Scalable Machine Learning on Compact Data Representations

With massive high-dimensional data now commonplace in research and industry, there is a strong and growing demand for more scalable computational techniques for data analysis and knowledge discovery. Key to turning these data into knowledge is the ability to learn statistical models with high interpretability. Current methods for learning statistical models either produce models that are not interpretable or have prohibitive computational costs when applied to massive data. In this paper we address this need by presenting a scalable algorithm for partial least squares regression (PLS), which we call compression-based PLS (cPLS), to learn predictive linear models with a high interpretability from massive high-dimensional data. We propose a novel grammar-compressed representation of data matrices that supports fast row and column access while the data matrix is in a compressed form. The original data matrix is grammar-compressed and then the linear model in PLS is learned on the compressed data matrix, which results in a significant reduction in working space, greatly improving scalability. We experimentally test cPLS on its ability to learn linear models for classification, regression and feature extraction with various massive high-dimensional data, and show that cPLS performs superiorly in terms of prediction accuracy, computational efficiency, and interpretability.

Massive high dimensional data has a large sample size and high dimensionality. However, all the features of the massive high dimensional data are used for identification or classification, which will increase the calculation time and reduce the accuracy of identification or classification. Therefore, extracting features with strong expressive ability is very important for processing massive high-dimensional data. In order to solve this problem, this paper presents a feature extraction method which based on Stacked Denoising Autoencoder (SDAE). SDAE trains each layer of neural network by unsupervised layer-by-layer greedy training. Then supervised training Softmax classifier. And finally uses Back Propagation (BP) algorithm to optimize the entire model. In this paper, the ISOLET data set is taken as experimental data. The experimental results demonstrated that our proposed method can extract feature subsets with strong expressive ability. And using this feature subset for classification, the classification accuracy is significantly improved.

Due to the intricate relationship between different dimensions of high-dimensional data, subspace analysis is often conducted to decompose dimensions and give prominence to certain subsets of dimensions, i.e. subspaces. Exploring and comparing subspaces are important to reveal the underlying features of subspaces, as well as to portray the characteristics of individual dimensions. To date, most of the existing high-dimensional data exploration and analysis approaches rely on dimensionality reduction algorithms (e.g. principal component analysis and multi-dimensional scaling) to project high-dimensional data, or their subspaces, to two-dimensional space and employ scatterplots for visualization. However, the dimensionality reduction algorithms are sometimes difficult to fine-tune and scatterplots are not effective for comparative visualization, making subspace comparison hard to perform. In this article, we aggregate high-dimensional data or their subspaces by computing pair-wise distances between all data items and showing the distances with matrix visualizations to present the original high-dimensional data or subspaces. Our approach enables effective visual comparisons among subspaces, which allows users to further investigate the characteristics of individual dimensions by studying their behaviors in similar subspaces. Through subspace comparisons, we identify dominant, similar, and conforming dimensions in different subspace contexts of synthetic and real-world high-dimensional data sets. Additionally, we present a prototype that integrates parallel coordinates plot and matrix visualization for high-dimensional data exploration and incremental dimensionality analysis, which also allows users to further validate the dimension characterization results derived from the subspace comparisons.

Decision analysis and optimization are examples of inverse management problems which can be effectively handled by means of Artificial Intelligence (AT) and simulation techniques. However, the application of these techniques in specific industrial contexts finds major constraints in the lack of formalized methodologies and systematic procedures to retrieve and handle the data required for their implementation. The paper proposes Knowledge Discovery (KD) techniques as means of data pre-processing for the development of Al-based decision support systems (DSSs). This use of Knowledge Discovery techniques is illustrated in relation to the design of a hybrid DSS - combining AI and simulation techniques - for the management of maintenance services in the public transport sector.

In recent years both the number and the size of organisational databases have increased rapidly. However, although available processing power has also grown, the increase in stored data has not necessarily led to a corresponding increase in useful information and knowledge. This has led to a growing interest in the development of tools capable of harnessing the increased processing power available to better utilise the potential of stored data. The terms Knowledge Discovery in Databases and Mining have been adopted for a field of research dealing with the automatic discovery of knowledge implicit within databases. Data mining is useful in situations where the volume of data is either too large or too complicated for manual processing or, to a lesser extent, where human experts are unavailable to provide knowledge. The success already attained by a wide range of data mining applications has continued to prompt further investigation into alternative data mining techniques and the extension of data mining to new domains. This paper surveys, from the standpoint of the database systems community, current issues in data mining research by examining the architectural and process models adopted by knowledge discovery systems, the different types of discovered knowledge, the way knowledge discovery systems operate on different data types, various techniques for knowledge discovery and the ways in which discovered knowledge is used.

