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

Deep learning has revolutionized many machine learning tasks in recent years, ranging from image classification and video processing to speech recognition and natural language understanding. The data in these tasks are typically represented in the Euclidean space. However, there is an increasing number of applications, where data are generated from non-Euclidean domains and are represented as graphs with complex relationships and interdependency between objects. The complexity of graph data has imposed significant challenges on the existing machine learning algorithms. Recently, many studies on extending deep learning approaches for graph data have emerged. In this article, we provide a comprehensive overview of graph neural networks (GNNs) in data mining and machine learning fields. We propose a new taxonomy to divide the state-of-the-art GNNs into four categories, namely, recurrent GNNs, convolutional GNNs, graph autoencoders, and spatial–temporal GNNs. We further discuss the applications of GNNs across various domains and summarize the open-source codes, benchmark data sets, and model evaluation of GNNs. Finally, we propose potential research directions in this rapidly growing field.

/pdf/a-comprehensive-survey-on-graph-neural-networks-4lw9i4v183.pdf

A Comprehensive Survey on Graph Neural Networks

On robust estimation of the location parameter

Even though human movement and mobility patterns have a high degree of freedom and variation, they also exhibit structural patterns due to geographic and social constraints. Using cell phone location data, as well as data from two online location-based social networks, we aim to understand what basic laws govern human motion and dynamics. We find that humans experience a combination of periodic movement that is geographically limited and seemingly random jumps correlated with their social networks. Short-ranged travel is periodic both spatially and temporally and not effected by the social network structure, while long-distance travel is more influenced by social network ties. We show that social relationships can explain about 10% to 30% of all human movement, while periodic behavior explains 50% to 70%. Based on our findings, we develop a model of human mobility that combines periodic short range movements with travel due to the social network structure. We show that our model reliably predicts the locations and dynamics of future human movement and gives an order of magnitude better performance than present models of human mobility.

/pdf/friendship-and-mobility-user-movement-in-location-based-2u0rrxuc9l.pdf

Friendship and mobility: user movement in location-based social networks

Multi-label learning studies the problem where each example is represented by a single instance while associated with a set of labels simultaneously. During the past decade, significant amount of progresses have been made toward this emerging machine learning paradigm. This paper aims to provide a timely review on this area with emphasis on state-of-the-art multi-label learning algorithms. Firstly, fundamentals on multi-label learning including formal definition and evaluation metrics are given. Secondly and primarily, eight representative multi-label learning algorithms are scrutinized under common notations with relevant analyses and discussions. Thirdly, several related learning settings are briefly summarized. As a conclusion, online resources and open research problems on multi-label learning are outlined for reference purposes.

https://romisatriawahono.net/lecture/rm/survey/machine%20learning/Zhang%20-%20Multi-Label%20Learning%20Algorithms%20-%202013.pdf

A Review On Multi-Label Learning Algorithms

Traffic signal control in cities today is not well optimized according to the feedback received from the real world. And such an inefficiency in traffic signal control results in people’s waste of time in commuting, road rage in the traffic jam, and high cost for city operation. Recently, deep reinforcement learning (DRL) approaches shed lights to better optimize traffic signal plans according to the feedback received from the environment. Most of these methods are evaluated in a simulated environment, but can not be applied to intersections in the real world directly, as the training of DRL relies on a great amount of samples and takes a long time to converge. In this paper, we propose a batch learning framework where the targeted transfer reinforcement learning (TTRL-B) is introduced to speed up learning. Specifically, a separate unsupervised method is designed to measure the similarities of traffic conditions to select the suitable source intersection for transfer. The proposed framework allows batch learning and this is the first work to consider the impact of slow learning in RL on real-world applications. Experiments on real traffic data demonstrate that our model accelerates learning with good performance.

