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

The Compact Muon Solenoid (CMS) detector is described. The detector operates at the Large Hadron Collider (LHC) at CERN. It was conceived to study proton-proton (and lead-lead) collisions at a centre-of-mass energy of 14 TeV (5.5 TeV nucleon-nucleon) and at luminosities up to 10(34)cm(-2)s(-1) (10(27)cm(-2)s(-1)). At the core of the CMS detector sits a high-magnetic-field and large-bore superconducting solenoid surrounding an all-silicon pixel and strip tracker, a lead-tungstate scintillating-crystals electromagnetic calorimeter, and a brass-scintillator sampling hadron calorimeter. The iron yoke of the flux-return is instrumented with four stations of muon detectors covering most of the 4 pi solid angle. Forward sampling calorimeters extend the pseudo-rapidity coverage to high values (vertical bar eta vertical bar <= 5) assuring very good hermeticity. The overall dimensions of the CMS detector are a length of 21.6 m, a diameter of 14.6 m and a total weight of 12500 t.

/pdf/the-cms-experiment-at-the-cern-lhc-4d2umuzboo.pdf

The CMS experiment at the CERN LHC

An Introduction to MultiAgent Systems.

The Smart Grid, regarded as the next generation power grid, uses two-way flows of electricity and information to create a widely distributed automated energy delivery network. In this article, we survey the literature till 2011 on the enabling technologies for the Smart Grid. We explore three major systems, namely the smart infrastructure system, the smart management system, and the smart protection system. We also propose possible future directions in each system. colorred{Specifically, for the smart infrastructure system, we explore the smart energy subsystem, the smart information subsystem, and the smart communication subsystem.} For the smart management system, we explore various management objectives, such as improving energy efficiency, profiling demand, maximizing utility, reducing cost, and controlling emission. We also explore various management methods to achieve these objectives. For the smart protection system, we explore various failure protection mechanisms which improve the reliability of the Smart Grid, and explore the security and privacy issues in the Smart Grid.

/pdf/smart-grid-the-new-and-improved-power-grid-a-survey-4jypbbgv3j.pdf

Smart Grid — The New and Improved Power Grid: A Survey

Smart Grid - The New and Improved Power Grid:

An expert system called SEEKS (security evaluation knowledge-based system) has been developed to apply operational planning knowledge to the static security assessment problem. SEEKS performs the contingency selection and remedial action tasks usually conducted offline by an operational planner. Rule-based techniques are used to select contingencies expected to cause steady state bus voltage violations. The selection identifies the most severe outage in each local area such that power flows need only be run for these worst-case outages. Both single and double branch outages are considered. Double outages are defined as scenarios where two branches share one bus, thereby providing coverage for events where a single branch outage evolves into a double branch outage due to protection, relay, or breaker failure. This system has been tested using the IEEE 30-bus system and a Los Angeles Department of Water and Power 182-bus system. >

Voltage security assessment using generalized operational planning knowledge

Malware authors attempt in an endless effort to find new methods to evade the malware detection engines. A popular method is the use of obfuscation technologies that change the syntax of malicious code while preserving the execution semantics. This leads to the evasion of signatures that are built based on the code syntax. In this paper, we propose a novel approach to develop an evasion-resistant malware signature. This signature is based on the malware's execution profiles extracted from kernel data structure objects and neither uses malicious code syntax specific information code execution flow information. Thus, proposed signature is more resistant to obfuscation methods and resilient in detecting malicious code variants. To evaluate the effectiveness of the proposed approach, a prototype signature generation tool called SigGENE is developed. The effectiveness of signatures generated by SigGENE evaluated using an experimental root kit-simulation tool that employs techniques commonly found in rootkits. This simulationtool is obfuscated using several different methods. In further experiments, real-world malware samples that have different variants with the same behavior used to verify the real-world applicability of the approach. The experiments show that the proposed approach is effective, not only in generating a signature that detects the malware and its variants and defeats different obfuscation methods, but also, in producing an execution profiles that can be used to characterize different malicious attacks.

Evasion-resistant malware signature based on profiling kernel data structure objects

System restoration involves status assessment, optimization of generation capability and load pickup. The optimization problem needs to take complex constraints into consideration and, therefore, it is not practical to formulate the problem as one single optimization problem. The other critical consideration for the development of decision support tools is its generality, i.e., the tools should be portable from a system to another with minimal customization. This paper reports a practical methodology for construction of system restoration strategies. The strategy adopted by each power system differs, depending on system characteristics and policies. A new method based on the concept of “Generic Restoration Milestones (GRMs)” is proposed. A specific restoration strategy can be synthesized by a combination of GRMs based on the actual system conditions. The decision support tool is expected to reduce the restoration time, thereby improving system reliability. The proposed decision support tool has been validated with cases based on a simplified Western Electricity Coordinating Council (WECC) 200-Bus system and Hawaiian Electric Company system.

/pdf/computation-of-milestones-for-decision-support-during-system-258zxo9i8a.pdf

Computation of milestones for decision support during system restoration

The supervisory control and data acquisition (SCADA) network provides adversaries with an opportunity to perform coordinated cyber attacks on power system equipment as it presents an increased attack surface. Coordinated attacks, when smartly structured, can not only have severe physical impacts, but can also potentially nullify the effect of system redundancy and other defense mechanisms. This chapter proposes a vulnerability assessment framework to quantify risk due to intelligent coordinated attacks, where risk is defined as the product of probability of successful cyber intrusion and resulting power system impact. The cyber network is modeled using Stochastic Petri Nets and the steady-state probability of successful intrusion into a substation is obtained using this. The model employs a SCADA network with firewalls and password protection schemes. The impact on the power system is estimated by load unserved after a successful attack. The New England 39-bus system is used as a test model to run Optimal Power Flow (OPF) simulations to determine load unserved. We conduct experiments creating coordinated attacks from our attack template on the test system and evaluate the risk for every case. Our attack cases include combinations of generation units and transmission lines that form coordinated attack pairs. Our integrated risk evaluation studies provide a methodology to assess risk from different cyber network configurations and substation capabilities. Our studies identify scenarios, where generation capacity, cyber vulnerability, and the topology of the grid together could be used by attackers to cause significant power system impact.

Risk Analysis of Coordinated Cyber Attacks on Power Grid

Area partitioning that splits a power network into self-sufficient islands is an emergency control to stop the propagation of disturbances and avoid cascading failures. This paper provides a survey of the state-of-the-art of power system islanding techniques and proposes the application of a multilevel graph area partitioning algorithm that are applicable to very large power grids. As an emergency control, network partitioning enhances the capability of a power system to withstand extreme and vulnerable operating conditions. The proposed algorithm has been simulated on two test systems, one with 200 buses and the other with 22,000 buses. The results indicate that the proposed algorithm is highly efficient.

Chen-Ching Liu

Papers

Voltage security assessment using generalized operational planning knowledge

Evasion-resistant malware signature based on profiling kernel data structure objects

Computation of milestones for decision support during system restoration

Risk Analysis of Coordinated Cyber Attacks on Power Grid

Power system reconfiguration based on multilevel graph partitioning