Mathematical models of classes of objects can significantly contribute to the analysis of digital images. A major problem in modelling is to establish suitable descriptions that cover not only a single object but also the variation that is usually present within a class of objects. The objective of this thesis is to develop more general modelling strategies than commonly used today. In particular, the impact of the human factor in the model creation process should be minimised. It is presumed that the human ability of abstraction imposes undesired constraints on the description. In comparison, common approaches are discussed from the viewpoint of generality. The technique considered introduces appearance space as a common framework to represent both shapes and images. In appearance space, an object is represented by a single point in a high-dimensional vector space. Accordingly, objects subject to variation appear as nonlinear manifolds in appearance space. These manifolds are often characterised by only a few intrinsic dimensions. A model of a class of objects is therefore considered equal to the mathematical description of this manifold. The presence of nonlinearity motivates the use of artificial auto-associative neural networks in the modelling process. The network extracts nonlinear modes of variation from a number of training examples. The procedure is evaluated on both synthetic and natural data of shapes and images and shows promising results as a general approach to object modelling.

Typische assoziative Verdicker für wässrige Systeme basieren auf linearen, doppelt hydrophob endmodifizierten Poly(ethylenglykolen) (PEGs). Diese Polymere aggregieren aufgrund ihrer Struktur in wässriger Lösung und bilden ein Netzwerk aus verbrückten Polymer-Mizellen. Dabei kann ein Polymer-Molekül maximal zwei Mizellen miteinander verbinden. Bisher ist unklar, wie die Anzahl der Endgruppen eines verzweigten, mehrfach hydrophob endmodifizierten hydrophilen Polymers die Struktur und Dynamik solcher Netzwerke beeinflusst. Die Synthese verzweigter Polymere auf PEG-Basis erfolgt mittels lebender ionischer Polymerisation und ist experimentell aufwändig. Das Einführen hydrophober Endgruppen erfordert zusätzliche Synthese-Schritte. In dieser Arbeit wurden hydrophile Sternpolymere mit hydrophoben Endgruppen in einem Schritt hergestellt. Dazu wurde die Technik der radikalischen Polymerisation unter Kettenübertragung durch reversible Addition und anschließende Fragmentierung (reversible addition-fragmentation chain transfer, RAFT) genutzt. Die Synthese der Sternpolymere erfolgte von einem multifunktionalen Kern, der die R-Gruppe der RAFT-Kettenüberträger (chain transfer agents, CTAs) bildete. Die dazu benötigten CTAs wurden so konzipiert, dass mit ihrer Hilfe sowohl die Anzahl der Arme des Sternpolymers (von 2 bis 4), als auch die Länge der hydrophoben Endgruppe (C4, C12, C18) variiert werden konnte. Der große Vorteil der RAFT-Polymerisation ist, dass sie viele polare Monomere für die Synthese der hydrophilen Arme des Sternpolymers toleriert. In dieser Arbeit wurden als Modell-Monomere Oligo(ethylenglykol)methylether-acrylat (OEGA) und N,N-Dimethylacrylamid (DMA) eingesetzt. Beide Monomere bilden nicht-ionische hydrophile Polymere. Poly(OEGA) ist ein Kammpolymer, das auf PEG basiert. Poly(DMA) besitzt dagegen eine deutlich kompaktere Struktur. Die erhaltenen amphiphilen Sternpolymere wurden umfassend molekular charakterisiert. Die Molmassen wurden mit verschiedenen GPC-Systemen bestimmt und der Grad der Endgruppenfunktionalisierung wurde mittels UV/Vis- und 1H-NMR-Spektroskopie überprüft. Die Polymerisation von OEGA zeigt mit den CTAs einige Charakteristika der Polymerisation mit reversibler Deaktivierung (RDRP, auch „kontrollierte radikalische Polymerisation“), wird aber durch Kettenübertragung zum Monomer bzw. Polymer gestört. Diese Nebenreaktion ist auf die Struktur des Monomers als Oligoether zurückzuführen. Bei allen untersuchten Polymerisationen von DMA mit den multifunktionalen CTAs steigt die Molmasse linear mit dem Umsatz. Die erhaltenen Polymere zeigen durchweg monomodale und enge Molmassenverteilungen (PDI ≤ 1,2). Die Molmassen lassen sich in einem weiten Bereich von 25 kg/mol bis 150 kg/mol einstellen und die Endgruppen der Polymere bleiben zu 90 % erhalten. Während die Polymerisation von DMA sowohl mit den di- als auch den trifunktionalen CTAs innerhalb von 3 h zu quantitativen Umsätzen verläuft, wird der quantitative Umsatz des Monomers bei der Polymerisation mit tetrafunktionalen CTAs erst nach 4 h erreicht. Diese Verzögerung ist auf eine Retardierung in der Anfangsphase der Polymerisation zurückzuführen, die sich aus der besonderen Struktur der tetrafunktionalen CTAs erklärt. Auf dem System zur Polymerisation von DMA aufbauend ließen sich Gradienten-Block-Copolymere in Eintopfreaktionen herstellen. Dazu wurde nach Erreichen des quantitativen Umsatzes von DMA ein zweites Monomer zur Reaktionsmischung gegeben. Mit Ethylacrylat (EtA) wurden so lineare amphiphile symmetrische Triblock-Copolymere erhalten. Dabei wurde die Länge des hydrophoben Blocks durch unterschiedliche Mengen