The 'moving targets' training algorithm

A simple method for training the dynamical behavior of a neural network is derived. It is applicable to any training problem in discrete-time networks with arbitrary feedback. The algorithm resembles back-propagation in that an error function is minimized using a gradient-based method, but the optimization is carried out in the hidden part of state space either instead of, or in addition to weight space. Computational results are presented for some simple dynamical training problems, one of which requires response to a signal 100 time steps in the past.

Sustained delivery by leucine-modified chitosan spray-dried respirable powders

The controlled co-delivery of multiple agents to the lung offers potential benefits to patients. This study investigated the preparation and characterisation of highly respirable spray-dried powders displaying the sustained release of two chemically distinct therapeutic agents. Spray-dried powders were produced from 30% (v/v) aqueous ethanol formulations that contained hydrophilic (terbutaline sulphate) and hydrophobic (beclometasone dipropionate) model drugs, chitosan (as a drug release modifier) and leucine (aerosolisation enhancer). The influence of chitosan molecular weight on spray-drying thermal efficiency, aerosol performance and drug release profile was investigated. Resultant powders were physically characterised: with in vitro aerosolisation performance and drug release profile investigated by the Multi-Stage Liquid Impinger and modified USP II dissolution apparatus, respectively. It was found that increased chitosan molecular weight gave increased spray-drying thermal efficiency. The powders generated were of a suitable size for inhalation—with emitted doses over 90% and fine particle fractions up to 72% of the loaded dose. Sustained drug release profiles were observed in dissolution tests for both agents: increased chitosan molecular weight associated with increased duration of drug release. The controlled co-delivery of hydrophilic and hydrophobic entities underlines the capability of spray drying to produce respirable particles with sustained release for delivery to the lung. (c) 2009 Elsevier B.V. All rights reserved.

"Minimálbér-paradoxon" versenyzői munkaerőpiacon? ("Minimum-wage paradox" in a competitive labour market? Lessons from an experimental line of thought)

A versenyzői munkaerőpiac hagyományos kereslet-kínálati modellje az egyensúlyi bérszintet meghaladó minimálbér következményeként az egyensúlyi bérszint mellettinél alacsonyabb foglalkoztatást jósol; minél magasabb a minimálbér, annál alacsonyabbat. Empirikus vizsgálatok szerint ugyanakkor a minimálbér-emelés nem feltétlenül csökkenti a foglalkoztatást - ezt nevezik minimálbér-paradoxonnak -, ami legkézenfekvőbben a munkáltatók munkaerő-piaci monopszonerejével látszik magyarázhatónak. Ezzel szemben az a gondolatkísérlet, amelyről ez a cikk beszámol, általánosabb érvényű, versenyzői munkaerőpiacot feltételező magyarázat kidolgozására irányul. / === / In the conventional textbook demand/supply model of competitive labour markets, the introduction of a minimum wage above market-clearing level must reduce employment. Empirical findings suggest, however, that this may not always be the case, a finding most readily explained by monopsonistic competition in the labour market. The experimental line of thought reported here explores an alternative root, interpreting the "minimum-wage paradox" as the outcome of a competitive labour market that displays friction.

Methane and sulphate consumption in different habitats of the Håkon Mosby Mud Volcano (HMMV)

The Hakon Mosby Mud Volcano is a highly active methane seep hosting different chemosynthetic communities such as thiotrophic bacterial mats and siboglinid tubeworm assemblages. This study focuses on in situ measurements of methane fluxes to and from these different habitats, in comparison to benthic methane and oxygen consumption rates. By quantifying in situ oxygen, methane, and sulfide fluxes in different habitats, a spatial budget covering areas of 10-1000 -m diameter was established. The range of dissolved methane efflux (770-2 mmol m-2 d-1) from the center to the outer rim was associated with a decrease in temperature gradients from 46°C to < 1°C m-1, indicating that spatial variations in fluid flow control the distribution of benthic habitats and activities. Accordingly, total oxygen uptake (TOU) varied between the different habitats by one order of magnitude from 15 mmol m-2 d-1 to 161 mmol m-2 d-1. High fluid flow rates appeared to suppress benthic activities by limiting the availability of electron acceptors. Accordingly, the highest TOU was associated with the lowest fluid flow and methane efflux. This was most likely due to the aerobic oxidation of methane, which may be more relevant as a sink for methane as previously considered in submarine ecosystems.

Relative abundance of prokaryotes in sediments of the Håkon Mosby mud volcano (Z1-Z3, all habitats)

CARD-FISH was performed as previously described in Ruff et al., (2013; doi:10.1371/journal.pone.0072627) with the following modifications. 4-6 µl of 25-fold diluted sediment were used for filtration. Archaeal cell walls were permeabilized with 0.1M HCl for 2 min to detect ANME-3 cells, or Proteinase K solution (15 µg ml-1 (Merck, Darmstadt, Germany) in 0.05 M EDTA (pH 8), 0.1 M Tris-HCl (pH 8), 0.5 M NaCl) for 2-4 min at room temperature for all other archaea. Bacterial cell walls were permeabilized with lysozyme solution (1000kU/ml) for 60 min at 37°. Cells were stained with DAPI (1µg/ml), embedded in mounting medium and counted in 40-60 independent microscopic fields using an Axiophot II epifluorescence microscope (Carl Zeiss, Jena, Germany).

Multiple-retrieval case-based reasoning for course timetabling problems

The structured representation of cases by attribute graphs in a Case-Based Reasoning (CBR) system for course timetabling has been the subject of previous research by the authors. In that system, the case base is organised as a decision tree and the retrieval process chooses those cases which are sub attribute graph isomorphic to the new case. The drawback of that approach is that it is not suitable for solving large problems. This paper presents a multiple-retrieval approach that partitions a large problem into small solvable sub-problems by recursively inputting the unsolved part of the graph into the decision tree for retrieval. The adaptation combines the retrieved partial solutions of all the partitioned sub-problems and employs a graph heuristic method to construct the whole solution for the new case. We present a methodology which is not dependant upon problem specific information and which, as such, represents an approach which underpins the goal of building more general timetabling systems. We also explore the question of whether this multiple-retrieval CBR could be an effective initialisation method for local search methods such as Hill Climbing, Tabu Search and Simulated Annealing. Significant results are obtained from a wide range of experiments. An evaluation of the CBR system is presented and the impact of the approach on timetabling research is discussed. We see that the approach does indeed represent an effective initialisation method for these approaches.