用于阿尔茨海默病诊断的权值分布特征学习

针对当前基于机器学习的早期阿尔茨海默病（AD）诊断中有标记训练样本不足的问题,提出一种基于多模态特征数据的权值分布稀疏特征学习方法,并将其应用于早期阿尔茨海默病的诊断.具体来说,该诊断方法主要包括两大模块：基于权值分布的Lasso特征选择模型（WDL）和大间隔分布分类机模型（LDM）.首先,为了获取多模态特征之间的数据分布信息,对传统Lasso模型进行改进,引入权值分布正则化项,从而构建出基于权值分布的Lasso特征选择模型;然后,为了有效地利用多模态特征之间的数据分布信息,以保持多模态特征之间的互补性,直接采用大间隔分布学习算法训练分类器.选取国际阿尔茨海默症数据库（ADNI）中202个多模态特征的被试者样本进行实验,分类AD最高平均精度为97.5%,分类轻度认知功能障碍（MCI）最高平均精度为83.1%,分类轻度认知功能障碍转化为AD（pMCI）最高平均精度为84.8%.实验结果表明,所提WDL特征学习方法可从串联的多模态特征学到性能更优的特征子集,并能根据权值分布获取多模态特征之间的数据分布信息,从而提高早期阿尔茨海默病诊断的性能.     

An intelligent hierarchical residual attention learning-based conjoined twin neural network for Alzheimer's stage detection and prediction

Alzheimer's disorder (AD) causes permanent impairment in the brain's memory of the cellular system, leading to the initiation of dementia. Earlier detection of Alzheimer's disease in the initial stages is challenging for researchers. Deep learning and machine learning-based techniques can help resolve many issues associated with brain imaging exploration. Brain MR Images (Brain-MRI) are used to detect Alzheimer's in computable research work. To correctly categorize the stages of Alzheimer's disease, discriminative features need to be extracted from the MR images. Recently, many studies have used deep learning methods for the early detection of this disorder. However, overfitting degrades the deep learning method's performance because the dataset's selection images are smaller and imbalanced. Some studies could not reach more discriminative and effectual attention-aware features for Alzheimer's stage classification to increase the model performance. In this paper, we develop a novel hierarchical residual attention learning-inspired multistage conjoined twin network (HRAL-CTNN) to classify the stages of Alzheimer's. We used augmentation approaches to scale insufficient and imbalanced data. The HRAL-CTNN is efficiently overcoming the issues of not obtaining efficient attention-aware and generative features for Alzheimer's stage classification. The proposed model solved the problem of redundant features by extracting attentive discriminant features, and scaling imbalance data by data augmentation, after that training and validation using HRAL-CTNN. The execution of this proposed work has been performed on the ADNI MRI dataset. This work achieved outstanding accuracy of 99.97 $\pm$ 0.01% and F1 score of 99.30 $\pm$ 0.02% for Alzheimer's stage classification. This model proposed by our group outperformed the existing related studies in terms of the model's performance score.     

A novel method for the classification of Alzheimer's disease from normal controls using magnetic resonance imaging

Alzheimer's disease (AD) is the most prevalent form of dementia. Although fewer people, who suffer from AD are correctly and promptly diagnosed, due to a lack of knowledge of its cause and unavailability of treatment, AD is more manageable if the symptoms of mild cognitive impairment (MCI) are in an early stage. In recent years, computer‐aided diagnosis has been widely used for the diagnosis of AD. The main motive of this paper is to improve the classification and prediction accuracy of AD. In this paper, a novel approach is developed to classify MCI, normal control (NC), and AD using structural magnetic resonance imaging (sMRI) from the Alzheimer's disease Neuroimaging Initiative (ADNI) dataset (50 AD, 50 NC, 50 MCI subjects). FreeSurfer is used to process these MRI data and obtain cortical features such as volume, surface area, thickness, white matter (WM), and intrinsic curvature of the brain regions. These features are modified by normalizing each cortical region's features using the absolute maximum value of that region's features from all subjects in each group of MCI, NC, and AD independently. A total of 420 features are obtained. To address the curse of dimensionality, the obtained features are reduced to 30 features using a sequential feature selection technique. Three classifiers, namely the twin support vector machine (TSVM), least squares TSVM (LSTSVM), and robust energy‐based least squares TSVM (RELS‐TSVM), are used to evaluate the classification accuracy from the obtained features. Five‐fold and 10‐fold cross‐validation are used to validate the proposed method. Experimental results show an accuracy of 100% for the studied database. The proposed approach is innovative due to its higher classification accuracy compared to methods in the existing literature.     

