Stability and model selection in k-means clustering

Clustering stability methods are a family of widely used model selection techniques for data clustering. Their unifying theme is that an appropriate model should result in a clustering which is robust with respect to various kinds of perturbations. Despite their relative success, not much is known theoretically on why or when do they work, or even what kind of assumptions they make in choosing an ‘appropriate’ model. Moreover, recent theoretical work has shown that they might ‘break down’ for large enough samples. In this paper, we focus on the behavior of clustering stability using k-means clustering. Our main technical result is an exact characterization of the distribution to which suitably scaled measures of instability converge, based on a sample drawn from any distribution in ? ⁿ satisfying mild regularity conditions. From this, we can show that clustering stability does not ‘break down’ even for arbitrarily large samples, at least for the k-means framework. Moreover, it allows us to identify the factors which eventually determine the behavior of clustering stability. This leads to some basic observations about what kind of assumptions are made when using these methods. While often reasonable, these assumptions might also lead to unexpected consequences.     

Cooperative TSP

In this paper we introduce and study cooperative variants of the Traveling Salesperson Problem. In these problems a salesperson has to make deliveries to customers who are willing to help in the process. Customer cooperativeness may be manifested in several modes: they may assist by approaching the salesperson, by reselling the goods they purchased to other customers, or by doing both.Several objectives are of interest: minimizing the total distance traveled by all of the participants, minimizing the maximal distance traveled by a participant and minimizing the total time until all of the deliveries are made.All of the combinations of cooperation modes and objective functions are considered, both in weighted undirected graphs and in Euclidean space. We show that most of the problems have a constant approximation algorithm, many of the others admit a PTAS, and a few are solvable in polynomial time. On the intractability side we provide NP-hardness proofs and inapproximability factors, some of which are tight.     

Fuel blend fraction estimation for fuel-flexible combustion control: Uncertainty analysis

“Fuel-flexible combustion control” strategies require accurate fuel composition information. A physics-based fuel blend fraction estimator for lean-burn combustion is used to assess the impact of uncertain variables on the estimator accuracy. For biodiesel blends in a diesel engine, the strategy is shown to be minimally affected by biodiesel feedstock variations, and expected fuel flow and oxygen sensor errors. However, observed airflow estimate errors are expected to lead to large estimator errors. When applied to ethanol blends, the estimator error is lower due to a higher stoichiometric mixture fraction for ethanol, and inherent differences in typical lean-burn gasoline engine operating conditions.     

A review of different pinch techniques

Individuals often pinch with preferred pinch techniques based on their experience and familiarity. However, researchers believe that pinch force exertions are not comparable across different pinch techniques. Hence, it is possible to pinch objects with incompatible pinch techniques during manual handling. This increases risks of developing musculoskeletal disorders. The preceding problem substantiates the need for a comprehensive understanding of definitions, categorisations and capacities of different pinch techniques. Nevertheless, the knowledge on pinch techniques appears to be highly dispersed and not compiled to a level systematic enough for further research. Thus, this review aims to develop a consolidated theoretical groundwork on pinch techniques. It presents an anthology of terminologies, previous findings and concepts underpinning different pinch techniques. While much research is still required, it is believed that an improved understanding on pinch techniques is now within sight, due to, in part, a structured review of their definitions, classifications and capacities.     

Asymmetric Catalysis in Confined Space Provided by l-Proline Functionalized Mesoporous Silica with Plugs in the Pore

Mesoporous silica having l-proline functionality and plugged pore structure was synthesized by direct microwave synthesis under acidic condition using Pluronic P123 as template and sodium metasilicate nonahydrate as silica source. The direct synthesis resulted in the prismatic or disk types with short channels and plugs in the pore structure. Due to the presence of plugs, dual mesoporosities with 7 and 3 nm pore diameter were generated leading to an existence of cavitation. Tri-functionality of this materials i.e. chiral organofunctional group, plugs, and short channel, would be useful for the asymmetric diethyl malonate addition reaction and asymmetric epoxidation reaction of α,β-unsaturated aldehydes. We proved that the formation of plugs in the mesoporous silica governs the formation of dual mesoporosities providing an enhancement of enantiomeric excess in the asymmetric catalysis. This finding might open an important application of the amino-acid organo multifunctionalized mesoporous silica for asymmetric reactions.     

