Technical University of Dortmund

About: Technical University of Dortmund is a education organization based out in Dortmund, Nordrhein-Westfalen, Germany. It is known for research contribution in the topics: Large Hadron Collider & Neutrino. The organization has 13028 authors who have published 27666 publications receiving 615557 citations. The organization is also known as: Dortmund University & University of Dortmund.

Experimental-study of opposite-sign dimuons produced in neutrino and antineutrino interactions

Abstract: A large sample of opposite-sign dimuons, produced by the interaction of neutrinos and antineutrinos in iron, is analysed. The data agree very well with the hypothesis that the extra muon is the product of charm decay. They yield information on the strength and space-time structure of the charmproducing weak current. The strange-sea structure functionxs(x) is determined. The difference between neutrino and antineutrino dimuon production is analysed to provide a value of the Kobayashi-Maskawa weak mixing angleθ2.

Exciton dephasing via phonon interactions in InAs quantum dots: dependence on quantum confinement

Abstract: We report systematic measurements of the dephasing of the excitonic ground-state transition in a series of InGaAs?GaAs quantum dots having different quantum confinement potentials. Using a highly sensitive four-wave mixing technique, we measure the polarization decay in the temperature range from 5 to 120 K on nine samples having the energy distance from the dot ground-state transition to the wetting layer continuum (confinement energy) tuned from 332 to 69 meV by thermal annealing. The width and the weight of the zero-phonon line in the homogeneous line shape are inferred from the measured polarization decay and are discussed within the framework of recent theoretical models of the exciton-acoustic phonon interaction in quantum dots. The weight of the zero-phonon line is found to decrease with increasing lattice temperature and confinement energy, consistently with theoretical predictions by the independent Boson model. The temperature-dependent width of the zero-phonon line is well reproduced by a thermally activated behavior having two constant activation energies of 6 and 28 meV, independent of confinement energy. Only the coefficient to the 6-meV activation energy shows a systematic increase with increasing confinement energy. These findings rule out that the process of one-phonon absorption from the excitonic ground state into higher energy states is the underlying dephasing mechanism.

The TTT Algorithm: A Redundancy-Free Approach to Active Automata Learning

Abstract: In this paper we present TTT, a novel active automata learning algorithm formulated in the Minimally Adequate Teacher (MAT) framework. The distinguishing characteristic of TTT is its redundancy-free organization of observations, which can be exploited to achieve optimal (linear) space complexity. This is thanks to a thorough analysis of counterexamples, extracting and storing only the essential refining information. TTT is therefore particularly well-suited for application in a runtime verification context, where counterexamples (obtained, e.g., via monitoring) may be excessively long: as the execution time of a test sequence typically grows with its length, this would otherwise cause severe performance degradation. We illustrate the impact of TTT’s consequent redundancy-free approach along a number of examples.

Robust trainability of single neurons

Abstract: We investigate the problem of learning concepts by presenting labeled and randomly chosen training–examples to single neurons. It is well-known that linear halfspaces are learnable by the method of linear programming. The corresponding (Mc-Culloch-Pitts) neurons are therefore efficiently trainable to learn an unknown halfspace from examples. We want to analyze how fast the learning performance degrades when the representational power of the neuron is overstrained, i.e., if more complex concepts than just halfspaces are allowed. We show that a neuron cannot efficently find its probably almost optimal adjustment (unless RP = NP). If the weights and the threshold of the neuron have a fixed constant bound on their coding length, the situation is even worse: There is in general no polynomial time training method which bounds the resulting prediction error of the neuron by k.opt for a fixed constant k (unless RP = NP). Other variants of learning more complex concepts than halfspaces by single neurons are also investigated. We show that neither heuristical learning nor learning by sigmoidal neurons with a constant reject-rate is efficiently possible (unless RP = NP).