust distributed state estimation algorithms against false data injection. They are built upon an existing distributed Kalman filtering algorithm. In the first algorithm, we use variational Bayesian learning to estimate attack parameters and achieve performance similar to a centralized majority voting rule, without causing extra communication overhead. In the second algorithm, we introduce heterogeneity into the network by utilizing a subset of pre-trusted nodes to achieve performance better than majority voting. We show that as long as there is a path connecting each node to some of the pre-trusted nodes, the attackers can not subvert the network. Experimental results demonstrate the effectiveness of our proposed schemes.

In order to meet the constantly increasing demand for ubiquitous, mobile access to the internet, next-generation mobile broadband communications systems based on OFDMA, such as IEEE 802.16 m, require a significant performance increase over previous generation systems, such as IEEE 802.16e-2005, particularly in cell-edge and average spectral efficiency. In this paper, we address the downlink adaptive frequency reuse (AFR) and multi-user MIMO (MU-MIMO) techniques which are considered to be the most promising candidates for meeting the requirements on cell-edge and average spectral efficiency of next-generation mobile broadband systems.