Visualizing “typical” and “exotic” Internet traffic data

The threat of cyber attacks motivates the need to monitor Internet traffic data for potentially abnormal behavior. Due to the enormous volumes of such data, statistical process monitoring tools, such as those traditionally used on data in the product manufacturing arena, are inadequate. “Exotic” data may indicate a potential attack; detecting such data requires a characterization of “typical” data. We devise some new graphical displays, including a “skyline plot,” that permit ready visual identification of unusual Internet traffic patterns in “streaming” data, and use appropriate statistical measures to help identify potential cyberattacks. These methods are illustrated on a moderate-sized data set (135,605 records) collected at George Mason University.     

Skill based manufacturing vs. “unmanned factory”—which is superior?

Past development of small batch production has led to advanced horizontal and vertical division of labor, job shop manufacturing, and the use of NC machines. This has caused serious difficulties such as long and variant throughput times, an unfavorable ratio of unproductive to productive workers, and low quality of work. Additionally confronted with deeply changed market conditions, the management is split over the right answer to this challenge. Most managers and production planners try to replace human work still further by enforced use of computers on the shop floor and in the technical office in an integrated manner (CIM). Since this strategy is very likely to create new severe problems, the minority seeks to avoid them by reorganizing production and rearranging the division of functions between humans and machines profoundly (skill based manufacturing). This position is outlined in this paper in some detail and examples of realized subsystems, as first steps on this development path, are included. Finally, reasons for its superiority and resistance to implement it are discussed.     

“Paraprogramming” manufacturing information systems

Computer integrated manufacturing (CIM) is creating unexpected problems for a growing number of manufacturing companies. Manufacturers are finding it especially difficult to attract programmers who are both willing and able to develop the highly complex software that integrates existing accounting, sales, production, engineering, and quality control information subsystems. Consequently, many companies abdicate their responsibility for manufacturing information systems and seek third party support ranging from consulting assistance to a total takeover of the company's information resources and operations. Companies that “give away” their internal information system capabilities to third parties will ultimately lose control of their enterprise information, a danger to be avoided. Off-the-shelf software for desktop computers has become sufficiently powerful to help solve a major portion of this serious problem. We hypothesize that manufacturing engineers (and others) can be trained to use packaged software to leverage their company's systems programming capabilities. In effect they would become “paraprogrammers” who would help design, develop, and maintain manufacturing information systems. This new type of professional would not require a computer science or similar educational background, but could be trained to satisfy many specialized programming needs in a manner similar to how paramedics and paralegals are trained and used in the medical and legal professions, respectively. This paper reports on the early stages of research to determine whether or not product design engineers can use a desktop relational database management system and its various command languages to develop a master bill of material information system (BOMIS). The purpose of the research is to evaluate the amount of programming complexity reduction and increased operational effectiveness that can be achieved through paraprogramming by manufacturing engineers.     

Modelling rhetorical legal “logic”—a double syllogism

This paper looks at legal reasoning from the point of view of the work of the lawyer, rather than the law itself. In the case of Common Law systems, this means a more flexible view of how tasks are divided between the humans and the computer system, with an emphasis on decision support rather than complete automation. A process-based model of the lawyer's work is proposed in the form of a double syllogism, which displays an aesthetically pleasing symmetry, but also a significant asymmetry in the role played by perceived precedents. This arises from the use of inductive, rather than deductive, reasoning. The potential complications arising from the issue of the perception of precedents are discussed in depth.The double-syllogism model is then considered in the light of CommonKADS terminology and models. It is suggested that decision support systems using knowledge-based techniques, as required to support lawyers working under Common Law jurisdiction, raise a stronger form of the interaction problem that is well known in knowledge-based systems. This means that such systems are not well catered for in the existing CommonKADS Organisational, Agent, Task and Communication Models. The double-syllogism model is suggested as a supplement to CommonKADS in the development of such systems, at least until a more generic addition is available.     

The depth of monotone Boolean functions with multi input “AND” and “OR” gates

The problem of maximum depth of monotone Boolean functions is investigated when k-input AND and OR gates are used for realisation.     

“Probability transforms” of digital pictures

Let g be any local property (e.g., gray level or gradient magnitude) defined on a digital picture. Let p_g(z) be the relative frequency with which g has value z. At each point (x,y) of the picture we can display p_g[g(x,y)], appropriately scaled; the result is called the p_g transform of the picture. Alternatively, we can use joint or conditional frequencies of pairs of local properties to define transforms. This note gives examples of such transforms for various gs and discusses their possible uses and limitations.     

“BILDLIB:” the image analysis software at IPA

The image analysis software for visual inspection purposes developed at IPA, the Fraunhofer-Institute for Manufacturing Engineering and Automation, is described. Its basic mathematical framework, its organization scheme, some of its contents, as well as some examples are given. Current research and future developments are discussed.     

Information retrieval expert system “matchmaker”

This paper describes the information system used to search for a potential matrimonial partner. The search is based on comparison of the subject's record, which consists of his/her answers to about 400 items of a specially designed questionnaire, to the records of the potential partners. The basic principle of the system is representation of the set of candidates for the client with psychological warnings about potential “conflict zones” in relationships between client and candidate rather than a ranking of candidates based on hypothetical “psychological compatibility” indices.     

Decision support for “messy” problems

Problems that are non-quantitative and not bound to a narrow knowledge domain have been served unsatisfactorily by decision support and expert systems. Alternative techniques that address this type of problem are explained here using two key concepts: problem type dependent process support and domain related knowledge. Process support refers to the program steps and the data items useful in finding the solution. Domain related knowledge is knowledge drawn from a specific domain, yet through abstraction applicable to a wider range of problems. Results of preliminary empirical analyses suggest that both concepts are useful.