Hybrid Approach to Document Anomaly Detection: An Application to Facilitate RPA in Title Insurance

Anomaly detection (AD) is an important aspect of various domains and title insurance (TI) is no exception. Robotic process automation (RPA) is taking over manual tasks in TI business processes, but it has its limitations without the support of artificial intelligence (AI) and machine learning (ML). With increasing data dimensionality and in composite population scenarios, the complexity of detecting anomalies increases and AD in automated document management systems (ADMS) is the least explored domain. Deep learning, being the fastest maturing technology can be combined along with traditional anomaly detectors to facilitate and improve the RPAs in TI. We present a hybrid model for AD, using autoencoders (AE) and a one-class support vector machine (OSVM). In the present study, OSVM receives input features representing real-time documents from the TI business, orchestrated and with dimensions reduced by AE. The results obtained from multiple experiments are comparable with traditional methods and within a business acceptable range, regarding accuracy and performance.

     

融入平滑组稀疏化的脑部MRI图像分类

目的 阿尔茨海默症（Alzheimer’s disease,AD）是主要的老年病之一,并正向年轻化发展。早期通过核磁共振（magnetic resonance imaging,MRI）图像识别AD的发病阶段,有助于在AD初期及时采取相关干预措施和治疗手段,控制和延缓AD疾病恶化。为此,提出了基于平滑函数的组L1/2稀疏正则化（smooth group L1/2,SGL1/2）方法。方法 通过引入平滑组L1/2正则化实现组内稀疏,并将原先组L1/2方法中含有的非平滑的绝对值函数向平滑函数逼近,解决了组L1/2方法中数值计算振荡和收敛难的缺点。SGL1/2方法能够在保持分类精度的前提下,加速对模型的求解。同时在分类方法中,引入一个校准hinge函数（calibrated hinge,Chinge）代替标准支持向量机（support vector machine,SVM）中的hinge函数,形成校准SVM （calibrated SVM,C-SVM）用于疾病的分类,使处于分类平面附近的样本更倾向于分类的正确一侧,对一些难以区分的样本能够进行更好的分类。结果 与其他组级别上的正则化方法相比,SGL1/2与校准支持向量机结合的分类模型对AD的识别具有更高的分类性能,分类准确率高达94.70%。结论 本文提出的组稀疏分类模型,实现了组间稀疏和组内稀疏的优点,为未来AD的自动诊断提供了客观参照。     

Hybrid-augmented intelligence: collaboration and cognition

The long-term goal of artificial intelligence (AI) is to make machines learn and think like human beings. Due to the high levels of uncertainty and vulnerability in human life and the open-ended nature of problems that humans are facing, no matter how intelligent machines are, they are unable to completely replace humans. Therefore, it is necessary to introduce human cognitive capabilities or human-like cognitive models into AI systems to develop a new form of AI, that is, hybrid-augmented intelligence. This form of AI or machine intelligence is a feasible and important developing model. Hybrid-augmented intelligence can be divided into two basic models: one is human-in-the-loop augmented intelligence with human-computer collaboration, and the other is cognitive computing based augmented intelligence, in which a cognitive model is embedded in the machine learning system. This survey describes a basic framework for human-computer collaborative hybrid-augmented intelligence, and the basic elements of hybrid-augmented intelligence based on cognitive computing. These elements include intuitive reasoning, causal models, evolution of memory and knowledge, especially the role and basic principles of intuitive reasoning for complex problem solving, and the cognitive learning framework for visual scene understanding based on memory and reasoning. Several typical applications of hybrid-augmented intelligence in related fields are given.     

