Self-clustered GAN for precipitation nowcasting.

This paper proposes a novel GAN framework with self-clustering approach for precipitation nowcasting (ClusterCast). Previous studies have primarily captured the motion vector using only a single latent space, making the models difficult to adapt to disparate space-time distribution of precipitation. Environmental factors (e.g., regional characteristics and precipitation scale) have an impact on precipitation systems and can cause non-stationary distribution. To tackle this problem, our key idea is to train a generator network to predict future radar frames by learning a sub-network that automatically labels precipitation types from a generative model. The training process consists of (i) clustering the hierarchical features derived from the generator stem using a sub-network and (ii) predicting future radar frames according to the self-supervised labels, enabling heterogeneous latent representation. Additionally, we attempt an ensemble forecast that prescribes random perturbations to improve performance. With the flexibility of representation learning, ClusterCast enables the model to learn precipitation distribution more accurately. Results indicate that our method generates non-blurry future frames by preventing mode collapse, and the proposed method demonstrates robustness across various precipitation scenarios. Extensive experiments demonstrate that our method outperforms four benchmarks on a 2-h prediction basis with a mean squared error (MSE) of 8.9% on unseen datasets.

DaGzang: a synthetic data generator for cross-domain recommendation services.

Research on cross-domain recommendation systems (CDRS) has shown efficiency by leveraging the overlapping associations between domains in order to generate more encompassing user models and better recommendations. Nonetheless, if there is no dataset belonging to a specific domain, it is a challenge to generate recommendations in CDRS. In addition, finding these overlapping associations in the real world is generally tricky, and it makes its application to actual services hard. Considering these issues, this study aims to present a synthetic data generation platform (called DaGzang) for cross-domain recommendation systems. The DaGzang platform works according to the complete loop, and it consists of the following three steps: (i) detecting the overlap association (data distribution pattern) between the real-world datasets, (ii) generating synthetic datasets based on these overlap associations, and (iii) evaluating the quality of the generated synthetic datasets. The real-world datasets in our experiments were collected from Amazon's e-commercial website. To validate the usefulness of the synthetic datasets generated from DaGzang, we embed these datasets into our cross-domain recommender system, called DakGalBi. We then evaluate the recommendations generated from DakGalBi with collaborative filtering (CF) algorithms, user-based CF, and item-based CF. Mean absolute error (MAE) and root mean square error (RMSE) metrics are measured to evaluate the performance of collaborative filtering (CF) CDRS. In particular, the highest performance of the three recommendation methods is user-based CF when using 10 synthetic datasets generated from DaGzang (0.437 at MAE and 0.465 at RMSE).

SocioPedia+: a visual analytics system for social knowledge graph-based event exploration.

In the recent era of information explosion, exploring event from social networks has recently been a crucial task for many applications. To derive valuable comprehensive and thorough insights on social events, visual analytics (VA) system have been broadly used as a promising solution. However, due to the enormous social data volume with highly diversity and complexity, the number of event exploration tasks which can be enabled in a conventional real-time visual analytics systems has been limited. In this article, we introduce SocioPedia+, a real-time visual analytics system for social event exploration in time and space domains. By introducing the dimension of social knowledge graph analysis into the system multivariate analysis, the process of event explorations in SocioPedia+ can be significantly enhanced and thus enabling system capability on performing full required tasks of visual analytics and social event explorations. Furthermore, SocioPedia+ has been optimized for visualizing event analysis on different levels from macroscopic (events level) to microscopic (knowledge level). The system is then implemented and investigated with a detailed case study for evaluating its usefulness and visualization effectiveness for the application of event explorations.

Understanding asymmetric synergistic effect between movie actors.

