iATC_Deep-mISF: A Multi-Label Classifier for Predicting the Classes of Anatomical Therapeutic Chemicals by Deep Learning

The recent worldwide spreading of pneumonia-causing virus, such as Coronavirus, COVID-19, and H1N1, has been endangering the life of human beings all around the world. To provide useful clues for developing antiviral drugs, information of anatomical therapeutic chemicals is vitally important. In view of this, a CNN based predictor called “iATC_Deep-mISF” has been developed. The predictor is particularly useful in dealing with the multi-label systems in which some chemicals may occur in two or more different classes. To maximize the convenience for most experimental scientists, a user-friendly web-server for the new predictor has been established at http://www.jci-bioinfo.cn/iATC_Deep-mISF/, which will become a very powerful tool for developing effective drugs to fight pandemic coronavirus and save the mankind of this planet.

Multi-label learning of face demographic classification for correlation analysis

In this paper, we utilize the framework of multi-label learning for face demographic classification. We also attempt t;o explore the suitable classifiers and features for face demographic classification. Three most popular demographic information, gender, ethnicity and age are considered in experiments. Based on the results from demographic classification, we utilize statistic analysis to explore the correlation among various face demographic information. Through the analysis, we draw several conclusions on the correlation and interaction among these high-level face semantic, and the obtained results can be helpful in automatic face semantic annotation and other face analysis tasks.

Optimization Model and Algorithm for Multi-Label Learning

This paper studies a kind of urban security risk assessment model based on multi-label learning, which is transformed into the solution of linear equations through a series of transformations, and then the solution of linear equations is transformed into an optimization problem. Finally, this paper uses some classical optimization algorithms to solve these optimization problems, the convergence of the algorithm is proved, and the advantages and disadvantages of several optimization methods are compared.

Feature Selection for Multi-label Classification Using Neighborhood Preservation

Multi-label learning deals with data associated with a set of labels simultaneously. Dimensionality reduction is an important but challenging task in multi-label learning. Feature selection is an efficient technique for dimensionality reduction to search an optimal feature subset preserving the most relevant information. In this paper, we propose an effective feature evaluation criterion for multi-label feature selection, called neighborhood relationship preserving score. This criterion is inspired by similarity preservation, which is widely used in single-label feature selection. It evaluates each feature subset by measuring its capability in preserving neighborhood relationship among samples. Unlike similarity preservation, we address the order of sample similarities which can well express the neighborhood relationship among samples, not just the pairwise sample similarity. With this criterion, we also design one ranking algorithm and one greedy algorithm for feature selection problem. The proposed algorithms are validated in six publicly available data sets from machine learning repository. Experimental results demonstrate their superiorities over the compared state-of-the-art methods.

pLoc_Deep-mHum: Predict Subcellular Localization of Human Proteins by Deep Learning

Recently, the life of human beings around the entire world has been endangering by the spreading of pneumonia-causing virus, such as Coronavirus, COVID-19, and H1N1. To develop effective drugs against Coronavirus, knowledge of protein subcellular localization is indispensable. In 2019, a predictor called “pLoc_bal-mHum” was developed for identifying the subcellular localization of human proteins. Its predicted results are significantly better than its counterparts, particularly for those proteins that may simultaneously occur or move between two or more subcellular location sites. However, more efforts are definitely needed to further improve its power since pLoc_bal-mHum was still not trained by a “deep learning”, a very powerful technique developed recently. The present study was devoted to incorporate the “deep-learning” technique and develop a new predictor called “pLoc_Deep-mHum”. The global absolute true rate achieved by the new predictor is over 81% and its local accuracy is over 90%. Both are overwhelmingly superior to its counterparts. Moreover, a user-friendly web-server for the new predictor has been well established at http://www.jci-bioinfo.cn/pLoc_Deep-mHum/, which will become a very useful tool for fighting pandemic coronavirus and save the mankind of this planet.

pLoc_Deep-mEuk: Predict Subcellular Localization of Eukaryotic Proteins by Deep Learning

Recently, the life of worldwide human beings has been endangering by the spreading of pneu- monia-causing virus, such as Coronavirus, COVID-19, and H1N1. To develop effective drugs against Coronavirus, knowledge of protein subcellular localization is prerequisite. In 2019, a predictor called “pLoc_bal-mEuk” was developed for identifying the subcellular localization of eukaryotic proteins. Its predicted results are significantly better than its counterparts, particularly for those proteins that may simultaneously occur or move between two or more subcellular location sites. However, more efforts are definitely needed to further improve its power since pLoc_bal-mEuk was still not trained by a “deep learning”, a very powerful technique developed recently. The present study was devoted to incorporating the “deep- learning” technique and developed a new predictor called “pLoc_Deep-mEuk”. The global absolute true rate achieved by the new predictor is over 81% and its local accuracy is over 90%. Both are overwhelmingly superior to its counterparts. Moreover, a user-friendly web- server for the new predictor has been well established at http://www.jci-bioinfo.cn/pLoc_Deep-mEuk/, by which the majority of experimental scientists can easily get their desired data.