Waste container weighing data processing to create reliable information of household waste generation

The development of advanced recording and measurement devices in scientific fields is producing high-dimensional time series data. Vector autoregressive (VAR) models are well suited for inferring Granger-causal networks from high dimensional time series data sets, but accurate inference at scale remains a central challenge. We have recently introduced a flexible and scalable statistical machine learning framework, Union of Intersections (UoI), which enables low false-positive and low false-negative feature selection along with low bias and low variance estimation, enhancing interpretation and predictive accuracy. In this paper, we scale the UoI framework for VAR models (algorithm UoI V AR ) to infer network connectivity from large time series data sets (TBs). To achieve this, we optimize distributed convex optimization and introduce novel strategies for improved data read and data distribution times. We study the strong and weak scaling of the algorithm on a Xeon-phi based supercomputer (100,000 cores). These advances enable us to estimate the largest VAR model as known (1000 nodes, corresponding to 1M parameters) and apply it to large time series data from neurophysiology (192 neurons) and finance (470 companies).

In this paper, we give an overview of knowledge discovery (KD) and information extraction (IE) techniques on the World Wide Web (WWW). We intend to answer the following questions: What kind of additional uncertainty challenges are introduced by the WWW setting to basic KD and IE techniques? What are the fundamental techniques that can be used to reduce such uncertainty and achieve reasonable KD and IE performance on the WWW? What is the impact of each novel method? What types of interactions can be conducted between these techniques and information networks to make them benefit from each other? In what way can we utilize the results in more interesting applications? What are the remaining challenges and what are the possible ways to address these challenges? We hope this can provide a road map to advance KD and IE on the WWW to a higher level of performance, portability and utilization.

Contemporary single-cell technologies produce data with a vast number of variables at a rapid pace, making large volumes of high-dimensional data available. The exploratory analysis of such high dimensional data can be aided by intuitive low dimensional visualizations. In this work, we investigate how both discrete and continuous structures in single cell data can be captured using the recently proposed dimensionality reduction method SONG, and compare the results with commonly used methods UMAP and PHATE. Using simulated and real-world datasets, we observed that SONG preserves a variety of patterns including discrete clusters, continuums, and branching structures. More importantly, SONG produced more/equally insightful visualizations compared to UMAP and PHATE in all considered datasets. We also quantitatively validate the high-dimensional pairwise distance preservation ability of these visualization methods in the low dimensional space for the generated visualizations.

Applications of medical image analysis are often faced with the challenge of modelling high-dimensional data with relatively few samples. In many settings, normal or healthy samples are prevalent while pathological samples are rarer, highly diverse, and/or difficult to model. In such cases, a robust model of the normal population in the high-dimensional space can be useful for characterizing pathologies. In this context, there is utility in hybrid models, such as probabilistic PCA, which learns a low-dimensional model, commensurates with the available data, and combines it with a generic, isotropic noise model for the remaining dimensions. However, the isotropic noise model ignores the inherent correlations that are evident in so many high-dimensional data sets associated with images and shapes in medicine. This paper describes a method for estimating a Gaussian model for collections of images or shapes that exhibit underlying correlations, e.g., in the form of smoothness. The proposed method incorporates a Gaussian-process noise model within a generative formulation. For optimization, we derive a novel expectation maximization (EM) algorithm. We demonstrate the efficacy of the method on synthetic examples and on anatomical shape data.

This chapter covers the state-of-the-art multivariate statistical methods designed for high dimensional multiset omics data analysis. Recent biotechnological developments have enabled large-scale measurement of various biomolecular data, such as genotypic and phenotypic data, dispersed over various omics domains. An emergent research direction is to analyze these data sources using an integrated approach to better model and understand the underlying biology of complex disease conditions. However, comprehensive analysis techniques that can handle both the size and complexity, and at the same time can account for the hierarchical structure of such data, are lacking. An overview of some of the developments in multivariate techniques for high-dimensional omics data analysis, highlighting two well-known multivariate methods, canonical correlation analysis (CCA) and redundancy analysis (RDA), is provided in this chapter. Penalized versions of CCA are widespread in the omics data analysis field, and there is recent work on multiset penalized RDA that is applicable to multiset omics data. How these methods meet the statistical challenges that come with high-dimensional multiset omics data analysis and help to further our understanding of the human condition in terms of health and disease are presented. Additionally, the current challenges to be resolved in the field of omics data analysis are discussed.