Synergistic effects between movie actors have been regarded as an important indicator when they are casted in a new movie. People simply assume that synergistic effect is symmetric. The aim of this study is to understand the asymmetric synergy between actors. We propose an asymmetric synergy measurement method for actor's star-power-based costarring movies to understand the synergistic effect. When measuring the synergy, we designed it so that it would be possible to measure the synergy that varies with the time of the costarring movie's release and the synergy of new actors. Measured synergies were analyzed on an actor's synergy and asymmetric synergy between actors to examine the characteristics of highly synergistic actors and the asymmetric synergy between actors. Moreover, we confirmed that measuring synergies asymmetrically demonstrated better prediction performance in various evaluation metrics (accuracy, precision, recall, and F1-score) than measuring synergies symmetrically through the synergy prediction experiment using synergy and asymmetric synergy.

DNN-based multi-output model for predicting soccer team tactics.

In modern sports, strategy and tactics are important in determining the game outcome. However, many coaches still base their game tactics on experience and intuition. The aim of this study is to predict tactics such as formations, game styles, and game outcome based on soccer dataset. In this paper, we propose to use Deep Neural Networks (DNN) based on Multi-Layer Perceptron (MLP) and feature engineering to predict the soccer tactics of teams. Previous works adopt simple machine learning techniques, such as Support Vector Machine (SVM) and decision tree, to analyze soccer dataset. However, these often have limitations in predicting tactics using soccer dataset. In this study, we use feature selection, clustering techniques for the segmented positions and Multi-Output model for Soccer (MOS) based on DNN, wide inputs and residual connections. Feature selection selects important features among features of soccer player dataset. Each position is segmented by applying clustering to the selected features. The segmented positions and game appearance dataset are used as training dataset for the proposed model. Our model predicts the core of soccer tactics: formation, game style and game outcome. And, we use wide inputs and embedding layers to learn sparse, specific rules of soccer dataset, and use residual connections to learn additional information. MLP layers help the model to generalize features of soccer dataset. Experimental results demonstrate the superiority of the proposed model, which obtain significant improvements comparing to baseline models.

Seizure detection from multi-channel EEG using entropy-based dynamic graph embedding.

An epileptic seizure is a chronic disease with sudden abnormal discharge of brain neurons, which leads to transient brain dysfunction. To detect epileptic seizures, we propose a novel idea based on a dynamic graph embedding model. The dynamic graph is built by identifying the correlation among the multi-channel EEG signals. Graph entropy measurement is exploited to calculate the similarity among the graph at each time interval and construct the graph embedding space. Since the abnormal electrical brain activity causes the epileptic seizure, the graph entropy during the seizure time interval is different from other time intervals. Therefore, we propose an entropy-based dynamic graph embedding model to cluster the graphs, and the graphs with epileptic seizures are discriminated. We applied the proposed approach to the Children Hospital Boston-Massachusetts Institute of Technology Scalp EEG database. The results have shown that the proposed approach outperformed the baselines by 1.4% with respect to accuracy.

Entropy-based dynamic graph embedding for anomaly detection on multiple climate time series.

Abnormal climate event is that some meteorological conditions are extreme in a certain time interval. The existing methods for detecting abnormal climate events utilize supervised learning models to learn the abnormal patterns, but they cannot detect the untrained patterns. To overcome this problem, we construct a dynamic graph by discovering the correlation among the climate time series and propose a novel dynamic graph embedding model based on graph entropy called EDynGE to discriminate anomalies. The graph entropy measurement quantifies the information of the graphs and constructs the embedding space. We conducted experiments on synthetic datasets and real-world meteorological datasets. The results showed that EdynGE model achieved a better F1-score than the baselines by 43.2%, and the number of days of abnormal climate events has increased by 304.5 days in the past 30 years.

Dynamic graph embedding for outlier detection on multiple meteorological time series.

