Non-redundant data clustering

Data clustering is a popular approach for automatically finding classes, concepts, or groups of patterns. In practice, this discovery process should avoid redundancies with existing knowledge about class structures or groupings, and reveal novel, previously unknown aspects of the data. In order to deal with this problem, we present an extension of the information bottleneck framework, called coordinated conditional information bottleneck, which takes negative relevance information into account by maximizing a conditional mutual information score subject to constraints. Algorithmically, one can apply an alternating optimization scheme that can be used in conjunction with different types of numeric and non-numeric attributes. We discuss extensions of the technique to the tasks of semi-supervised classification and enumeration of successive non-redundant clusterings. We present experimental results for applications in text mining and computer vision.     

Erfolgsfaktoren bei der Einführung von Wissensmanagementsystemen in die Praxis

Neben der Gestaltung technischer Systeme zur Unterstützung des Wissensmanagements ist die Ausgestaltung organisatorischer Ma?nahmen gleichbedeutend wichtig.     

Affordances in the design of enactive systems

Enactive interfaces must incorporate intuitive activity that characterizes naturalistic perception. However, the manner in which information is presented is not more important than the contents: what information is presented. In this contribution, we address the contents of perception. We argue that people perceive affordances, that is, the possible actions that are available in any given situation. We further argue that enactive interfaces should be designed to optimize presentation of information about the possible actions that are available to a person using the enactive interface. The design of enactive interfaces might be guided by an extension of the theory of ecological interface design (Vicente in Hum Factors 44:62–78, 2002) to include multimodal information that is accessed through fast, intuitive exploratory movement. We review two empirical studies that illustrate our arguments. Careful analysis of affordances, together with our increasing understanding of the enactive perception of affordances, should influence the design of enactive interfaces.     

Components of genetic variance and covariance for linear type traits in Jersey cattle

First available appraisals for 67,644 Jersey cows were used to estimate relative magnitudes of additive and selected sources of nonadditive genetic variance and covariance for 13 type traits scored linearly from 50 to 99 points. Covariances among four types of relatives were used to estimate components of genetic variance and covariance for each of the 13 traits. Direct additive, direct dominance, and additive maternal genetic variances and the direct-maternal additive genetic covariance were estimated using covariances among paternal half-sibs, full-sibs, daughters and dams, and maternal half-sibs. Estimates of additive genetic variance were 11 to 36% of variance within herd-appraiser subclasses with largest values for stature, strength, rump angle, rump width, and udder depth. Dominance components were two to four times the magnitude of additive components for dairy character, rump width, fore udder, rear udder height, rear udder width, and suspensory ligament but much smaller for other traits. Maternal effects and the direct-maternal additive genetic covariance appeared small for all linear type traits.     

Slabpose Columnsort: A New Oblivious Algorithm for Out-of-Core Sorting on Distributed-Memory Clusters

Our goal is to develop a robust out-of-core sorting program for a distributed-memory cluster. The literature contains two dominant paradigms for out-of-core sorting algorithms: merging-based and partitioning-based. We explore a third paradigm, that of oblivious algorithms. Unlike the two dominant paradigms, oblivious algorithms do not depend on the input keys and therefore lead to predetermined I/O and communication patterns in an out-of-core setting. Predetermined I/O and communication patterns facilitate overlapping I/O, communication, and computation for efficient implementation. We have developed several out-of-core sorting programs using the paradigm of oblivious algorithms. Our baseline implementation, 3-pass columnsort, was based on Leighton's columnsort algorithm. Though efficient in terms of I/O and communication, 3-pass columnsort has a restriction on the maximum problem size. As our first effort toward relaxing this restriction, we developed two implementations: subblock columnsort and M-columnsort. Both of these implementations incur substantial performance costs: subblock columnsort performs additional disk I/O, and M-columnsort needs substantial amounts of extra communication and computation. In this paper we present slabpose columnsort, a new oblivious algorithm that we have designed explicitly for the out-of-core setting. Slabpose columnsort relaxes the problem-size restriction at no extra I/O or communication cost. Experimental evidence on a Beowulf cluster shows that unlike subblock columnsort and M-columnsort, slabpose columnsort runs almost as fast as 3-pass columnsort. To the best of our knowledge, our implementations are the first out-of-core multiprocessor sorting algorithms that make no assumptions about the keys and produce output that is perfectly load balanced and in the striped order assumed by the Parallel Disk Model.     

Mutagenicity of three agricultural soils

A chemical and biological testing protocol was employed to evaluate the mutagenic potential of the organic compounds extracted from three agricultural soils. The analytical procedures used included bioassays with Salmonella typhimurium and Aspergillus nidulans for the detection of point mutations and a gas chromatography/mass spectrometry/computer system to identify major organic constituents. The extracts of all three soils exhibited mutagenic response in the bioassays. At a dose level of 1000 micrograms per plate, the organic extract of the Bastrop clay induced 434 net revertants; while at the same dose level the Norwood sandy clay and the Sassafrass sandy loam induced 35 and 178 net revertants, respectively, in the Salmonella assay with metabolic activation. In the Aspergillus assay, the extract of the Norwood and Bastrop soils induced a positive response without metabolic activation; this effect was reduced or eliminated in the presence of metabolic activation. Chemical analysis identified a variety of initiators, promotors, inhibitors, and cocarcinogens; however, there were no mutagenic compounds identified in any of the soil extracts. The results of this combined testing protocol indicate that the agricultural soils tested had an inherent level of mutagenic activity, which was not detected by GC/MS analysis alone, and this activity may be related to the past history of agricultural practices, including biocide applications, fertilization, and cultivation.     

Using mechatronics to teach mechanical design and technical communication

It is commonly accepted that hands-on experiences increase both learning and enjoyment during coursework. Mechatronics projects provide both interesting and relevant hands-on experiences for a wide range of topics including design processes, basic mechatronics concepts, technical communication, and working in a group environment. ME2110: Creative Decisions and Design at Georgia Tech integrates mechatronics and technical communication into a sophomore level mechanical design class. This paper describes the course in detail, highlighting the course goals and layout, tools provided to the students, industry involvement, and the main challenges of administering such a course.