A wavelet extreme learning machine

Extreme learning machine (ELM) has been widely used in various fields to overcome the problem of low training speed of the conventional neural network. Kernel extreme learning machine (KELM) introduces the kernel method to ELM model, which is applicable in Stat ML. However, if the number of samples in Stat ML is too small, perhaps the unbalanced samples cannot reflect the statistical characteristics of the input data, so that the learning ability of Stat ML will be influenced. At the same time, the mix kernel functions used in KELM are conventional functions. Therefore, the selection of kernel function can still be optimized. Based on the problems above, we introduce the weighted method to KELM to deal with the unbalanced samples. Wavelet kernel functions have been widely used in support vector machine and obtain a good classification performance. Therefore, to realize a combination of wavelet analysis and KELM, we introduce wavelet kernel functions to KELM model, which has a mix kernel function of wavelet kernel and sigmoid kernel, and introduce the weighted method to KELM model to balance the sample distribution, and then we propose the weighted wavelet–mix kernel extreme learning machine. The experimental results show that this method can effectively improve the classification ability with better generalization. At the same time, the wavelet kernel functions perform very well compared with the conventional kernel functions in KELM model.     

Patent classification system using a new hybrid genetic algorithm support vector machine

Nowadays, decision-making activities of knowledge-intensive enterprises depend heavily on the successful classification of patents. A considerable amount of time is required to achieve successful classification because of the complexity associated with patent information and of the large number of potential patents. Several different patent classification approaches have been developed in the past, but most of these studies focus on using computational models for the International Patent Classification (IPC) system rather than using these models in real-world cases of patent classification. In contrast to previous studies that combined algorithms and the IPC system directly without using expert screening, this study proposes a novel artificial intelligence (AI)-aided patent decision-making process. In this process, an expert screening approach is integrated with a hybrid genetic-based support vector machine (HGA-SVM) model for developing a patent classification system with the high classification accuracy and generalization ability for real-world patent searching cases. The proposed approach is tested on a real-world case—an expert's patent document searching history that contains 234 patent documents of semiconductor equipment components. The research results demonstrate that our proposed hybrid genetic algorithm approach can optimize all the parameters of the SVM for developing a patent classification system with a high accuracy. The proposed HGA-SVM model is able to dynamically and automatically classify patent documents by recording and learning the experts’ knowledge and logic. Finally, we propose a new decision-making process for improving the development of the SVM patent classification and searching system.     

多媒体技术研究:2017——记忆驱动的媒体学习与创意

目的 借鉴大脑的工作机理来发展人工智能是当前人工智能发展的重要方向之一。注意力与记忆在人的认知理解过程中扮演了重要的角色。由于"端到端"深度学习在识别分类等任务中表现了优异性能,因此如何在深度学习模型中引入注意力机制和外在记忆结构,以挖掘数据中感兴趣的信息和有效利用外来信息,是当前人工智能研究的热点。方法 本文以记忆和注意力等机制为中心,介绍了这些方面的3个代表性工作,包括神经图灵机、记忆网络和可微分神经计算机。在这个基础上,进一步介绍了利用记忆网络的研究工作,其分别是记忆驱动的自动问答、记忆驱动的电影视频问答和记忆驱动的创意（文本生成图像）,并对国内外关于记忆网络的研究进展进行了比较。结果 调研结果表明：1）在深度学习模型中引入注意力机制和外在记忆结构,是当前人工智能研究的热点;2）关于记忆网络的研究越来越多。国内外关于记忆网络的研究正在蓬勃发展,每年发表在机器学习与人工智能相关的各大顶级会议上的论文数量正在逐年攀升;3）关于记忆网络的研究越来越热。不仅每年发表的论文数量越来越多,且每年的增长趋势并没有放缓,2015年增长了9篇,2016年增长了4篇,2017年增长了9篇,2018年增长了14篇;4）基于记忆驱动的手段和方法十分通用。记忆网络已成功地运用于自动问答、视觉问答、物体检测、强化学习、文本生成图像等领域。结论 数据驱动的机器学习方法已成功运用于自然语言、多媒体、计算机视觉、语音等领域,数据驱动和知识引导将是人工智能未来发展的趋势之一。     

