Data Mining - Concepts and Techniques.

What is Twitter

Anomaly detection is an important problem that has been well-studied within diverse research areas and application domains. The aim of this survey is two-fold, firstly we present a structured and comprehensive overview of research methods in deep learning-based anomaly detection. Furthermore, we review the adoption of these methods for anomaly across various application domains and assess their effectiveness. We have grouped state-of-the-art research techniques into different categories based on the underlying assumptions and approach adopted. Within each category we outline the basic anomaly detection technique, along with its variants and present key assumptions, to differentiate between normal and anomalous behavior. For each category, we present we also present the advantages and limitations and discuss the computational complexity of the techniques in real application domains. Finally, we outline open issues in research and challenges faced while adopting these techniques.

Deep Learning for Anomaly Detection: A Survey.

Emotion recognition from speech signals is an important but challenging component of Human-Computer Interaction (HCI). In the literature of speech emotion recognition (SER), many techniques have been utilized to extract emotions from signals, including many well-established speech analysis and classification techniques. Deep Learning techniques have been recently proposed as an alternative to traditional techniques in SER. This paper presents an overview of Deep Learning techniques and discusses some recent literature where these methods are utilized for speech-based emotion recognition. The review covers databases used, emotions extracted, contributions made toward speech emotion recognition and limitations related to it.

/pdf/speech-emotion-recognition-using-deep-learning-techniques-a-zl8wl0vs5t.pdf

Speech Emotion Recognition Using Deep Learning Techniques: A Review

Emotion recognition using deep learning approach from audio–visual emotional big data

Intelligent video surveillance is a challenging issue due to complicated scenes. Based on empirical and experimental explorations, we propose a multi-person tracking-by-detection framework to achieve pedestrian counting at run time. This framework is integrated with a stream based cloud computing paradigm to improve tracking performance. We evaluated our approach which shows improved time performance compared with those classical approaches.

Online Multiperson Tracking and Counting with Cloud Computing

Pyruvate kinase M2 (PKM2), as the terminal and last rate-limiting enzyme of the glycolytic pathway, is an ideal enzyme for regulating metabolic phenotype. PKM2 tetramer activation has shown a protective role against diabetic kidney disease (DKD). However, the molecular mechanisms involved in diabetic tubular have not been investigated so far. In this study, we performed transcriptome gene expression profiling in human renal proximal tubular epithelial cell line (HK-2 cells) treated with 25 mM high D-glucose (HG) for 7 days before the addition of 10 μM TEPP-46, an activator of PKM2 tetramerization, for a further 1 day in the presence of HG. Afterwards, we analyzed the differentially expressed (DE) genes and investigated gene relationships based on weighted gene co-expression network analysis. The results showed that 2,902 DE genes were identified (adjusted P-value ≤ 0.05), where 2,509 DE genes (86.46%) were co-expressed in the key module. Four extremely downregulated DE genes (HSPA8, HSPA2, HSPA1B, and ARRB1) and three extremely upregulated DE genes (GADD45A, IGFBP3, and SIAH1) enriched in the downregulated endocytosis (hsa04144) and upregulated p53 signaling pathway (hsa04115), respectively, were validated by qRT-PCR experiments. The qRT-PCR results showed that the relative expression levels of HSPA8 [adjusted P-value = 4.45 × 10-34 and log2(FC) = -1.12], HSPA2 [adjusted P-value = 6.09 × 10-14 and log2(FC) = -1.27], HSPA1B [adjusted P-value = 1.14 × 10-11 and log2(FC) = -1.02], and ARRB1 [adjusted P-value = 2.60 × 10-5 and log2(FC) = -1.13] were significantly different (P-value < 0.05) from the case group to the control group. Furthermore, the interactions and predicted microRNAs of the key genes (HSPA8, HSPA2, HSPA1B, and ARRB1) were visualized in networks. This study identified the key candidate transcriptomic biomarkers and biological pathways in hyperglycemic HK-2 cells responding to the PKM2 activator TEPP-46 that can highlight a possibility of PKM2 tetramerization reshaping the interplay among endocytic trafficking through the versatile networks of Hsp70s and rewiring the crosstalk between EGFR signal transduction circuits and metabolic stress to promote resilience, which will be valuable for further research on PKM2 in DKD.

/pdf/transcriptomic-signatures-responding-to-pkm2-activator-tepp-irurzh9u.pdf

Transcriptomic signatures responding to PKM2 activator TEPP-46 in the hyperglycemic human renal proximal epithelial tubular cells

The parallel intelligence era characterized by virtual reality interaction is coming, and the battle for the dominant power of online public opinion battlefield is becoming fiercer. It is urgent to establish an information processing system suitable for the characteristics of the times. In this paper, an artificial intelligence system based on multi-agent system is designed. Artificial intelligence system is composed of agents, agent interaction environment and agent interaction rules. Agents update their states according to behavior rules and the perception to the environment. Combined with the theme crawler, the real-time intercommunication between the artificial intelligence system and the online public opinion dissemination system is realized, and a real-time, online, closed-loop and adaptive parallel intelligence Communities is builded. Our experiments show that the multiagent artificial intelligence system has excellent performance in intelligence prediction and intelligence guidance.

MAS-based Parallel Intelligence Communities

Surveillance cameras are being rapidly deployed for facilitating smart transportation. Recognizing the vehicle license plate from massive videos becomes a challenge in context of system scalability and efficiency. This paper proposes a novel algorithm for scalable and efficient license plate recognition (SELPR). The SELPR algorithm first locates the license plate using a YOLO (You Look Only Once) network and recognizes the license plate using multi-label convolutional neural network (Multi-label CNN). We deploy the SELPR algorithm to the Apache Spark framework to evaluate its scalability and efficiency in parallel processing. The results demonstrates that SELPR can achieve synthesized performance with 95% recognition accuracy, better processing efficiency and scalability on a Spark cluster.

/pdf/a-scalable-and-efficient-multi-label-cnn-based-license-plate-1w92pnk2du.pdf

A Scalable and Efficient Multi-Label CNN-Based License Plate Recognition on Spark

Ship license plate recognition is challenging due to the diversity of plate locations and text types. This paper proposes a DCNN-based (deep convolutional neural network) online adaptive real-time ship license plate recognition approach, namely, DRASLPR, which consists of three steps. First, it uses a Single Shot MultiBox Detector (SSD) to detect a ship. Then, it detects the ship license plate with a designed detector. Third, DRASLPR recognizes the ship license plate. The proposed DRASLPR has been deployed at Dongying Port, China and the running results show the effectiveness of DRASLPR.

/pdf/dcnn-based-real-time-adaptive-ship-license-plate-recognition-3z98i1tkpb.pdf

Xin Liu

Papers

Online Multiperson Tracking and Counting with Cloud Computing

Transcriptomic signatures responding to PKM2 activator TEPP-46 in the hyperglycemic human renal proximal epithelial tubular cells

MAS-based Parallel Intelligence Communities

A Scalable and Efficient Multi-Label CNN-Based License Plate Recognition on Spark

DCNN Based Real-Time Adaptive Ship License Plate Recognition (DRASLPR)