Existing dynamic graph embedding-based outlier detection methods mainly focus on the evolution of graphs and ignore the similarities among them. To overcome this limitation for the effective detection of abnormal climatic events from meteorological time series, we proposed a dynamic graph embedding model based on graph proximity, called DynGPE. Climatic events are represented as a graph where each vertex indicates meteorological data and each edge indicates a spurious relationship between two meteorological time series that are not causally related. The graph proximity is described as the distance between two graphs. DynGPE can cluster similar climatic events in the embedding space. Abnormal climatic events are distant from most of the other events and can be detected using outlier detection methods. We conducted experiments by applying three outlier detection methods (i.e., isolation forest, local outlier factor, and box plot) to real meteorological data. The results showed that DynGPE achieves better results than the baseline by 44.3% on average in terms of the F-measure. Isolation forest provides the best performance and stability. It achieved higher results than the local outlier factor and box plot methods, namely, by 15.4% and 78.9% on average, respectively.

Implicit Stochastic Gradient Descent Method for Cross-Domain Recommendation System.

The previous recommendation system applied the matrix factorization collaborative filtering (MFCF) technique to only single domains. Due to data sparsity, this approach has a limitation in overcoming the cold-start problem. Thus, in this study, we focus on discovering latent features from domains to understand the relationships between domains (called domain coherence). This approach uses potential knowledge of the source domain to improve the quality of the target domain recommendation. In this paper, we consider applying MFCF to multiple domains. Mainly, by adopting the implicit stochastic gradient descent algorithm to optimize the objective function for prediction, multiple matrices from different domains are consolidated inside the cross-domain recommendation system (CDRS). Additionally, we design a conceptual framework for CDRS, which applies to different industrial scenarios for recommenders across domains. Moreover, an experiment is devised to validate the proposed method. By using a real-world dataset gathered from Amazon Food and MovieLens, experimental results show that the proposed method improves 15.2% and 19.7% in terms of computation time and MSE over other methods on a utility matrix. Notably, a much lower convergence value of the loss function has been obtained from the experiment. Furthermore, a critical analysis of the obtained results shows that there is a dynamic balance between prediction accuracy and computational complexity.

Learning Hierarchical Representations of Stories by Using Multi-layered Structures in Narrative Multimedia.

Narrative works (e.g., novels and movies) consist of various utterances (e.g., scenes and episodes) with multi-layered structures. However, the existing studies aimed to embed only stories in a narrative work. By covering other granularity levels, we can easily compare narrative utterances that are coarser (e.g., movie series) or finer (e.g., scenes) than a narrative work. We apply the multi-layered structures on learning hierarchical representations of the narrative utterances. To represent coarser utterances, we consider adjacency and appearance of finer utterances in the coarser ones. For the movies, we suppose a four-layered structure (character roles ∈ characters ∈ scenes ∈ movies) and propose three learning methods bridging the layers: Char2Vec, Scene2Vec, and Hierarchical Story2Vec. Char2Vec represents a character by using dynamic changes in the character's roles. To find the character roles, we use substructures of character networks (i.e., dynamic social networks of characters). A scene describes an event. Interactions between characters in the scene are designed to describe the event. Scene2Vec learns representations of a scene from interactions between characters in the scene. A story is a series of events. Meanings of the story are affected by order of the events as well as their content. Hierarchical Story2Vec uses sequential order of scenes to represent stories. The proposed model has been evaluated by estimating the similarity between narrative utterances in real movies.

Is Performance of Scholars Correlated to Their Research Collaboration Patterns?

This study aims to validate whether the research performance of scholars correlates with how the scholars work together. Although the most straightforward approaches are centrality measurements or community detection, scholars mostly participate in multiple research groups and have different roles in each group. Thus, we concentrate on the subgraphs of co-authorship networks rooted in each scholar that cover (i) overlapping of the research groups on the scholar and (ii) roles of the scholar in the groups. This study calls the subgraphs "collaboration patterns" and applies subgraph embedding methods to discover and represent the collaboration patterns. Based on embedding the collaboration patterns, we have clustered scholars according to their collaboration styles. Then, we have examined whether scholars in each cluster have similar research performance, using the quantitative indicators. The coherence of the indicators cannot be solid proofs for validating the correlation between collaboration and performance. Nevertheless, the examination for clusters has exhibited that the collaboration patterns can reflect research styles of scholars. This information will enable us to predict the research performance more accurately since the research styles are more consistent and sustainable features of scholars than a few high-impact publications.