High performance accelerators for deep neural networks: A review

The availability of huge structured and unstructured data, advanced highly dense memory and high performance computing machines have provided a strong push for the development in artificial intelligence (AI) and machine learning (ML) domains. AI and machine learning has rekindled the hope of efficiently solving complex problems which was not possible in the recent past. The generation and availability of big-data is a strong driving force for the development of AI/ML applications, however, several challenges need to be addressed, like processing speed, memory requirement, high bandwidth, low latency memory access, and highly conductive and flexible connections between processing units and memory blocks. The conventional computing platforms are unable to address these issues with machine learning and AI. Deep neural networks (DNNs) are widely employed for machine learning and AI applications, like speech recognition, computer vison, robotics, and so forth, efficiently and accurately. However, accuracy is achieved at the cost of high computational complexity, sacrificing energy efficiency and throughput like performance measuring parameters along with high latency. To address the problems of latency, energy efficiency, complexity, power consumption, and so forth, a lot of state of the art DNN accelerators have been designed and implemented in the form of application specific integrated circuits (ASICs) and field programmable gate arrays (FPGAs). This work provides the state of the art of all these DNN accelerators which have been developed recently. Various DNN architectures, their computing units, emerging technologies used in improving the performance of DNN accelerators will be discussed. Finally, we will try to explore the scope for further improvement in these accelerator designs, various opportunities and challenges for the future research.     

基于多样真实任务生成的鲁棒小样本分类方法

随着大数据、计算机与互联网等技术的不断进步,以机器学习和深度学习为代表的人工智能技术取得了巨大成功,尤其是最近不断涌现的各种大模型,极大地加速了人工智能技术在各个领域的应用.但这些技术的成功离不开海量训练数据和充足的计算资源,大大限制了这些方法在一些数据或计算资源匮乏领域的应用.因此,如何利用少量样本进行学习,也就是小样本学习成为以人工智能技术引领新一轮产业变革中一个十分重要的研究问题.小样本学习中最常用的方法是基于元学习的方法,这类方法通过在一系列相似的训练任务上学习解决这类任务的元知识,在新的测试任务上利用元知识可以进行快速学习.虽然这类方法在小样本分类任务上取得了不错的效果,但是这类方法的一个潜在假设是训练任务和测试任务来自同一分布.这意味着训练任务需要足够多才能使模型学到的元知识泛化到不断变化的测试任务中.但是在一些真正数据匮乏的应用场景,训练任务的数量也是难以保证的.为此,提出一种基于多样真实任务生成的鲁棒小样本分类方法(DATG).该方法通过对已有少量任务进行Mixup,可以生成更多的训练任务帮助模型进行学习.通过约束生成任务的多样性和真实性,该方法可以有效提高小样本分类方...     

