Search

Topic models are a prevailing tool for exploring latent topics in documents, and for helping to complete many NLP tasks. To obtain good topics for a corpus, a preprocessing step is often needed to remove common stop words and identify topic general words (TGW) from the corpus. Such words can seriously harm the topic formation because they create spurious co-occurrence of unrelated words. Also, they are likely to occupy top positions of multiple topics, lead to many unrelated words being grouped under a topic, and consequently result in inscrutable and similar topics. In an application, one typically manually identifies and removes a list of TGWs in the corpus. This is a time consuming process and very hard to do by a layman user. In this paper, we aim to solve this problem automatically. The proposed approaches can be based on the current corpus alone or multiple corpora. In the latter case, a novel continuous learning method is proposed that learns from past results of multiple domain corpora to help identify TGWs in the current domain. We conduct experiments in two real-world datasets, and the experimental results show that the proposed approaches achieve superior results. [ABSTRACT FROM AUTHOR]

In traditional federated learning such as FedAvg, the associations among clients are often ignored when executing on non-independent or heterogeneously distributed datasets, resulting in unsatisfactory accuracy. Although some previous works on clustered federated learning have been proposed to address such problems, most of them have a polarized problem. When the number of clustering is small, the model performs poorly and fails to accurately capture the distinction between clients. While a large number of clustering times tends to lead to higher communication costs. Therefore, a critical need is to design an efficient clustered federated solution that can both better capture the diversity between local clients and minimize the communication and computation costs. To this end, we propose an efficient one-off clustered federated learning framework called FedEOC. FedEOC exploits the "learning-to-learn" characteristic of meta-learning to enhance the generalization of the model across different clients so that only a small number of iterations are needed for each client to quickly obtain locally adapted weights. Based on the well-initially trained weights on all clients, we can cluster the clients only once to achieve the effect of one-off clustering and multiple-round applying. Additionally, to alleviate the issue of cluster imbalance, FedEOC is equipped with a Decomposition and Consolidation (Dec-Con) mechanism to decompose the clients from the extreme clusters and consolidate them into the most similar ones. The comprehensive experiments conducted on two real-world datasets demonstrate the superior capability of FedEOC from both aspects of accuracy and efficiency. [ABSTRACT FROM AUTHOR]

Traditional topic modeling has been widely studied and popularly employed in expert systems and information systems. However, traditional topic models cannot discover structural relations among topics, thus losing the chance to explore the data more deeply. Hierarchical topic modeling has the capability of learning topics, as well as discovering the hierarchical topic structure from text data. But purely unsupervised models tend to generate weak topic hierarchies. To solve this problem, we propose a novel knowledge-based hierarchical topic model (KHTM), which can incorporate prior knowledge into topic hierarchy building. A key novelty of this model is that it can mine prior knowledge automatically from the topic hierarchies of multiple domains corpora. In this paper, the knowledge is represented as the word pairs which satisfy the requirement of frequent co-occurrence, and knowledge is organized in form of hierarchical structure. We also propose an iterative learning algorithm. For evaluation, we crawled two new multi-domain datasets and conducted comprehensive experiments. The experimental results show that our algorithm and model can generate more coherent topics, and more reasonable hierarchical structure. [ABSTRACT FROM AUTHOR]

YouTube offers a great opportunity for people to entertain, advertise, gain popularity, and generate revenue. How to increase views for a video has become the key question for anyone who wish to be famous or gain more revenue. Recognizing that a recommendation system is a major view source for videos, our goal in this paper is to increase video views in YouTube by leveraging on the recommendation system. We first perform measurements to understand how views are propagated among videos through recommendation links and identify factors that influence the recommendation produced by the system. Our measurement results show that similarity in video meta-data is a crucial ingredient in connecting videos. We then propose a keyword suggestion method for a video with the aim to raise video views through the recommendation system. The keyword suggestion method utilizes video clusters on a referrer video graph to obtain relevant keywords and ranks keywords based on both their relevance and their potential to attract video views. The effectiveness of the keyword suggestion method is demonstrated through a case study, showing that using the keywords suggested by our method leads to a larger number of video views and higher average watching time per video playback compared to initial user-given keywords. [ABSTRACT FROM AUTHOR]

With the explosive growth of products and multimedia contents on the Internet, the desire of users to find these online resources matching their interests makes it imperative to develop high quality recommendation systems. In this paper, we propose one novel neural network, namely Dynamic Heterogeneous Graph Convolutional Network (DHGCN), for item recommendation. Specifically, our proposed DHGCN consists of two components, namely the graph learner and heterogeneous graph convolution. The graph learner considers not only the user–item interactions but also user–user and item–item interactions, which are dynamically established during the graph evolution process. The heterogeneous graph convolution relies on a novel cross gating strategy to aggregate the representations yielded by the convolution over the learned heterogeneous graph and the item content information. Through comprehensive experiments on two real-world datasets, the proposed model is demonstrated to be effective on item recommendation task, outperforming the existing state-of-the-art models. [ABSTRACT FROM AUTHOR]