Abstract: Deep neural networks can now generate real-world data. To overcome the posterior distribution’s intractability, implicit probabilistic models that generate data stochastically have been introduced. However, data modeling requires deep statistical dependence knowledge, which can…
Abstract: Many node embedding and network representation methods use random walks. However, using traditional random walks results in embeddings that capture proximity (communities) rather than structural similarity (roles). Due to node identity, embeddings cannot be…
Abstract: The lifelong machine learning paradigm learns a sequence of tasks based on previous experiences, such as knowledge library or deep network weights. Most recent lifelong learning models have knowledge libraries or deep networks of…
Abstract: Event-based Social Network (EBSN) connects online and offline social relationships. EBSNs publish more events, making personalized event recommendation essential to help users choose appealing events. Most event recommendation algorithms fail to distinguish constraint factors…
Abstract: Momentum technique has improved deep learning and regularized learning by accelerating gradient descent (GD) convergence. NAG and HB are momentum methods. Most acceleration analyses have focused on NAG, with a few on HB. This…
Abstract: Ensemble learning, which combines multiple models, has performed well in many tasks. Unsupervised ensemble classification is the focus here. Unsupervised means the ensemble combiner does not know the ground-truth labels each classifier was trained…
Abstract: Security applications employ machine learning models. However, attackers can adapt to avoid detection. Machine learning models have been strengthened against such attacks. However, randomization is underutilized. Most studies use models with differentiable error functions,…
Abstract: This paper introduces an unsupervised feature learning architecture based on a multi-clustering integration module and MIRBM, a variant of RBM. The multi-clustering integration module uses three clusterers—K-means, affinity propagation, and spectral—to generate three clustering…
Abstract: Unsupervised domain adaptation is effective in leveraging rich information from a labeled source domain to an unlabeled target domain. Though deep learning and adversarial strategy made a significant breakthrough in the adaptability of features,…
Abstract: In this paper, we study an automatic hypothesis generation (HG) problem, which involves finding meaningful implicit connections between scientific terms like diseases, chemicals, drugs, and genes from biomedical publication databases. Most previous studies used…
Abstract: To serve a user, an intelligent robot needs episodic memory to recall a sequence of events from past experiences. ART networks have been used to design episodic memories that incrementally learn new tasks without…
Abstract: Detecting small low-contrast airspace targets is crucial and difficult. This article proposes a simple and effective data-driven support vector machine (SVM)-based spatiotemporal feature fusion detection method for small low-contrast targets. A spatiotemporal profile, a…
WhatsApp us