Ontology driven decision support for the diagnosis of mild cognitive impairment

In recent years, mild cognitive impairment (MCI) has attracted significant attention as an indicator of high risk for Alzheimer's disease (AD), and the diagnosis of MCI can alert patient to carry out appropriate strategies to prevent AD. To avoid subjectivity in diagnosis, we propose an ontology driven decision support method which is an automated procedure for diagnosing MCI through magnetic resonance imaging (MRI). In this approach, we encode specialized MRI knowledge into an ontology and construct a rule set using machine learning algorithms. Then we apply these two parts in conjunction with reasoning engine to automatically distinguish MCI patients from normal controls (NC). The rule set is trained by MRI data of 187 MCI patients and 177 normal controls selected from Alzheimer's Disease Neuroimaging Initiative (ADNI) using C4.5 algorithm. By using a 10-fold cross validation, we prove that the performance of C4.5 with 80.2% sensitivity is better than other algorithms, such as support vector machine (SVM), Bayesian network (BN) and back propagation (BP) neural networks, and C4.5 is suitable for the construction of reasoning rules. Meanwhile, the evaluation results suggest that our approach would be useful to assist physicians efficiently in real clinical diagnosis for the disease of MCI.     

超级计算支撑的新冠肺炎CT影像综合分析辅助系统应用

目的 新冠肺炎（COVID-19）已经成为全球大流行疾病,在全球范围数百万人确诊。基于计算机断层扫描（computed tomography,CT）数据的影像学分析是临床诊断的重要手段。为了实现快速高效高精度地检测,提出了一种超级计算支撑的新冠肺炎CT影像综合分析辅助系统构建方法。方法 系统整个处理流程依次包括输入处理模块、预处理模块、影像学分析子系统和人工智能（artifiaial intelligence,AI）分析子系统4部分。其中影像学分析子系统通过分析肺实变、磨玻璃影和铺路石等影像学典型特征检测是否有肺炎和典型新冠肺炎特征,给出肺炎影像分析结论;AI分析子系统通过构建深度学习模型来区分普通病毒肺炎与新冠肺炎,增加肺炎的筛查甄别能力。结果 系统发布以来,持续稳定地为国内外超过三十家医院与一百多家科研机构提供了新冠肺炎辅助诊断服务和科研支撑,为抗击疫情提供重要支撑。结论 本文提出的超级计算支撑的新冠肺炎CT影像综合分析辅助系统构建方法,取得了应用效果,是一种有效实现快速部署服务、对突发疫情提供高效支撑的服务方式。     

Automated Machine Learning Enabled Cybersecurity Threat Detection in Internet of Things Environment

Recently, Internet of Things (IoT) devices produces massive quantity of data from distinct sources that get transmitted over public networks. Cybersecurity becomes a challenging issue in the IoT environment where the existence of cyber threats needs to be resolved. The development of automated tools for cyber threat detection and classification using machine learning (ML) and artificial intelligence (AI) tools become essential to accomplish security in the IoT environment. It is needed to minimize security issues related to IoT gadgets effectively. Therefore, this article introduces a new Mayfly optimization (MFO) with regularized extreme learning machine (RELM) model, named MFO-RELM for Cybersecurity Threat Detection and classification in IoT environment. The presented MFO-RELM technique accomplishes the effectual identification of cybersecurity threats that exist in the IoT environment. For accomplishing this, the MFO-RELM model pre-processes the actual IoT data into a meaningful format. In addition, the RELM model receives the pre-processed data and carries out the classification process. In order to boost the performance of the RELM model, the MFO algorithm has been employed to it. The performance validation of the MFO-RELM model is tested using standard datasets and the results highlighted the better outcomes of the MFO-RELM model under distinct aspects.     

Creating human digital memories with the aid of pervasive mobile devices

The abundance of mobile and sensing devices, within our environment, has led to a society in which any object, embedded with sensors, is capable of providing us with information. A human digital memory, created with the data from these pervasive devices, produces a more dynamic and data rich memory. Information such as how you felt, where you were and the context of the environment can be established. This paper presents the DigMem system, which utilizes distributed mobile services, linked data and machine learning to create such memories. Along with the design of the system, a prototype has also been developed, and two case studies have been undertaken, which successfully create memories. As well as demonstrating how memories are created, a key concern in human digital memory research relates to the amount of data that is generated and stored. In particular, searching this set of big data is a key challenge. In response to this, the paper evaluates the use of machine learning algorithms, as an alternative to SPARQL, and treats searching as a classification problem. In particular, supervised machine learning algorithms are used to find information in semantic annotations, based on probabilistic reasoning. Our approach produces good results with 100% sensitivity, 93% specificity, 93% positive predicted value, 100% negative predicted value, and an overall accuracy of 97%.     

