As aggregators, online news portals face great challenges in continuously selecting a pool of candidate articles to be shown their users. Typically, those are recommended manually by platform editors from much larger aggregated multiple sources. Such hand-pick process is labor intensive and time-consuming. In this paper, we study the editor article selection behavior propose learning demonstration system automatically select subset large pool. Our data analysis shows that (i) editors'...

Reinforcement learning (RL) has recently been introduced to interactive recommender systems (IRS) because of its nature from dynamic interactions and planning for long-run performance. As IRS is always with thousands items recommend (i.e., actions), most existing RL-based methods, however, fail handle such a large discrete action space problem thus become inefficient. The work that tries deal the by utilizing deep deterministic policy gradient framework suffers inconsistency between...

Link prediction is a fundamental problem with wide range of applications in various domains, which predicts the links that are not yet observed or may appear future. Most existing works this field only focus on modeling single network, while real-world networks actually aligned each other. Network alignments contain valuable additional information for understanding networks, and provide new direction addressing data insufficiency alleviating cold start problem. However, there rare leveraging...

With the development of urban road traffic, noise pollution is becoming a public concern. Controlling and reducing harm caused by traffic have been hot spots management research. The subjective annoyance level has become one most important measurements for evaluating pollution. There are experimental methods objective prediction to assess noise: method usually uses social surveys or listening experiments in laboratories directly level, which highly reliable, but often requires lot time...

Reinforcement learning (RL) has recently been introduced to interactive recommender systems (IRS) because of its nature from dynamic interactions and planning for long-run performance. As IRS is always with thousands items recommend (i.e., actions), most existing RL-based methods, however, fail handle such a large discrete action space problem thus become inefficient. The work that tries deal the by utilizing deep deterministic policy gradient framework suffers inconsistency between...

Model combination, often regarded as a key sub-field of ensemble learning, has been widely used in both academic research and industry applications. To facilitate this process, we propose implement an easy-to-use Python toolkit, combo, to aggregate models scores under various scenarios, including classification, clustering, anomaly detection. In nutshell, combo provides unified consistent way combine raw pretrained from popular machine learning libraries, e.g., scikit-learn, XGBoost,...

Medical image based computer aided diagnosis is considers to be an important and challenging task, it has extracted more research work in recent years. Due its interdisciplinarity complexity, there remain many problems not solved. In this paper, a novel method named SeCED proposed, which utilized as the core mechanism of our medical encephalopathy system. The built on two-level architecture, where kM-DBSCAN algorithm employ base clusterer each level k-Medoids select subset for ensemble....

Although several automatic computer systems have been proposed to address facial expression recognition problems, the majority of them still fail cope with some requirements many practical application scenarios. In this paper, one most influential and common issues raised in scenarios when applying system, head pose variation, is comprehensively explored investigated. order do this, two novel texture feature representations are for implementing multi-view environments. These combine...

In the context of facial expression recognition (FER), this paper reviews fundamental theories emotions and further explains key dimensions a defined emotional space. The main contribution is to propose set novel categorization methods for expressions be used in design an automatic FER system. This enables interpreted better way that more effective practical applications systems. order validate feasibility proposed methods, experiments reported which investigates analyzes influence brings multi-view

Subway engineering construction is characterized by large scale, cross-temporal and cross-region, etc. In view of the problems many data sources, poor business collaboration, visualization uneven information management in current projects. From perspective integrated

To ensure the security of top-level design gateway, we proposed a method formally designing and verifying typical gateway. Firstly, designed gateway’s policy according to its requirements. Secondly, modeled verified model’s internal consistency by means BLP model. In end, between functional specifications make sure reasoning procedure’s correctness, used theorem prover Isabelle/HOL describe above work help us deduce. Our ensures gateway in terms exerts certain referential significance on formal

Objective To investigate the value of combined digital mammography and Color Doppler Mammasonography in diagnosis breast carcinoma with ROC analysis.Methods Fifty female patients fifty benign lesions confirmed by means surgery histopathological examinations were studied.All examined Digital Mammagraphy Mammasonography.Images all sorted divided into three groups,Digital radiographs was first group,and second group photographs,and third included above two photographs.Each evaluated...

Facial expressions can be seen as a form of non-verbal communication well primary means conveying social information among humans.Automatic facial expression recognition (FER) applied to wide range scenarios in human-computer interaction, animation, entertainment, and psychology studies. For feature representation FER system, various texture descriptors have been employed derive an effective solution for this system. However, these individual descriptor-based systems often failed achieve...