Network embedding (NE) is playing a critical role in network analysis, due to its ability represent vertices with efficient low-dimensional vectors. However, existing NE models aim learn fixed context-free for each vertex and neglect the diverse roles when interacting other vertices. In this paper, we assume that one usually shows different aspects neighbor vertices, should own embeddings respectively. Therefore, present Context-Aware Embedding (CANE), novel model address issue. CANE learns...

In this research, we propose two variants of the Firefly Algorithm (FA), namely inward intensified exploration FA (IIEFA) and compound (CIEFA), for undertaking obstinate problems initialization sensitivity local optima traps K-means clustering model. To enhance capability both exploitation exploration, matrix-based search parameters dispersing mechanisms are incorporated into proposed models. We first replace attractiveness coefficient with a randomized control matrix in IIEFA model to...

We develop a cyclical blockwise coordinate descent algorithm for the multi-task Lasso that efficiently solves problems with thousands of features and tasks. The main result shows closed-form Winsorization operator can be obtained sup-norm penalized least squares regression. This allows to find solutions very large-scale far more than existing methods. complements pioneering work Friedman, et al. (2007) single-task Lasso. As case study, we use as variable selector discover semantic basis...

Most existing deep reinforcement learning (DRL) frameworks consider either discrete action space or continuous solely. Motivated by applications in computer games, we the scenario with discrete-continuous hybrid space. To handle space, previous works approximate discretization, relax it into a set. In this paper, propose parametrized Q-network (P- DQN) framework for without approximation relaxation. Our algorithm combines spirits of both DQN (dealing space) and DDPG seamlessly integrating...

Abstract National governments now recognize online hate speech as a pernicious social problem. In the wake of political votes and terror attacks, incidents offline are known to peak in tandem. This article examines whether an association exists between both forms hate, independent ‘trigger’ events. Using Computational Criminology that draws on data science methods, we link police crime, census Twitter establish temporal spatial targets race religion, racially religiously aggravated crimes...

Due to the vast and rapid increase in size of data, machine learning has become an increasingly more popular approach for purpose knowledge discovery predictive modelling. For both above purposes, it is essential have a data set partitioned into training test set. In particular, used towards model then evaluating performance learned from The split two sets, however, influence on performance, only been investigated with respect optimal proportion no attention paid characteristics within sets....

A rule-based system is a special type of expert system, which typically consists set if–then rules. Such rules can be used in the real world for both academic and practical purposes. In general, systems are involved knowledge discovery tasks purposes predictive modeling latter purpose. context granular computing, each that make up seen as granule. This due to fact granulation general means decomposition whole into several parts. Similarly, rule number terms. From this point view, term also...

Handwritten digits recognition has been treated as a multi-class classification problem in the machine learning context, where each of ten (0–9) is viewed class and task essentially to train classifier that can effectively discriminate classes. In practice, it very usual performance single trained using standard algorithm varied on different datasets, which indicates same may strong classifiers some datasets but weak be other datasets. It also possible shows test sets, especially when...

Sentiment analysis is a very popular application area of text mining and machine learning. The methods include support vector machine, naive bayes, decision trees, deep neural networks. However, these generally belong to discriminative learning, which aims distinguish one class from others with clear-cut outcome, under the presence ground truth. In context classification, instances are naturally fuzzy (can be multilabeled in some areas) thus not considered clear-cut, especially given fact...

In recent years, the increasing prevalence of hate speech in social media has been considered as a serious problem worldwide. Many governments and organizations have made significant investment detection techniques, which also attracted attention scientific community. Although plenty literature focusing on this issue is available, it remains difficult to assess performances each proposed method, its own advantages disadvantages. A general way improve overall results classification by fusing...

It is essential to accurately estimate the state of health (SOH) for lithium-ion batteries from perspectives safety and reliability. However, most existing data-driven methods are based on charging or discharging data, which relatively difficult apply. This paper proposes a novel SOH estimation approach relaxation voltage reconstruction long short-term memory network with an attention mechanism (AM-LSTM) method. First, complete curve reconstructed gated recurrent unit (GRU) neural network....

User intent classification plays a vital role in dialogue systems. Since user may frequently change over time many realistic scenarios, unknown (new) detection has become an essential problem, where the study just begun. This paper proposes semantic-enhanced Gaussian mixture model (SEG) for detection. In particular, we utterance embeddings with distribution and inject dynamic class semantic information into means, which enables learning more class-concentrated that help to facilitate...

Abstract The use of skeleton data for human posture recognition is a key research topic in the human-computer interaction field. To improve accuracy recognition, new algorithm based on multiple features and rule learning proposed this paper. Firstly, 219-dimensional vector that includes angle distance defined. Specifically, are defined terms local relationship between joints global spatial location joints. Then, during classification, method used together with Bagging random subspace methods...

Due to the rapid development of human–computer interaction, affective computing has attracted more and attention in recent years. In emotion recognition, Electroencephalogram (EEG) signals are easier be recorded than other physiological experiments not easily camouflaged. Because high dimensional nature EEG data diversity human emotions, it is difficult extract effective features recognize patterns. This paper proposes a multi-feature deep forest (MFDF) model identify emotions. The firstly...

Sentiment analysis aims to identify the polarity of a document through natural language processing, text and computational linguistics. Over last decade, there has been much focus on sentiment as data available on-line grown exponentially include many based documents (reviews, feedback, articles). Many approaches consider machine learning techniques or statistical analysis, but little use fuzzy classifiers in this field especially considering ambiguity suitability deal with ambiguity. This...

Sentiment analysis, which is also known as opinion mining, aims to recognise the attitude or emotion of people through natural language processing, text analysis and computational linguistics.In recent years, many studies have focused on sentiment classification in context machine learning, e.g. identify that a positive negative.In particular, bag-of-words method has been popularly used transform textual data into structured data, order enable direct use learning algorithms for...

Starcraft II (SC2) is widely considered as the most challenging Real Time Strategy (RTS) game. The underlying challenges include a large observation space, huge (continuous and infinite) action partial observations, simultaneous move for all players, long horizon delayed rewards local decisions. To push frontier of AI research, Deepmind Blizzard jointly developed StarCraft Learning Environment (SC2LE) testbench complex decision making systems. SC2LE provides few mini games such MoveToBeacon,...

Offensive or antagonistic language targeted at individuals and social groups based on their personal characteristics (also known as cyber hate speech cyberhate) has been frequently posted widely circulated via the World Wide Web. This can be considered a key risk factor for individual societal tension surrounding regional instability. Automated Web-based cyberhate detection is important observing understanding community tension—especially in online networks where posts rapidly viewed...

Handwritten character recognition has been profoundly studied for many years in the field of pattern recognition. Due to its vast practical applications and financial implications, handwritten is still an important research area. In this research, a Ethiopian Character Recognition (HECR) dataset prepared train model. Images HECR were organized with more than one color pen RGB main spaces that are size normalized 28 × pixels. The combination scripts (Fidel Ethiopia), numerical...