Combining color and depth for enhanced image segmentation and retargeting

As depth cameras become more popular, pixel depth information becomes easier to obtain. This information can clearly enhance many image processing applications. However, combining depth and color information is not straightforward as these two signals can have different noise characteristics, differences in resolution, and their boundaries do not generally agree. We present a technique that combines depth and color image information from real devices in synergy. In particular, we focus on combining them to improve image segmentation. We use color information to fill and clean depth and use depth to enhance color image segmentation. We demonstrate the utility of the combined segmentation for extracting layers and present a novel image retargeting algorithm for layered images.     

Signal amelioration of electrophoretically deposited whole-cell biosensors using external electric fields

This paper presents an integrated whole-cell biochip system where functioning cells are deposited on the solid micro-machined surfaces while specially designed indium tin oxide electrodes that can be used to apply controllable electric fields during various stages; for example during cell deposition. The electrodes can be used also for sensing currents associated with the sensing mechanisms of electrochemical whole-cell biosensors. In this work a new approach integrating live bacterial cells on a biochip using electrophoretic deposition is presented. The biomaterial deposition technique was characterized under various driving potentials and chamber configurations. An analytical model of the electrophoretic deposition kinetics was developed and presented here. The deposited biomass included genetically engineered bacterial cells that may respond to toxic material exposure by expressing proteins that react with specific analytes generating electrochemically active byproducts. In this study the effect of external electric fields on the whole-cell biochips has been successfully developed and tested. The research hypothesis was that by applying electric fields on bacterial whole-cells, their permeability to the penetration of external analytes can be increased. This effect was tested and the results are shown here. The effect of prolonged and short external electric fields on the bioelectrochemical signal generated by sessile bacterial whole-cells in response to the presence of toxins was studied. It was demonstrated that relatively short 10 ms external DC electric pulse improves the performance of bacterial biosensors by 15% relative to un-biased biosensors. The application of prolonged 1 h external alternating electric fields deteriorated the whole-cell performance in the presence of toxins. In this paper we present the electrode apparatus and methods, as well as the characterization results, e.g. signal vs. time and induction factor, of such chips and discussing the highlight and problems of this new concept.     

Enhanced removal of humic acid from water by micelle-montmorillonite composites: Comparison to granulated activated carbon

Removal of humic acid (HA) from water by octadecyltrimethyl-ammonium (ODTMA) micelle–montmorillonite (MMT) composites and by granulated activated carbon (GAC) was studied in dispersion and by filtration. FTIR measurements emphasized that the ODTMA micelle–clay-mineral composite differs from the ODTMA monomer-clay-mineral one. HA adsorption by GAC in dispersion was moderately reduced from 100% initially to 75% with an increase in HA concentrations. In contrast, the fractions adsorbed by the composite increased monotonically with HA concentrations from very low adsorption up to 80%. This effect was explained by the presence of a small concentration of ODTMA monomers in dispersion, which adsorbed on the negatively charged HA and partially neutralized it, or caused charge reversal, which in turn inhibited HA adsorption by the positively charged composite. This effect was verified by mobility measurements and was found to be more pronounced at low HA concentrations. This trend of increase in HA removal with an increase in HA concentration (3 and 10 ppm) was also observed in filtration (50% and 85% removal, respectively), when employing columns of the composite mixed with sand (to improve flow). Throughout the experiment the removal of high HA concentration (10 ppm) by the composite filter remained ~ 80% whereas, the removal by the GAC filter decreased from an initial 80% removal to a complete recovery of the HA (no removal) after the passage of 20 L (200 pore volumes). Upon applying sand of higher quality in the columns the removal of low HA concentrations was nearly complete by the composite column while the GAC only slightly contributed to its removal (for a passage of 100 L (1000 pore volumes)).     

M-learning in review： Technology, standard and evaluation

Nowadays E-learning as a form of learning depending on networks and computer devices has been expanding by leaps and bounds. As a special kind of E-learning, M-learning （mobile learning） aims at the use of mobile devices anywhere at anytime by anybody. In this paper, the technologies to develop a M-learning system are deeply analyzed according to the layer model at first. Then, from the view of application, the standardization in M-learning is researched in order to accelerate the development and popularization of M-learning. Finally, quality estimation of M-learning system is discussed from the view of QoE （quality of experiences）. The experience form end-user is the sole effective norm to judge the result of M-learning technology. It is no doubt that technologies, standardization and evaluation will play very important parts in the course of M-leaming development.