an EtA variiert. Mit N,N-Diethylacrylamid (DEA) wurden lineare symmetrische Triblock-Copolymere sowie 3-Arm Stern-Diblock-Copolymere hergestellt, die über einen thermisch schaltbaren zweiten Block verfügen. Bei diesen Polymeren lässt sich die Länge des hydrophoben Teils in situ durch Veränderung der Temperatur variieren. Das Verhalten der amphiphilen Sternpolymere in wässriger Lösung und in Mikroemulsion wurde im Rahmen einer Kooperation an der TU Berlin mit Hilfe von Kleinwinkel-Neutronenstreuung (SANS), dynamischer Lichtstreuung (DLS) und Rheologie untersucht. Die Polymere wirken durch Assoziation der hydrophoben Endgruppen als effektive Verdicker sowohl allein in wässriger Lösung als auch in Mikroemulsion. Die Struktur des gebildeten Netzwerks hängt dabei von der Konzentration des Polymers in der Lösung und der Länge der Endgruppe (Hydrophobie) ab. Die dynamischen Eigenschaften der Lösungen werden außerdem durch die Anzahl der Arme der Polymere bestimmt. / Typically, associative thickeners for aqueous system consist of linear, hydrophobically α,ω-end-capped poly(ethylene glycols) (PEGs). Owing to their structure, these polymers aggregate in aqueous solution, forming a network of bridged micelles. Thus, one polymer molecule can link not more than two micelles. Until now it is unclear whether the structure and dynamics of such networks are influenced by the number of end groups of a branched multiply hydrophobically end-capped hydrophilic polymers. Branched PEG-based polymers are synthesized using the laborious and limited techniques of living ionic polymerization. Introducing hydrophobic end groups demands a multiple-step process. This work presents the one-step synthesis of hydrophilic star polymers with hydrophobic end groups, using reversible addition fragmentation chain transfer (RAFT) polymerization. This radical polymerization method is easy to use and tolerates a large number of polar monomers for the synthesis of the hydrophilic arms of the star polymers. The arms of the polymer were grown from a multifunctional core that formed the R-group of the chain transfer agents (CTAs). The CTAs where tailored to be able to vary the number of arms of the star polymers from 2 to 4 and to vary the length (and therefore the hydrophobicity) of the end groups (C4, C12, C18). Two different polar monomers where used as model monomers: Oligo(ethylene glycol)methyl ether acrylate (OEGA) and N,N-Dimethylacrylamide (DMA). Both monomers yield non-ionic hydrophilic polymers. While poly(OEGA) is a comb polymer based on PEG, poly(DMA) exhibits a more compact structure. The amphiphilic star polymers were characterized extensively. The molar masses were determined using GPC in various solvents and the degree of end functionalisation was monitored using 1H NMR and UV/Vis spectroscopy. The polymerization of OEGA shows some of the expected characteristics of reversible deactivation radical polymerization (RDRP). However, chain transfer to monomer and polymer is a prominent side reaction, limiting the use of this monomer for the fabrication of well-defined material. This reaction can be attributed to the structure of the monomer being an oligoether. For all examined polymerizations of DMA with the multifunctional CTAs the molar mass increased linearly with conversion. The molar mass distributions were monomodal and narrow (PDI ≤ 1.2). Expected values were reached for molar masses from 25 to 150 kg/mol and the end group functionality was about 90 % in all cases. While the polymerization of DMA using di- and trifunctional CTAs proceeded to quantitative conversion within 3 h, an initial retardation period of about 60 min was observed for the polymerization using tetrafunctional CTAs. This retardation was attributed to the peculiar molecular structure of these CTAs. Owing to the well-controlled features of the polymerization of DMA using the multifunctional CTAs, this system was used to obtain tapered block copolymers in a one-pot process. These structures were achieved by adding a second monomer to the reaction mixture after the quantitative conversion of DMA. Using ethyl acrylate (EtA), linear amphiphilic symmetrical triblock copolymers were synthesized. The length of the hydrophobic block was tailored by the addition of varying amounts of EtA. With N,N-Diethylacrylamide as a second monomer, linear symmetric triblock copolymers as well as 3-arm star diblock copolymers were obtained that contain a thermosensitve block. Altering the temperature of aqueous solutions of these polymers varies the length of the hydrophobic block in situ. At the TU Berlin, the behavior of the polymers was studied in aqueous solution as well as in microemulsion. The solutions were characterized by small angle neutron scattering (SANS), dynamic light scattering (DLS) and rheology. The end groups of the polymers aggregate, making the polymers efficient thickeners both in aqueous solution and in microemulsion. The structure of the formed network depends on the concentration of the polymer in solution and on the length of the end group. The dynamic properties of the solutions are governed additionally by the number of arms.

