Evaporation ducts are a particular type of atmospheric stratification that frequently appears on the sea surface. The accurate and timely prediction of the evaporation duct height (EDH) is significant for the practical application of electromagnetic communication equipment. Due to the typical time-series (TS) characteristics of the measured evaporation duct data, we construct an EDH prediction model based on a long short-term memory network (LSTM-EDH model) for the first time. The experimental results show that the LSTM-EDH model’s root-mean-square error (RMSE) is dramatically reduced and can achieve a better fit of the measured EDH compared with the Babin–Young–Carton (BYC), Naval-Postgraduate-School (NPS), and eXtreme Gradient Boosting (XGB) EDH models. Compared with the XGB model, the generalization ability is also greatly improved.

Sea surface temperature (SST) prediction has attracted increasing attention, due to its crucial role in understanding the Earth’s climate and ocean system. Existing SST prediction methods are typically based on either physics-based numerical methods or data-driven methods. Physics-based numerical methods rely on marine physics equations and have stable and explicable outputs, while data-driven methods are flexible in adapting to data and are capable of detecting unexpected patterns. We believe that these two types of method are complementary to each other, and their combination can potentially achieve better performances. In this paper, a space-time partial differential equation (PDE) is employed to form a novel physics-based deep learning framework, named the space-time PDE-guided neural network (STPDE-Net), to predict daily SST. Comprehensive experiments for SST prediction were conducted, and the results proved that our method could outperform the traditional finite-difference forecast method and several state-of-the-art deep learning and physics-guided deep learning methods.

Nowadays, people are overwhelmingly exposed to various kinds of information from different information networks. In order to recommend users with the information entities that match their interests, many recommendation methods have been proposed so far. And some of these methods have explored different ways to utilize different kinds of auxiliary information to deal with the information sparsity problem of user feedbacks. However, as a special kind of information sparsity problem, the “cold start” problem is still a big challenge not well-solved yet in the recommendation problem. In order to tackle the “cold start” challenge, in this paper, we propose a novel recommendation model, which integrates the auxiliary information in multiple heterogeneous information networks (HINs), namely the Cross-HIN Recommendation System (CHRS) . By utilizing the rich heterogeneous information from meta-paths, the CHRS is able to calculate the similarities of information entities and apply the calculated similarity scores in the recommendation process. For the information entities shared among multiple information networks, CHRS transfers item latent information from other networks to help the recommendation task in a given network. During the information transfer process, CHRS applies a domain adaptation matrix to tackle the domain difference problem. We conduct experiments to compare our CHRS method with several widely employed or the state-of-art recommendation models, and the experimental results demonstrate that our method outperforms the baseline methods in addressing the “cold start” recommendation problem.

Inferring the links among entities in networks is an important research problem for various disciplines. Depending on the specific application settings, the links to be inferred are usually subject to different cardinality constraints, like one-to-one, one-to-many and many-to-many. However, most existing research works on link prediction problems fail to consider such a kind of constraint. In this paper, we propose to study the link prediction problem with general cardinality constraints, which is formally defined as the CLP (Cardinality Constrained Link Prediction) problem. By minimizing the projection loss of links from feature vectors to labels, the CLP problem is formulated as an optimization problem involving multiple variables, where the cardinality constraints are modeled as mathematical constraints on node degrees. The objective function is shown to be not jointly convex and the optimal solution subject to the cardinality constraints can be very time-consuming to achieve. To solve the optimization problem, an iterative variable updating based link prediction framework ITERCLIPS (Iterative Constrained Link Prediction & Selection) is introduced in this paper, which involves the steps on link updating and selection alternatively. To overcome the high time cost problem, a greedy link selection step is introduced in this paper, which picks links greedily while preserving the link cardinality constraints simultaneously. Meanwhile, to ensure the effectiveness of ITERCLIPS on large-scale networks, a distributed implementation of ITERCLIPS is further presented as a scalable solution to the CLP problem. Extensive experiments have been done on three real-world network datasets with different types of cardinality constraints, and the experimental results achieved by ITERCLIPS on all these datasets can demonstrate the effectiveness and advantages of ITERCLIPS in solving the CLP problem.