pLoc_Deep-mVirus: A CNN Model for Predicting Subcellular Localization of Virus Proteins by Deep Learning

The recent worldwide spreading of pneumonia-causing virus, such as Coronavirus, COVID-19, and H1N1, has been endangering the life of human beings all around the world. In order to really understand the biological process within a cell level and provide useful clues to develop antiviral drugs, information of virus protein subcellular localization is vitally important. In view of this, a CNN based virus protein subcellular localization predictor called “pLoc_Deep-mVirus” was developed. The predictor is particularly useful in dealing with the multi-sites systems in which some proteins may simultaneously occur in two or more different organelles that are the current focus of pharmaceutical industry. The global absolute true rate achieved by the new predictor is over 97% and its local accuracy is over 98%. Both are transcending other existing state-of-the-art predictors significantly. It has not escaped our notice that the deep-learning treatment can be used to deal with many other biological systems as well. To maximize the convenience for most experimental scientists, a user-friendly web-server for the new predictor has been established at http://www.jci-bioinfo.cn/pLoc_Deep-mVirus/.

pLoc_Deep-mPlant: Predict Subcellular Localization of Plant Proteins by Deep Learning

Current coronavirus pandemic has endangered mankind life. The reported cases are increasing exponentially. Information of plant protein subcellular localization can provide useful clues to develop antiviral drugs. To cope with such a catastrophe, a CNN based plant protein subcellular localization predictor called “pLoc_Deep-mPlant” was developed. The predictor is particularly useful in dealing with the multi-sites systems in which some proteins may simultaneously occur in two or more different organelles that are the current focus of pharmaceutical industry. The global absolute true rate achieved by the new predictor is over 95% and its local accuracy is about 90%?-?100%. Both have substantially exceeded the?other existing state-of-the-art predictors. To maximize the convenience for most?experimental scientists, a user-friendly web-server for the new predictor has been established?at?http://www.jci-bioinfo.cn/pLoc_Deep-mPlant/, by which the majority of experimental?scientists can easily obtain their desired data without the need to go through the?mathematical details.

pLoc_Deep-mGneg: Predict Subcellular Localization of Gram Negative Bacterial Proteins by Deep Learning

The recent worldwide spreading of pneumonia-causing virus, such as Coronavirus, COVID-19, and H1N1, has been endangering the life of human beings all around the world. In order to really understand the biological process within a cell level and provide useful clues to develop antiviral drugs, information of Gram negative bacterial protein subcellular localization is vitally important. In view of this, a CNN based protein subcellular localization predictor called “pLoc_Deep-mGnet” was developed. The predictor is particularly useful in dealing with the multi-sites systems in which some proteins may simultaneously occur in two or more different organelles that are the current focus of pharmaceutical industry. The global absolute true rate achieved by the new predictor is over 98% and its local accuracy is around 94% - 100%. Both are transcending other existing state-of-the-art predictors significantly. To maximize the convenience for most experimental scientists, a user-friendly web-server for the new predictor has been established at http://www.jci-bioinfo.cn/pLoc_Deep-mGneg/, which will become a very useful tool for fighting pandemic coronavirus and save the mankind of this planet.

pLoc_Deep-mGpos: Predict Subcellular Localization of Gram Positive Bacteria Proteins by Deep Learning

The recent worldwide spreading of pneumonia-causing virus, such as Coronavirus, COVID-19, and H1N1, has been endangering the life of human beings all around the world. In order to really understand the biological process within a cell level and provide useful clues to develop antiviral drugs, information of Gram positive bacteria protein subcellular localization is vitally important. In view of this, a CNN based protein subcellular localization predictor called “pLoc_Deep-mGpos” was developed. The predictor is particularly useful in dealing with the multi-sites systems in which some proteins may simultaneously occur in two or more different organelles that are the current focus of pharmaceutical industry. The global absolute true rate achieved by the new predictor is over 99% and its local accuracy is around 92% - 99%. Both are transcending other existing state-of-the-art predictors significantly. To maximize the convenience for most experimental scientists, a user-friendly web-server for the new predictor has been established at http://www.jci-bioinfo.cn/pLoc_Deep-mGpos/, which will become a very powerful tool for developing effective drugs to fight pandemic coronavirus and save the mankind of this planet.

