Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

Data Mining - Concepts and Techniques.

Social Network Analysis

We study the problem of representation learning in heterogeneous networks. Its unique challenges come from the existence of multiple types of nodes and links, which limit the feasibility of the conventional network embedding techniques. We develop two scalable representation learning models, namely metapath2vec and metapath2vec++. The metapath2vec model formalizes meta-path-based random walks to construct the heterogeneous neighborhood of a node and then leverages a heterogeneous skip-gram model to perform node embeddings. The metapath2vec++ model further enables the simultaneous modeling of structural and semantic correlations in heterogeneous networks. Extensive experiments show that metapath2vec and metapath2vec++ are able to not only outperform state-of-the-art embedding models in various heterogeneous network mining tasks, such as node classification, clustering, and similarity search, but also discern the structural and semantic correlations between diverse network objects.

https://dl.acm.org/doi/pdf/10.1145/3097983.3098036

metapath2vec: Scalable Representation Learning for Heterogeneous Networks

The Analysis of Time Series: An Introduction, 4th edn. By C. Chatfield. ISBN 0 412 31820 2. Chapman and Hall, London, 1989. 242 pp. £13.50.

The Analysis of Time Series: An Introduction

In many data-intensive applications, workflow is often used as an important model for organizing data processing tasks and resource provisioning is an important and challenging problem for improving the performance of workflows. Recently, system variations in the cloud and large-scale clusters, such as those in I/O and network performances and failure events, have been observed to greatly affect the performance of workflows. Traditional resource provisioning methods, which overlook these variations, can lead to suboptimal resource provisioning results. In this article, we provide a general solution for workflow performance optimizations considering system variations. Specifically, we model system dynamics as time-dependent random variables and take their probability distributions as optimization input. Despite its effectiveness, this solution involves heavy computation overhead. Thus, we propose three pruning techniques to simplify workflow structure and reduce the probability evaluation overhead. We implement our techniques in a runtime library, which allows users to incorporate efficient probabilistic optimization into existing resource provisioning methods. Experiments show that probabilistic solutions can improve the performance by up to 65 percent compared to state-of-the-art static solutions, and our pruning techniques can greatly reduce the overhead of our probabilistic approach.

Taming System Dynamics on Resource Optimization for Data Processing Workflows: A Probabilistic Approach

In database systems that collect information about the external environment, such as temperature and location values, it is often infeasible to obtain accurate information due to measurement and sampling errors, and resource limitations. Queries evaluated over these inaccurate data can potentially yield incorrect results. To avoid these problems. the idea of using uncertainty models (such as an interval associated with a probability density function) instead of a single value for modeling a data item has been explored in recent years. These works have focussed on simple queries such as range and nearest-neighbor queries. Queries that join multiple relations have not been addressed in earlier work despite the significance of joins in databases. In this paper we address join queries over uncertain data. As with other queries over uncertain data, these joins return probabilistic answers. A probabilistic Join Query (PJQ) augments the results with probability guarantees to indicate the likelihood of each join tuple being part of the result. Traditional join operators, such as equality and inequality, need to be extended to support uncertain data. In this paper, we present the notion of equality and inequality operators for uncertainty. vVe also introduce the concept of "approximation" in these comparison operators. Although PJQs are more informative than traditional joins. they are expensive to evaluate. To overcome this problem, we observe that often it is only necessary to know whether the probability of the results exceeds a given threshold. instead of the precise probability value. By incorporating this constraint into PJQ, it is possible to achieve much better performance. In particular, we develop three sets of optimization techniques, namely item-leveL page-level and index-level pruning. for different join operators. These techniques facilitate pruning with little space and time overhead, and are easily adapted to most join algorithms. Extensive simulation results show that these techniques improve the performance of joins significantly.

/pdf/efficient-join-processing-over-uncertain-data-technical-54s9m1zlrf.pdf

Efficient Join Processing over Uncertain Data Technical Report

Given a large weighted directed graph where nodes are associated with attributes and edges are weighted, we study a new problem, called preferential nearest neighbors (NN) browsing, in this paper. In such browsing, a user may provide one or more source nodes and some keywords to retrieve the nearest neighbors of those source nodes that contain the given keywords. For example, when a tourist has a plan to visit several places (source nodes), he/she would like to search hotels with some preferred features (e.g., Internet and swimming pools). It is highly desirable to recommend a list of near hotels with those preferred features, in order of the road network distance to the places (source nodes) the tourist wants to visit. The existing approach by graph traversal at querying time requires long query processing time, and the approach by maintenance of the pre-computed all-pairs shortest distances requires huge storage space on disk. In this paper, we propose new approaches to support on-line preferential NN browsing. The data graphs we are dealing with are weighted directed graphs where nodes are associated with attributes, and the distances between nodes to be found are the exact distances in the graph. We focus ourselves on two-step approaches. In the first step, we identify a number of reference nodes (also called centers) which exist alone on some shortest paths between a source node and a preferential NN node that contains the user-given keywords. In the second step, we find the preferential NN nodes within a certain distance to the source nodes via the relevant reference nodes, using an index that supports both textural (attributes) and and the distance. Our approach tightly integrates NN search with the preference search, which is confirmed to be efficient and effective to find any preferential NN nodes.

On-line preferential nearest neighbor browsing in large attributed graphs

This paper focuses on the problem of tagging quality evaluation in collaborative tagging systems. By investigating the dynamics of tagging process, we find that high frequency tags almost cover the main aspects of a resource content and can be determined stable much earlier than a whole tag set. Motivated by this finding, we design the swapping index and smart moving index on tagging quality. We also study the correlations in tag usage and propose the semantic measurement on tagging quality. The proposed methods are evaluated against real datasets and the results show that they are more efficient than previous methods, which are appropriate for a large number of web resources. The effectiveness is justified by the results in tag based applications. The light weight metrics bring a little loss on the performance, while the semantic metric is better than current methods.

Fast and Semantic Measurements on Collaborative Tagging Quality

Crowdsourced ranking and top-k queries have attracted significant attention recently. Their goal is to combine human cognitive abilities and machine intelligence to rank computer hostile but human friendly items. Many task assignment algorithms and inference approaches have been proposed to publish suitable micro-tasks to the crowd, obtain informative answers, and aggregate the rank from noisy human answers. However, they are all focused on single query processing. To the best of our knowledge, no prior work helps users manage multiple crowdsourced top-k queries. We propose a toolkit, which seamlessly works with most existing inference and task assignment methods, for crowdsourced top-k query management. Our toolkit attempts to optimize human resource allocation and continuously monitors query quality at any stage of the crowdsourcing process. A user can terminate a query early, if the estimated quality already fulfills her requirements. Besides, the toolkit provides user-friendly interfaces for users to initialize queries, monitor execution status, and do more operations by hand.

Reynold Cheng

Papers

Taming System Dynamics on Resource Optimization for Data Processing Workflows: A Probabilistic Approach

Efficient Join Processing over Uncertain Data Technical Report

On-line preferential nearest neighbor browsing in large attributed graphs

Fast and Semantic Measurements on Collaborative Tagging Quality

A Toolkit for Managing Multiple Crowdsourced Top-K Queries