Training echo state networks for rotation-invariant bone marrow cell classification

The main principle of diagnostic pathology is the reliable interpretation of individual cells in context of the tissue architecture. Especially a confident examination of bone marrow specimen is dependent on a valid classification of myeloid cells. In this work, we propose a novel rotation-invariant learning scheme for multi-class echo state networks (ESNs), which achieves very high performance in automated bone marrow cell classification. Based on representing static images as temporal sequence of rotations, we show how ESNs robustly recognize cells of arbitrary rotations by taking advantage of their short-term memory capacity. The performance of our approach is compared to a classification random forest that learns rotation-invariance in a conventional way by exhaustively training on multiple rotations of individual samples. The methods were evaluated on a human bone marrow image database consisting of granulopoietic and erythropoietic cells in different maturation stages. Our ESN approach to cell classification does not rely on segmentation of cells or manual feature extraction and can therefore directly be applied to image data.

     

Das Wirken von Em.O.Univ.-Prof. Dipl.-Ing. Dr. mont. Hans J?rg Steiner aus Sicht eines seiner Schüler und des Nachfolgers

On the occasion of his 75th birthday, this publication pursues the goals to approach the personality Hans J?rg Steiner from the perspective of one of his students and to acknowledge his work from his successor’s point of view. The author regretfully, but with composure recognizes that the portrait of this great Austrian mineral processing engineer as drawn by him can only be fragmentary, and invites the gentle reader to do the same.     

A socially aware caching mechanism for encounter networks

Managing community content in mobile communities is challenging, since informal groups may emerge spontaneously whenever opportunities exist. To deal with group dynamics, encounter based community support systems can be a solution. In encounter networks, peers exchange content whenever they physically meet, i.e., whenever peers are near to each other. To assure content availability within the community network, content needs to be replicated (cached) and continuously updated. A?well known drawback of this scheme is the fact that the resource usage can be quite high. In this paper we present new caching strategies based on sociological knowledge, aiming at improving the overall content quality. We derive mobility and connectivity aware cache policies and simulate their benefit for knowledge building with urban life simulations. Results show that context precision and recall can be greatly improved in comparison to traditional cache strategies applied in the majority of encounter networks.     

Single‐step synthesis of a radically polymerizable 1,1′‐bi‐2‐naphthol derivative for asymmetric catalysis

A new polymerizable 1,1′‐bi‐2‐naphthol derivative for polymer‐supported catalytic asymmetric synthesis is presented. The synthesis is conducted within a single reaction step, which is a major advantage over other approaches presented in the literature. The ligand‐bearing polymer is prepared through copolymerization with N‐isopropylacrylamide. Preliminary experiments on the utility in catalytic asymmetric alkylation reactions reveal the accessibility and activity of the polymer‐attached catalysts. The stereoselectivity of the reaction is found to be somewhat lower than for reactions performed in the presence of free 1,1′‐bi‐2‐naphthol, and thus requires further optimization. The enantiomeric excess of the reaction products was determined via ¹H NMR spectroscopy after chiral derivatization with (R)‐α‐methylbenzyl isocyanate. © 2015 Society of Chemical Industry     

Finite element analyses of slope stability problems using non-associated plasticity

In recent years, finite element analyses have increasingly been utilized for slope stability problems. In comparison to limit equilibrium methods, numerical analyses do not require any definition of the failure mechanism a priori and enable the determination of the safety level more accurately. The paper compares the performances of strength reduction finite element analysis (SRFEA) with finite element limit analysis (FELA), whereby the focus is related to non-associated plasticity. Displacement-based finite element analyses using a strength reduction technique suffer from numerical instabilities when using non-associated plasticity, especially when dealing with high friction angles but moderate dilatancy angles. The FELA on the other hand provides rigorous upper and lower bounds of the factor of safety (FoS) but is restricted to associated flow rules. Suggestions to overcome this problem, proposed by Davis (1968), lead to conservative FoSs; therefore, an enhanced procedure has been investigated. When using the modified approach, both the SRFEA and the FELA provide very similar results. Further studies highlight the advantages of using an adaptive mesh refinement to determine FoSs. Additionally, it is shown that the initial stress field does not affect the FoS when using a Mohr-Coulomb failure criterion.     

Manuelle Kommissioniersysteme und die Rolle des Menschen

Picking processes have gained importance in recent years because of increasing online trade, increasing pressure on lead times in production, procurement and distribution of goods, and the introduction of just-in-time concepts. Typically, there is a trade-off between throughput times, quality, and cost. The article examines the contribution of alternative manual picking systems to achieve these objectives, and in particular looks at the role of humans in the process.