基于预训练模型与BiLSTM-CNN的多标签代码坏味检测方法

为了提高多标签代码坏味检测的准确率,提出一种基于预训练模型与BiLSTM-CNN的多标签代码坏味检测方法DMSmell(deep multi-smell)。首先,利用静态分析工具获取源代码中的文本信息和结构度量信息,并采用2种检测规则对代码坏味实例进行标记;其次,利用CodeBERT预训练模型生成文本信息对应的词向量,并分别采用BiLSTM和CNN对词向量和结构度量信息进行深度特征提取;最后,结合注意力机制和多层感知机,完成多标签代码坏味的检测,并对DMSmell方法进行了性能评估。结果表明:DMSmell方法在一定程度上提高了多标签代码坏味检测的准确率,与基于分类器链的方法相比,精确匹配率提高了1.36个百分点,微查全率提高了2.45个百分点,微F1提高了1.1个百分点。这表明,将文本信息与结构度量信息相结合,并利用深度学习技术进行特征提取和分类,可以有效提高代码坏味检测的准确性,为多标签代码坏味检测的研究和应用提供重要的参考。

基于多模态共享网络的自监督语音-人脸跨模态关联学习方法

现有的语音-人脸跨模态关联学习方法在语义关联和监督信息方面仍然面临挑战,尚未充分考虑语音与人脸之间的语义信息交互。为解决这些问题,提出一种基于多模态共享网络的自监督关联学习方法。首先,将语音和人脸模态的特征映射到单位球面,构建一个公共的特征空间;接着,通过多模态共享网络的残差块来挖掘复杂的非线性数据关系,并利用其中权重共享的全连接层来增强语音与人脸特征向量之间的关联性;最后,使用K均值聚类算法生成的伪标签作为监督信号来指导度量学习,从而完成4种跨模态关联学习任务。实验结果表明,本文提出的方法在语音-人脸跨模态验证、匹配和检索任务上均取得了良好的效果,多项评价指标相较于现有基线方法提升1%~4%的准确率。

基于ALBERT-Seq2Seq模型的多标签农业文本分类方法

针对多标签分类采用现有静态词向量模型无法捕获文本完整语义的问题,文中结合ALBERT与序列到序列模型,提出一种用于农业文本多标签分类的神经网络模型ALBERT-Seq2Seq。该模型采用ALBERT预训练语言模型动态获取农业文本语义信息,利用其内部多层双向Transformer架构挖掘农业文本信息的深层特征,接着引入Seq2Seq模型构造出多标签分类器并进行训练。在AGRI-ML2020农业文本多标签数据集上进行算法性能测试,实验结果表明,该模型分类F1值达89.5%,能够有效提升农业文本多标签分类效果。

Binary relevance for multi-label learning： an overview

Multi-label learning deals with problems where each example is represented by a single instance while being associated with multiple class labels simultaneously. Binary relevance is arguably the most intuitive solution for learning from multi-label examples. It works by decomposing the multi-label learning task into a number of independent binary learning tasks （one per class label）. In view of its potential weakness in ignoring correlations between labels, many correlation-enabling extensions to binary relevance have been proposed in the past decade. In this paper, we aim to review the state of the art of binary relevance from three perspectives. First, basic settings for multi-label learning and binary relevance solutions are briefly summarized. Second, representative strategies to provide binary relevance with label correlation exploitation abilities are discussed. Third, some of our recent studies on binary relevance aimed at issues other than label correlation exploitation are introduced. As a conclusion, we provide suggestions on future research directions.

Learnability of multi-instance multi-label learning

Multi-instance multi-label learning(MIML) is a new machine learning framework where one data object is described by multiple instances and associated with multiple class labels.During the past few years,many MIML algorithms have been developed and many applications have been described.However,there lacks theoretical exploration to the learnability of MIML.In this paper,through proving a generalization bound for multi-instance single-label learner and viewing MIML as a number of multi-instance single-label learning subtasks with the correlation among the labels,we show that the MIML hypothesis class constructed from a multi-instance single-label hypothesis class is PAC-learnable.