Uncovering driver genes is crucial for understanding heterogeneity in cancer. L1-type regularization approaches have been widely used for uncovering cancer driver genes based on genome-scale data. Although the existing methods have been widely applied in the field of bioinformatics, they possess several drawbacks: subset size limitations, erroneous estimation results, multicollinearity, and heavy time consumption. We introduce a novel statistical strategy, called a Recursive Random Lasso (RRLasso), for high dimensional genomic data analysis and investigation of driver genes. For time-effective analysis, we consider a recursive bootstrap procedure in line with the random lasso. Furthermore, we introduce a parametric statistical test for driver gene selection based on bootstrap regression modeling results. The proposed RRLasso is not only rapid but performs well for high dimensional genomic data analysis. Monte Carlo simulations and analysis of the “Sanger Genomics of Drug Sensitivity in Cancer dataset from the Cancer Genome Project” show that the proposed RRLasso is an effective tool for high dimensional genomic data analysis. The proposed methods provide reliable and biologically relevant results for cancer driver gene selection.

When looking at a scene, we frequently move our eyes to place consecutive interesting regions on the fovea, the retina centre. At each fixation, only this specific foveal region is analysed in detail by the visual system. The visual attention mechanisms control eye movements and depend on two types of factor: bottom-up and top-down factors. Bottom-up factors include different visual features such as colour, luminance, edges, and orientations. In this paper, we evaluate quantitatively the relative contribution of basic low-level features as candidate guiding factors to visual attention and hence to eye movements. We also study how these visual features can be combined in a bottom-up saliency model. Our work consists of three interactive parts: a functional saliency model, a statistical model and eye movement data recorded during free viewing of natural scenes. The functional saliency model, inspired by the primate visual system, decomposes a visual scene into different feature maps. The statistical model indicates which features best explain the recorded eye movements. We show an essential role of high frequency luminance and an important contribution of central fixation bias. The relative contribution of features, calculated by the statistical model, is then used to combine the different feature maps into a saliency map. Finally, the comparison between the saliency model and experimental data confirmed the influence of these contributions.

Overlay analysis is a common task in geographic computing that is widely used in geographic information systems, computer graphics, and computer science. With the breakthroughs in Earth observation technologies, particularly the emergence of high-resolution satellite remote-sensing technology, geographic data have demonstrated explosive growth. The overlay analysis of massive and complex geographic data has become a computationally intensive task. Distributed parallel processing in a cloud environment provides an efficient solution to this problem. The cloud computing paradigm represented by Spark has become the standard for massive data processing in the industry and academia due to its large-scale and low-latency characteristics. The cloud computing paradigm has attracted further attention for the purpose of solving the overlay analysis of massive data. These studies mainly focus on how to implement parallel overlay analysis in a cloud computing paradigm but pay less attention to the impact of spatial data graphics complexity on parallel computing efficiency, especially the data skew caused by the difference in the graphic complexity. Geographic polygons often have complex graphical structures, such as many vertices, composite structures including holes and islands. When the Spark paradigm is used to solve the overlay analysis of massive geographic polygons, its calculation efficiency is closely related to factors such as data organization and algorithm design. Considering the influence of the shape complexity of polygons on the performance of overlay analysis, we design and implement a parallel processing algorithm based on the Spark paradigm in this paper. Based on the analysis of the shape complexity of polygons, the overlay analysis speed is improved via reasonable data partition, distributed spatial index, a minimum boundary rectangular filter and other optimization processes, and the high speed and parallel efficiency are maintained.

Data mining, a key technique in knowledge discovery, is the process of identifying useful patterns from a collection of data. This process is made difficult for complex data combining, for example, numeric and symbolic attributes, or also when the number of observations is large. In this paper, we present a pattern mining approach to identify local correlations in the data, that is to say, sets of numerical attributes that strongly co-vary together in a subset of the data. The sets of numerical attributes and the subset of data are automatically (inductively) identified by the method. Whereas the space of patterns to be potentially explored is exponential, the complexity of the problem can be overcome by using sampling techniques that have several advantages: (1) reducing the computation cost, (2) identifying most important patterns, and (3) making possible to process large databases by distributed computing on multiple machines.