Detection of cognitive impairment using a machine-learning algorithm.

PURPOSE: The Mini-Mental State Examination (MMSE) is one of the most frequently used bedside screening measures of cognition. However, the Korean Dementia Screening Questionnaire (KDSQ) is an easier and more reliable screening method. Instead, other clinical variables and raw data were used for this study without the consideration of a cutoff value. The objective of this study was to develop a machine-learning algorithm for the detection of cognitive impairment (CI) based on the KDSQ and the MMSE. PATIENTS AND METHODS: The original dataset from the Clinical Research Center for Dementia of South Korea study was obtained. In total, 9,885 and 300 patients were randomly allocated to the training and test datasets, respectively. We selected up to 24 variables including sex, age, education duration, diabetes mellitus, and hypertension. We trained a machine-learning algorithm using TensorFlow based on the training dataset and then calculated its accuracy using the test dataset. The cost was calculated by conducting a logistic regression. RESULTS: The accuracy of the model in predicting CI based on the KDSQ only, the MMSE only, and the combination of the KDSQ and MMSE was 84.3%, 88.3%, and 86.3%, respectively. For the KDSQ, the sensitivity for detecting CI was 91.50% and the specificity for detecting normal cognition (NL) was 59.60%. The sensitivity of the MMSE was 94.35%, and the specificity was 59.62%. When combining the KDSQ and the MMSE, the sensitivity for detecting CI was 91.5% and the specificity for detecting NL was 61.5%. CONCLUSION: The algorithm predicting CI based on the MMSE is superior. However, the KDSQ can be administered more easily in clinical practice and the algorithm using KDSQ is a comparable screening tool.

Consensual Negotiation-Based Decision Making for Connected Appliances in Smart Home Management Systems.

Recently, the concept of Internet of Agent has been introduced as a potential technology that pushes intelligence, data processing, analytics and communication capabilities down to the point where the data originates. In this paper, we introduce a novel approach for a Decentralized Home Energy Management System by applying the Internet of Agent concept. In particular, we first present an Internet of Agent framework in terms of sensing, communicating and collaborating among connected appliances. Then, the decentralized management based on consensual negotiation mechanism with several intelligent techniques are proposed for dynamic scheduling connected appliance. Specifically, by applying the Internet of Agent framework, connected appliances are regarded as smart agents that are able to make individual decisions by reaching agreement over the exchange of operations on competitive resources. Furthermore, in this study, the load balancing problem in which load shifting is able to reduce the electricity demand during peak hours is taken into account in order to emphasize the effectiveness of our approach. For the experiment, we develop a simulation of smart home environment to evaluate our approach using NetLogo, a tool which provides real-time analysis in the modeling and simulation domain of complex systems.

Social big data: Recent achievements and new challenges.

Big data has become an important issue for a large number of research areas such as data mining, machine learning, computational intelligence, information fusion, the semantic Web, and social networks. The rise of different big data frameworks such as Apache Hadoop and, more recently, Spark, for massive data processing based on the MapReduce paradigm has allowed for the efficient utilisation of data mining methods and machine learning algorithms in different domains. A number of libraries such as Mahout and SparkMLib have been designed to develop new efficient applications based on machine learning algorithms. The combination of big data technologies and traditional machine learning algorithms has generated new and interesting challenges in other areas as social media and social networks. These new challenges are focused mainly on problems such as data processing, data storage, data representation, and how data can be used for pattern mining, analysing user behaviours, and visualizing and tracking data, among others. In this paper, we present a revision of the new methodologies that is designed to allow for efficient data mining and information fusion from social media and of the new applications and frameworks that are currently appearing under the "umbrella" of the social networks, social media and big data paradigms.

Intelligent advisory speed limit dedication in highway using VANET.

