In this work, we study algorithms for cooperative search and survey using a fleet of Autonomous Underwater Vehicles (AUVs). Due to the limited energy, communication range/bandwidth, and sensing range of the AUVs, underwater search and survey with multiple AUVs brings about several new challenges since a large amount of data needs to be collected by each AUV, and any AUV may fail unexpectedly. To address the challenges and meet our objectives of minimizing the total survey time and traveled distance of AUVs, we propose a cooperative rendezvous scheme called Synchronization-Based Survey (SBS) to facilitate cooperation among a large number of AUVs when surveying a large area. In SBS, AUVs form an intermittently connected network (ICN) in that they periodically meet each other for data aggregation, control signal dissemination, and AUV failure detection/recovery. Numerical analysis and simulations have been performed to compare the performance of three variants of SBS schemes, namely, Alternating Column Synchronization (ACS), Strict Line Synchronization (SLS), and X Synchronization (XS). The results show that XS can outperform other SBS schemes in terms of the survey time and the traveled distance of AUVs. We also compare XS with nonsynchronization-based survey and the lower bound on the survey time and traveled distance. The results show that XS achieves a close to optimal performance.

Deploying a multi-hop underwater acoustic sensor network (UASN) in a large area brings about new challenges in reliable data transmissions and survivability of network due to the limited underwater communication range/bandwidth and the limited energy of underwater sensor nodes. In order to address those challenges and achieve the objectives of maximization of data delivery ratio and minimization of energy consumption of underwater sensor nodes, this paper proposes a new underwater routing scheme, namely AURP (AUV-aided underwater routing protocol), which uses not only heterogeneous acoustic communication channels but also controlled mobility of multiple autonomous underwater vehicles (AUVs). In AURP, the total data transmissions are minimized by using AUVs as relay nodes, which collect sensed data from gateway nodes and then forward to the sink. Moreover, controlled mobility of AUVs makes it possible to apply a short-range high data rate underwater channel for transmissions of a large amount of data. To the best to our knowledge, this work is the first attempt to employ multiple AUVs as relay nodes in a multi-hop UASN to improve the network performance in terms of data delivery ratio and energy consumption. Simulations, which are incorporated with a realistic underwater acoustic communication channel model, are carried out to evaluate the performance of the proposed scheme, and the results indicate that a high delivery ratio and low energy consumption can be achieved.

Stock price prediction has always been an important application in time series predictions. Recently, deep neural networks have been employed extensively for financial time series tasks. The network typically requires a large amount of training samples to achieve high accuracy. However, in the stock market, the number of data points collected on a daily basis is limited in one year, which leads to insufficient training samples and accordingly results in an overfitting problem. Moreover, predicting stock price movement is affected by various factors in the stock market. Therefore, choosing appropriate input features for prediction models should be taken into account. To address these problems, this paper proposes a novel framework, named deep transfer with related stock information (DTRSI), which takes advantage of a deep neural network and transfer learning. First, a base model using long short-term memory (LSTM) cells is pre-trained based on a large amount of data, which are obtained from a number of different stocks, to optimize initial training parameters. Second, the base model is fine-tuned by using a small amount data from a target stock and different types of input features (constructed based on the relationship between stocks) in order to enhance performance. Experiments are conducted with data from top-five companies in the Korean market and the United States (US) market from 2012 to 2018 in terms of the highest market capitalization. Experimental results demonstrate the effectiveness of transfer learning and using stock relationship information in helping to improve model performance, and the proposed approach shows remarkable performance (compared to other baselines) in terms of prediction accuracy.

Given a probabilistic graph G representing an intermittently connected network and routing algorithm A, we wish to determine a delivery subgraph G[A] of G with at most k edges, such that the probability Conn₂ (G[A]) that there is a path from source s to destination t (in a graph H chosen randomly from the probability space defined by G[A]) is maximized. To the best of our knowledge, this problem and its complexity has not been addressed in the literature. Also, there is the corresponding distributed version of the problem where the delivery subgraph G[A] is to be constructed distributively, yielding a routing protocol. Our proposed solution to this routing problem is multi-fold: First, we prove the hardness of our optimization problem of finding a delivery subgraph that maximizes the delivery probability and discuss the hardness of computing the objective function Conn₂(G[A]); Second, we present an algorithm to approximate Conn₂(G[A]) and compare it with an optimal algorithm; Third, we focus on intermittently connected networks, and model the users' mobility within them; and Fourth, we propose an edge-constrained routing protocol (EC-SOLAR-KSP) based on the insights obtained from the first step and the contact probabilities computed in the third step. We then highlight the protocol's novelty and effectiveness by comparing it with a probabilistic routing protocol, and an epidemic routing protocol proposed in literature.

Facial expression recognition (FER) is a challenging problem in the fields of pattern recognition and computer vision. The recent success of convolutional neural networks (CNNs) in object detection and object segmentation tasks has shown promise in building an automatic deep CNN-based FER model. However, in real-world scenarios, performance degrades dramatically owing to the great diversity of factors unrelated to facial expressions, and due to a lack of training data and an intrinsic imbalance in the existing facial emotion datasets. To tackle these problems, this paper not only applies deep transfer learning techniques, but also proposes a novel loss function called weighted-cluster loss, which is used during the fine-tuning phase. Specifically, the weighted-cluster loss function simultaneously improves the intra-class compactness and the inter-class separability by learning a class center for each emotion class. It also takes the imbalance in a facial expression dataset into account by giving each emotion class a weight based on its proportion of the total number of images. In addition, a recent, successful deep CNN architecture, pre-trained in the task of face identification with the VGGFace2 database from the Visual Geometry Group at Oxford University, is employed and fine-tuned using the proposed loss function to recognize eight basic facial emotions from the AffectNet database of facial expression, valence, and arousal computing in the wild. Experiments on an AffectNet real-world facial dataset demonstrate that our method outperforms the baseline CNN models that use either weighted-softmax loss or center loss.

In emergency response operations, using uncrewed aerial vehicles (UAVs) has recently become a promising solution due to their flexibility and easy deployment. However, tasks performed by the UAVs, e.g., object detection and human pose recognition, usually require a high computation capacity and energy supply. Furthermore, offloading tasks to the edge server-equipped base stations may not always be possible because of a lack of infrastructure or distance. Therefore, UAV-aided edge servers can be deployed near UAV scouts to provide computing services. However, a UAV can not perform all types of tasks since it has limitations on memory, available software, central processing unit (CPU), and graphics processing unit (GPU) capacity. Therefore, this study focuses on task offloading (TO), power, and computation resource allocation (PRA) problems in a multi-layer MEC-enabled UAV network while taking into account CPU and GPU requirements of tasks, the capacity of the devices (i.e., computational resources, power, and energy), and limitations on the type of tasks a UAV can perform. The problem is formulated as a non-convex mixed-integer nonlinear problem to minimize the weighted sum of the maximum energy consumption ratio in the network and total task execution latency ratio, and then decomposed and converted into an integer and a convex problem. A messy genetic algorithm (mGA)-based TO and PRA strategy (mGA-TPR) is proposed to solve the problem, where two PRA strategies are based on the Karush–Kuhn–Tucker conditions used to solve the PRA problem. Simulation results verify that the proposed scheme can outperform the baseline methods.