Multi-label active learning by model guided distribution matching

Multi-label learning is an effective framework for learning with objects that have multiple semantic labels, and has been successfully applied into many real-world tasks, In contrast with traditional single-label learning, the cost of la- beling a multi-label example is rather high, thus it becomes an important task to train an effective multi-label learning model with as few labeled examples as possible. Active learning, which actively selects the most valuable data to query their labels, is the most important approach to reduce labeling cost. In this paper, we propose a novel approach MADM for batch mode multi-label active learning. On one hand, MADM exploits representativeness and diversity in both the feature and label space by matching the distribution between labeled and unlabeled data. On the other hand, it tends to query predicted positive instances, which are expected to be more informative than negative ones. Experiments on benchmark datasets demonstrate that the proposed approach can reduce the labeling cost significantly.

Classifying Syndromes in Chinese Medicine Using Multi-label Learning Algorithm with Relevant Features for Each Label

Objective: To develop an effective Chinese Medicine(CM) diagnostic model of coronary heart disease(CHD) and to confirm the scientific validity of CM theoretical basis from an algorithmic viewpoint. Methods: Four types of objective diagnostic data were collected from 835 CHD patients by using a selfdeveloped CM inquiry scale for the diagnosis of heart problems, a tongue diagnosis instrument, a ZBOX-I pulse digital collection instrument, and the sound of an attending acquisition system. These diagnostic data was analyzed and a CM diagnostic model was established using a multi-label learning algorithm(REAL). Results: REAL was employed to establish a Xin(Heart) qi deficiency, Xin yang deficiency, Xin yin deficiency, blood stasis, and phlegm five-card CM diagnostic model, which had recognition rates of 80.32%, 89.77%, 84.93%, 85.37%, and 69.90%, respectively. Conclusions: The multi-label learning method established using four diagnostic models based on mutual information feature selection yielded good recognition results. The characteristic model parameters were selected by maximizing the mutual information for each card type. The four diagnostic methods used to obtain information in CM, i.e., observation, auscultation and olfaction, inquiry, and pulse diagnosis, can be characterized by these parameters, which is consistent with CM theory.

一种基于Tri-training的半监督多标记学习文档分类算法

多标记学习主要用于解决因单个样本对应多个概念标记而带来的歧义性问题,而半监督多标记学习是近年来多标记学习任务中的一个新的研究方向,它试图综合利用少量的已标记样本和大量的未标记样本来提高学习性能。为了进一步挖掘未标记样本的信息和价值并将其应用于文档多标记分类问题,该文提出了一种基于Tritraining的半监督多标记学习算法(MKSMLT),该算法首先利用k近邻算法扩充已标记样本集,结合Tri-training算法训练分类器,将多标记学习问题转化为标记排序问题。实验表明,该算法能够有效提高文档分类性能。

基于共享背景主题的Labeled LDA模型

隐藏狄利克雷分配(Latent Dirichlet Allocation,LDA)模型被广泛应用于文本分析、图像识别等领域.但由于LDA及其扩展模型多为无监督学习模型,无法将其应用于分类任务中.本文通过研究文档标记与LDA模型中主题的映射关系,提出一种新的Labeled LDA模型(Shared Background Topics Labeled LDA,SBTL-LDA).在SBTL-LDA模型中每个标记除了存在若干个独享的局部主题外,还存在若干个共享的背景(Background)主题,这样可以有效分析不同标记所含主题之间的依赖关系,而文档标记被映射为局部主题和共享主题的组合,因此SBTL-LDA模型可以有效提升文档标记判别的准确性.同时SBTL-LDA模型还可以看成是一种半监督聚类模型,在对文档进行聚类分析的过程中模型可以有效的利用文档的标记信息提升文档聚类效果.实验证明SBTL-LDA模型能够有效解决PLDA模型中主题之间的相似性和依赖关系,具有良好的多标记判别能力,并且具有优于LDA、PLDA模型的文档聚类效果.

基于Tri-training的半监督多标记学习算法

传统的多标记学习是监督意义下的学习,它要求获得完整的类别标记.但是当数据规模较大且类别数目较多时,获得完整类别标记的训练样本集是非常困难的.因而,在半监督协同训练思想的框架下,提出了基于Tri-training的半监督多标记学习算法(SMLT).在学习阶段,SMLT引入一个虚拟类标记,然后针对每一对类别标记,利用协同训练机制Tri-training算法训练得到对应的分类器;在预测阶段,给定一个新的样本,将其代入上述所得的分类器中,根据类别标记得票数的多少将多标记学习问题转化为标记排序问题,并将虚拟类标记的得票数作为阈值对标记排序结果进行划分.在UCI中4个常用的多标记数据集上的对比实验表明,SMLT算法在4个评价指标上的性能大多优于其他对比算法,验证了该算法的有效性.