Variable speed limits (VSLs) as a mean for enhancing road traffic safety are studied for decades to modify the speed limit based on the prevailing road circumstances. In this study the pros and cons of VSL systems and their effects on traffic controlling efficiency are summarized. Despite the potential effectiveness of utilizing VSLs, we have witnessed that the effectiveness of this system is impacted by factors such as VSL control strategy used and the level of driver compliance. Hence, the proposed approach called Intelligent Advisory Speed Limit Dedication (IASLD) as the novel VSL control strategy which considers the driver compliance aims to improve the traffic flow and occupancy of vehicles in addition to amelioration of vehicle's travel times. The IASLD provides the advisory speed limit for each vehicle exclusively based on the vehicle's characteristics including the vehicle type, size, and safety capabilities as well as traffic and weather conditions. The proposed approach takes advantage of vehicular ad hoc network (VANET) to accelerate its performance, in the way that simulation results demonstrate the reduction of incident detection time up to 31.2% in comparison with traditional VSL strategy. The simulation results similarly indicate the improvement of traffic flow efficiency, occupancy, and travel time in different conditions.

Hunk negatively regulates c-myc to promote Akt-mediated cell survival and mammary tumorigenesis induced by loss of Pten.

The protooncogenes Akt and c-myc each positively regulate cell growth and proliferation, but have opposing effects on cell survival. These oncogenes cooperate to promote tumorigenesis, in part because the prosurvival effects of Akt offset the proapoptotic effects of c-myc. Akt's ability to counterbalance c-myc's proapoptotic effects has primarily been attributed to Akt-induced stimulation of prosurvival pathways that indirectly antagonize the effects of c-myc. We report a more direct mechanism by which Akt modulates the proapoptotic effects of c-myc. Specifically, we demonstrate that Akt up-regulates the adenosine monophosphate-associated kinase (AMPK)-related protein kinase, Hormonally up-regulated neu-associated kinase (Hunk), which serves as an effector of Akt prosurvival signaling by suppressing c-myc expression in a kinase-dependent manner to levels that are compatible with cell survival. Consequently, Akt pathway activation in the mammary glands of Hunk(-/-) mice results in induction of c-myc expression to levels that induce apoptosis. c-myc knockdown rescues the increase in apoptosis induced by Hunk deletion in cells in which Akt has been activated, indicating that repression of c-myc is a principal mechanism by which Hunk mediates the prosurvival effects of Akt. Consistent with this mechanism of action, we find that Hunk is required for c-myc suppression and mammary tumorigenesis induced by phosphatase and tensin homolog (Pten) deletion in mice. Together, our findings establish a prosurvival function for Hunk in tumorigenesis, define an essential mechanism by which Akt suppresses c-myc-induced apoptosis, and identify Hunk as a previously unrecognized link between the Akt and c-myc oncogenic pathways.

Hunk is required for HER2/neu-induced mammary tumorigenesis.

Understanding the molecular pathways that contribute to the aggressive behavior of human cancers is a critical research priority. The SNF1/AMPK-related protein kinase Hunk is overexpressed in aggressive subsets of human breast, ovarian, and colon cancers. Analysis of Hunk(­/­) mice revealed that this kinase is required for metastasis of c-myc­induced mammary tumors but not c-myc­induced primary tumor formation. Similar to c-myc, amplification of the proto-oncogene HER2/neu occurs in 10%­30% of breast cancers and is associated with aggressive tumor behavior. By crossing Hunk(­/­) mice with transgenic mouse models for HER2/neu-induced mammary tumorigenesis, we report that Hunk is required for primary tumor formation induced by HER2/neu. Knockdown and reconstitution experiments in mouse and human breast cancer cell lines demonstrated that Hunk is required for maintenance of the tumorigenic phenotype in HER2/neu-transformed cells. This requirement is kinase dependent and resulted from the ability of Hunk to suppress apoptosis in association with downregulation of the tumor suppressor p27(kip1). Additionally, we find that Hunk is rapidly upregulated following HER2/neu activation in vivo and in vitro. These findings provide what we believe is the first evidence for a role for Hunk in primary tumorigenesis and cell survival and identify this kinase as an essential effector of the HER2/neu oncogenic pathway.