Advancing an in-memory computing for a multi-accent real-time voice frequency recognition modeling: a comprehensive study of models & mechanism

Multimedia Tools and Applications - In this age of pervasive computing, numerous scientific accomplishments, such as artificial intelligence and machine learning [ML], have conveyed exciting...     

Optimal Machine Learning Enabled Performance Monitoring for Learning Management Systems

Learning Management System (LMS) is an application software that is used in automation, delivery, administration, tracking, and reporting of courses and programs in educational sector. The LMS which exploits machine learning (ML) has the ability of accessing user data and exploit it for improving the learning experience. The recently developed artificial intelligence (AI) and ML models helps to accomplish effective performance monitoring for LMS. Among the different processes involved in ML based LMS, feature selection and classification processes find beneficial. In this motivation, this study introduces Glowworm-based Feature Selection with Machine Learning Enabled Performance Monitoring (GSO-MFWELM) technique for LMS. The key objective of the proposed GSO-MFWELM technique is to effectually monitor the performance in LMS. The proposed GSO-MFWELM technique involves GSO-based feature selection technique to select the optimal features. Besides, Weighted Extreme Learning Machine (WELM) model is applied for classification process whereas the parameters involved in WELM model are optimally fine-tuned with the help of Mayfly Optimization (MFO) algorithm. The design of GSO and MFO techniques result in reduced computation complexity and improved classification performance. The presented GSO-MFWELM technique was validated for its performance against benchmark dataset and the results were inspected under several aspects. The simulation results established the supremacy of GSO-MFWELM technique over recent approaches with the maximum classification accuracy of 0.9589.     

IoTGuardEye:一种面向物联网服务的Web攻击检测方法

在包括物联网(Internet of Things,IoT)设备的绝大部分边缘计算应用中,基于互联网应用技术(通常被称为Web技术)开发的应用程序接口(Application Programming Interface,API)是设备与远程服务器进行信息交互的核心。相比传统的Web应用,大部分用户无法直接接触到边缘设备使用的API,使得其遭受的攻击相对较少。但随着物联网设备的普及,针对API的攻击逐渐成为热点。因此,文中提出了一种面向物联网服务的Web攻击向量检测方法,用于对物联网服务收到的Web流量进行检测,并挖掘出其中的恶意流量,从而为安全运营中心(Security Operation Center,SOC)提供安全情报。该方法在对超文本传输协议(Hypertext Transfer Protocol,HTTP)请求的文本序列进行特征抽取的基础上,针对API请求的报文格式相对固定的特点,结合双向长短期记忆网络(Bidirectional Long Short-Term Memory,BLSTM)实现对Web流量的攻击向量检测。实验结果表明,相比基于规则的Web应用防火墙(Web Application Firewall,WAF)和传统的机器学习方法,所提方法针对面向物联网服务API的攻击具有更好的识别能力。     

基于AI+IOT的智慧家居系统

利用人工智能技术和深度学习算法,设计开发了基于AI+IOT的智慧家居系统。基于百度提供的免费的语音识别云平台,该系统使用ZigBee网络,对家居环境数据进行采集、分析,并通过物联网技术和人工智能技术实现远程语音控制各种家电的功能。基于深度学习,系统通过百度语音识别技术对自然语言进行语音识别,通过搭建系统编译环境成功融合了AI技术和IOT技术实现了具有语音控制功能的智能家居系统,致力于为人们提供更加便捷智能的生活。