PAC-Bayesian Compression Bounds on the Prediction Error of Learning Algorithms for Classification

We consider bounds on the prediction error of classification algorithms based on sample compression. We refine the notion of a compression scheme to distinguish permutation and repetition invariant and non-permutation and repetition invariant compression schemes leading to different prediction error bounds. Also, we extend known results on compression to the case of non-zero empirical risk.We provide bounds on the prediction error of classifiers returned by mistake-driven online learning algorithms by interpreting mistake bounds as bounds on the size of the respective compression scheme of the algorithm. This leads to a bound on the prediction error of perceptron solutions that depends on the margin a support vector machine would achieve on the same training sample.Furthermore, using the property of compression we derive bounds on the average prediction error of kernel classifiers in the PAC-Bayesian framework. These bounds assume a prior measure over the expansion coefficients in the data-dependent kernel expansion and bound the average prediction error uniformly over subsets of the space of expansion coefficients.Editor Shai Ben-David     

The fast and the slow of skilled bimanual rhythm production: Parallel versus integrated timing.

Professional pianists performed 2 bimanual rhythms at a wide range of different tempos. The polyrhythmic task required the combination of 2 isochronous sequences (3 against 4) between the hands; in the syncopated rhythm task successive keystrokes formed intervals of identical (isochronous) durations. At slower tempos, pianists relied on integrated timing control merging successive intervals between the hands into a common reference frame. A timer–motor model is proposed based on the concepts of rate fluctuation and the distinction between target specification and timekeeper execution processes as a quantitative account of performance at slow tempos. At rapid rates expert pianists used hand-independent, parallel timing control. In alternative to a model based on a single central clock, findings support a model of flexible control structures with multiple timekeepers that can work in parallel to accommodate specific task constraints. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Does age make a difference? Predicting physical activity of South Koreans.

Social cognition models of health behavior are commonly understood as being universal, which implies that they are applicable to groups varying in age or cultural background, for example. Cultural uniqueness and characteristics of life-span development, however, necessitate the study of differential effects. Accordingly, the health action process approach (HAPA) was examined in younger and middle-aged/older adults from South Korea (N = 697) who participated in a longitudinal health screening study with a 6-month time lag. The HAPA model had a good fit within the middle-aged/older adult sample. Physical activity was predicted by planning, coping self-efficacy, and intention, which were, in turn, predicted by action self-efficacy, outcome expectancies, and risk perceptions. Conversely, the results indicated a poor model fit in the younger adult sample. The results suggest a different motivation for the involvement in physical activity as a function of age. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Prological Language Processing

We describe a Prolog-based approach to the development of language processors (such as preprocessors, frontends, evaluators, tools for software modification and analysis). The design of the corresponding environment Laptob for prological language processing is outlined. Language processor definitions in Laptob are basically Prolog programs. The programs might contain grammars, that is, we consider logic grammars. The programs can be typed, and they can be higher-order. The adaptation and composition of the logic programs themselves is supported by meta-programming. The environment offers tool support for efficient scanning, testing, and application development based on a make-system. We report on recent and ongoing applications of the Prolog-based approach.     

Semantic-Web-Technologien in der Arbeitsplatzvermittlung

Zusammenfassung The importance of the Internet for job procurement is increasing for the reason that three quarters of the people in the employment age are online. On the other hand because ever more companies are publishing their job offers on the Web. However, due to the large number of openings published online it is almost impossible for job seekers and job portals to gain an overview of the entire employment market. Since job offers lack semantically meaningful annotations, the search and integration into databases are made highly difficult. Applying Semantic Web technologies to the e-recruitment process provides advantages for all participants in the market. In this paper we describe a method for analysing the domain-specific language of an application domain. We use this method to describe the e-recruitment process and the necessary ontologies for annotating job offers and job applications. In conclusion, we present the prototypical implementation of the scenario based on Semantic Web, especially semantic matching.     

House Prices as a Result of Trading Activities: A Patient Trader Model

Computational Economics - We present a new modeling approach for house price movements as a consequence of the trading behavior of market agents. In our modeling approach, all agents are assumed to...     

Solving a linear multiperiod portfolio problem by interior-point methodology

In this report we formulate a linear multiperiod programming problem and show how it can be solved by a new interior point algorithm. The conditions of convergence and applicability of the algorithm of centers are explicitly connected to the multiperiod programming problem. For the former, effective application is supported by a set of simplifying conditions stated and proved in the text. For the latter, boundedness and nontriviality under real world conditions is demonstrated, allowing for its solution by the interior point algorithm. The use of a fast interior point algorithm is motivated by some empirical evidence from a Revised Simplex optimizer.     

A strong winner-take-all model based on integration of potential differences

A dynamic, strong winner-take-all model is presented. It is based on a limited integrator that is fed by potential differences between feedback from a single interneuron and the input value. The model does not require a special nonlinear function in the forward path. In equilibrium state, a single winner neuron transmits its input value to the output without change. For a two-neuron system the behaviour is shown graphically. Equilibrium states are determined for different types of input vectors.     

混凝土搅拌站和混凝土制品厂的污水处理

当你清洗混凝土搅拌车时 ,会产生大量的混凝土残留。回用这些砂子、石子和水泥浆水非常经济而又有利于环保。德国克莱斐克斯 (ＫＬＡＥＲＦＩＸ)公司经验成熟的回收系统 ,可以将清洗混凝土砂石后残余的带活性的水利用起来。这些经过标准化和模块化设计的系统 ,是针对各个不同客户的需要加以定制的。这份报告将告诉你在混凝土搅拌站和混凝土制品厂 ,将剩余残留的混凝土骨料完全回收和回收过程中将带活性的污水重新回收利用的过程。当人们开始对砂石料工厂清洗时产生的那些不带活性的污水进行处理的同时 ,人们对混凝土工厂、搅拌站中产生的含…     

Sanierungsforschung in regional kontaminierten Aquiferen (SAFIRA) – 1. Information zum Forschungsschwerpunkt am Standort Bitterfeld

The SAFIRA-project (remediation research in regionally contaminated aquifers) focuses on innovative in situ remediation technologies to treat complex groundwater contamination. For this purpose, a model site located in Bitterfeld, Germany, was selected. The site is heavily contaminated with aliphatic and aromatic chlorinated hydrocarbons. The goal of the project is to develop, test and optimize new technologies of reactive barriers under controlled in situ conditions in a pilot plant. In order to create the necessary basis for this model project, the hydrogeological and hydrogeochemical conditions at the selected field site were investigated and various technologies for the design of the in situ pilot plant were examinated. Small-scale laboratory studies and tests in a mobile test unit (scale 1 m) were carried out in order to test the technologies prior to their application in the field. The technologies now tested in the reactive columns at the pilot plant include physico-chemical, microbial and combined techniques.