Targeted Knowledge Transfer for Learning Traffic Signal Plans

Academic search engines and digital libraries provide convenient online search and access facilities for scientific publications. However, most existing systems do not include books in their collections although several books are freely available online. Academic books are different from papers in terms of their length, contents and structure. We argue that accounting for academic books is important in understanding and assessing scientific impact. We introduce an open-book search engine that extracts and indexes metadata, contents, and bibliography from online PDF book documents. To the best of our knowledge, no previous work gives a systematical study on building a search engine for books. We propose a hybrid approach for extracting title and authors from a book that combines results from CiteSeer, a rule based extractor, and a SVM based extractor, leveraging web knowledge. For "table of contents" recognition, we propose rules based on multiple regularities based on numbering and ordering. In addition, we study bibliography extraction and citation parsing for a large dataset of books. Finally, we use the multiple fields available in books to rank books in response to search queries. Our system can effectively extract metadata and contents from large collections of online books and provides efficient book search and retrieval facilities.

/pdf/searching-online-book-documents-and-analyzing-book-citations-2jh6kxdrjg.pdf

Searching online book documents and analyzing book citations

Traffic simulations can help to explore novel and efficient transportation solutions that overcome traffic problems such as traffic jams and road planning. Traditional traffic simulators usually leverage a car-following model to simulate the vehicle’s behavior in the real-world traffic environment. However, these calibrated simplified physical models often fail to accurately predict the pattern of vehicle’s movement in complicated real-world traffic environment. Considering the complexity and non-linearity of the real-world traffic, this paper unprecedentedly treat the problem of traffic simulation as a learning problem, and proposes learning to simulate (L2S) vehicle trajectory. We use the generative adversarial imitation learning framework to estimate the policy that provides sequential decisions for the vehicle given real-world demonstrations. The experiment on real-world traffic data shows the superior performance in simulating vehicle trajectories of our method compared to traditional traffic simulation approaches.

Learning to Simulate Vehicle Trajectories from Demonstrations

Representation leaning on networks aims to embed networks into a low-dimensional vector space, which is useful in many tasks such as node classification, network clustering, link prediction and recommendation. In reality, most real-life networks constantly evolve over time with various kinds of changes to the network structure, e.g., creation and deletion of edges. However, existing network embedding methods learn the representation vectors for nodes in a static manner, which are not suitable for dynamic network embedding. In this paper, we propose a dynamic network embedding approach for large-scale networks. The method incrementally updates the embeddings by considering the changes of the network structures and is able to dynamically learn the embedding for networks with millions of nodes within a few seconds. Extensive experimental results on three real large-scale networks demonstrate the efficiency and effectiveness of our proposed methods.

Representation Learning for Large-Scale Dynamic Networks

Real-time traffic monitoring becomes an essential part of an intelligent city. In recent years, the adoption of surveillance cameras is rapidly growing because they are helpful to manage and control the traffic. However, it is impossible to install cameras on every road in a city due to the high costs of deployment and maintenance. Given the information from limited surveillance cameras, can we infer the citywide traffic volume accurately? This is a challenging question because we have no historical data on the roads without cameras. In this paper, we propose a framework named CityVolInf to infer citywide traffic volume based on surveillance camera records. Our framework combines a semi-supervised learning-based similarity module with a novel simulation module to address the above challenges. While the similarity module focuses on spatiotemporal correlations of traffic volume between road segments, the simulation module utilizes incomplete trajectories to model transitions of traffic volume between adjacent road segments. Our framework bridges the conventional data-driven approach and transportation domain knowledge from the simulator. Extensive experiments on a real-world dataset, containing 405,370,631 camera records collected from 1,704 surveillance cameras over a period of 31 days in Jinan, China, demonstrate the effectiveness of CityVolInf compared with existing methods.

Zhenhui Li

Papers

Targeted Knowledge Transfer for Learning Traffic Signal Plans

Searching online book documents and analyzing book citations

Learning to Simulate Vehicle Trajectories from Demonstrations

Representation Learning for Large-Scale Dynamic Networks

Citywide Traffic Volume Inference with Surveillance Camera Records