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

物件導向軟體之架構(Object-Oriented Software Construction)探討

This paper studies the control of a class of discrete event processes, i.e. processes that are discrete, asynchronous and possibly nondeter-ministic. The controlled process is described as the generator of a formal language, while the controller, or supervisor, is constructed from the grammar of a specified target language that incorporates the desired closed-loop system behavior. The existence problem for a supervisor is reduced to finding the largest controllable language contained in a given legal language. Two examples are provided.

/pdf/supervisory-control-of-a-class-of-discrete-event-processes-45fsmhdhbo.pdf

Supervisory control of a class of discrete event processes

Since the subject of traffic dynamics has captured the interest of physicists, many surprising effects have been revealed and explained. Some of the questions now understood are the following: Why are vehicles sometimes stopped by ``phantom traffic jams'' even though drivers all like to drive fast? What are the mechanisms behind stop-and-go traffic? Why are there several different kinds of congestion, and how are they related? Why do most traffic jams occur considerably before the road capacity is reached? Can a temporary reduction in the volume of traffic cause a lasting traffic jam? Under which conditions can speed limits speed up traffic? Why do pedestrians moving in opposite directions normally organize into lanes, while similar systems ``freeze by heating''? All of these questions have been answered by applying and extending methods from statistical physics and nonlinear dynamics to self-driven many-particle systems. This article considers the empirical data and then reviews the main approaches to modeling pedestrian and vehicle traffic. These include microscopic (particle-based), mesoscopic (gas-kinetic), and macroscopic (fluid-dynamic) models. Attention is also paid to the formulation of a micro-macro link, to aspects of universality, and to other unifying concepts, such as a general modeling framework for self-driven many-particle systems, including spin systems. While the primary focus is upon vehicle and pedestrian traffic, applications to biological or socio-economic systems such as bacterial colonies, flocks of birds, panics, and stock market dynamics are touched upon as well.

Traffic and related self-driven many-particle systems

https://www.ufz.de/export/data/2/100065_GrimmODD.pdf

A standard protocol for describing individual-based and agent-based models

Markov Modeling is among the most commonly used forms of model expression and Markov concepts of states and state transitions are fully compatible with the DEVS characterization of discrete event systems. Besides their general usefulness, the Markov concepts of stochastic modeling are implicitly at the heart of most forms of discrete event simulation and are a natural basis for the extended and integrated Markov modeling facility discussed in this paper. DEVS Markov models are full-fledged DEVS models and can be coupled with other DEVS components in hierarchical compositions. Due to their explicit transition and time advance structure, DEVS Markov models can be individualized with specific transition probabilities and transition times/rates which can be changed during model execution for dynamic structural change. This paper presents the formal concepts underlying DEVS Markov models and how they are implemented in MS4 Me, also discussing how the facilities differ from other Markov M&S tools.

DEVS markov modeling and simulation: formal definition and implementation

DEVS representation and aggregation of spatially distributed systems: speed-versus-error tradeoffs

This paper describes the design of a model-based autonomous planning system that will enable robots to manage a space-borne chemical laboratory. In a model-based planning system, knowledge is encapsulated in the form of models at the various layers to support the predefined system objectives. Thus the model-based approach can be considered as an extended planning paradigm which is able to base its planning, control, diagnosis, repair, and other activities on a variety of objectives-related models. A System Entity Structure/Model Base framework is employed to support autonomous system design through the ability to generate a family of planning alternatives as well as to build hierarchical event-based control structures. The model base is a multilevel, multiabstraction, and multiformalism system organized through the use of system morphisms to integrate related models.

Model-based task planning system for a space laboratory environment

14 There is a rapidly growing demand to model and simulate complex large-scale distributed systems and to collaboratively share geographically dispersed data assets and computing resources to perform such distributed simulation with reasonable conmiunication and computation resources. Interest management schemes have been studied in the literature. In this dissertation we propose an interest-based quantization scheme that is created by combining a quantization scheme and an interest management scheme. We show that this approach provides a superior solution to reduce message traffic and network data transmission load. As an environmental platform for data distribution management, we extended the DEVS/HLA distributed modeling and simulation environment. This environment allows us to study interest-based quantization schemes in order to achieve effective reduction of data communication in distributed simulation. In this environment, system modeling is provided by the DEVS (Discrete Event System Specification) formalism and supports effective modeling based on hierarchical and modular object-oriented technology. Distributed simulation is performed by a highly reliable facility using the HLA (High Level Architecture). The extended DEVS/HLA environment, called DEVS/GDDM (Generic Data Distribution Management), provides a high level abstraction to specify a set of interest-based quantization schemes. This dissertation presents a performance analysis of centralized and distributed configurations to study the scalability of the interest-based quantization schemes. These 15 results illustrate the advantages of using space-based quantization in reducing both network load and overall simulation execution time. A real world application, relating to ballistic missiles simulation, demonstrates the operation of the DEVS/GDDM environment. Theoretical and empirical results of the ballistic missiles application show that the space-based quantization scheme, especially with predictive and multiplexing extensions, is very effective and scalable due to reduced local computation demands and extremely favorable conmiunication data reduction with a reasonably small potential for error. This realistic case study establishes that the DEVS/GDDM environment can provide scalable distributed simulation for practical, real-world applications.

Space-based data management for high performance distributed simulation

In this paper we address the question of the expressibility of discrete event specified systems (DEVS), i.e., we characterize the subclass of dynamical systems which can be homomorphically represented by DEVS models. We show that causal dynamical systems with piecewise constant input and output segment spaces are DEVS-representable. Moreover, DEVS-representable dynamical systems are closed under coupling, i.e., that a valid coupling of DEVS-representable dynamical systems is a DEVS-representable dynamical system. This justifies hierarchical, modular construction of both DEVS models and real-world (continuous or discrete) counterpart systems. Furthermore, we investigate a subclass of dynamical systems in detail, viz., discretely interacting continuous systems with internal dynamics specified by differential equations. This class is important in the rapidly developing field of hybrid autonomous systems control. As an application to CAST (computer-aided system technology), we show that this class is amenable to DEVS representation and therefore, also new forms of high performance parallel/distributed discrete event simulation.

Bernard P. Zeigler

Papers

DEVS markov modeling and simulation: formal definition and implementation

DEVS representation and aggregation of spatially distributed systems: speed-versus-error tradeoffs

Model-based task planning system for a space laboratory environment

Space-based data management for high performance distributed simulation

On the Expressibility of Discrete Event Specified Systems