The overall goal of this thesis is to better understand mechanisms governing associative phase separation within admixtures of plant proteins (e.g., pea and canola) and anionic polysaccharides (e.g., gum Arabic, alginate or é-carrageenan). The process involves the electrostatic attraction between two biopolymers of opposing charges, and typically results in the formation of both soluble and insoluble complexes during an acidic pH titration. If successful, polysaccharides could be triggered to coat the proteins surface to give novel, and hopefully improved functionality as ingredients for food and biomaterials.<p> In the first study, the effect of protein enrichment and pH on the formation of soluble and insoluble complexes in admixtures of pea legumin (Lg) and vicilin (Vn) isolates with gum Arabic (GA) was investigated by turbidimetric, surface charge and fluorometric measurements. The solubility of the protein isolates and mixed biopolymer systems was also studied as a function of pH. Enrichment of the crude Lg and Vn isolates by low pressure liquid chromatography led to a shift towards higher pHs at the onset of soluble complex formation in the presence of GA for both protein isolates, whereas the onset of insoluble complex formation was unaffected. Complexation of the Lg (or Vn) isolates with GA resulted in a shift in the pH where neutrality (zeta potential = 0 mV) occurred to lower pH values, relative to the Lg (or Vn) isolates alone. In the case of the enriched Vn isloate, changes to its tertiary structure were observed by fluorometry upon complexation with GA, whereas no changes were found for the enriched Lg isolate. Complexation of Lg and Vn isolates with GA also had little effect on their solubilities relative to protein alone solutions.<p> In the second study, the formation of soluble and insoluble complexes, and the nature of their interactions as determined by optical density analysis, were investigated in admixtures of canola protein isolate (CPI) and anionic polysaccharides (alginate and é-carrageenan) as a function of pH and biopolymer weight mixing ratio. The solubilities of formed complexes were also investigated versus protein alone. In both CPI-polysaccharide systems, critical pH associated with the onset of soluble and insoluble complexes shifted to higher pHs as the mixing ratios increased from 1:1 to 20:1 (CPI:polysaccharide), and then became constant. There complexes formed primarily through electrostatic attractive forces with secondary stabilization by hydrogen bonding. The solubilities of the CPI-alginate complexes were significantly enhanced relative to CPI alone or CPI-é-carrageenan, which were similar.

Ternary content addressable memories (TCAMs) are hardware-based parallel lookup tables with bit-level masking capability. They are attractive for applications such as packet forwarding and classification in network routers. Despite the attractive features of TCAMs, high power consumption is one of the most critical challenges faced by TCAM designers. This work proposes circuit techniques for reducing TCAM power consumption. The main contribution of this work is divided in two parts: (i) reduction in match line (ML) sensing energy, and (ii) static-power reduction techniques. The ML sensing energy is reduced by employing (i) positive-feedback ML sense amplifiers (MLSAs), (ii) low-capacitance comparison logic, and (iii) low-power ML-segmentation techniques. The positive-feedback MLSAs include both resistive and active feedback to reduce the ML sensing energy. A body-bias technique can further improve the feedback action at the expense of additional area and ML capacitance. The measurement results of the active-feedback MLSA show 50-56% reduction in ML sensing energy. The measurement results of the proposed low-capacitance comparison logic show 25% and 42% reductions in ML sensing energy and time, respectively, which can further be improved by careful layout. The low-power ML-segmentation techniques include dual ML TCAM and charge-shared ML. Simulation results of the dual ML TCAM that connects two sides of the comparison logic to two ML segments for sequential sensing show 43% power savings for a small (4%) trade-off in the search speed. The charge-shared ML scheme achieves power savings by partial recycling of the charge stored in the first ML segment. Chip measurement results show that the charge-shared ML scheme results in 11% and 9% reductions in ML sensing time and energy, respectively, which can be improved to 19-25% by using a digitally controlled charge sharing time-window and a slightly modified MLSA. The static power reduction is achieved by a dual-VDD technique and low-leakage TCAM cells. The dual-VDD technique trades-off the excess noise margin of MLSA for smaller cell leakage by applying a smaller VDD to TCAM cells and a larger VDD to the peripheral circuits. The low-leakage TCAM cells trade off the speed of READ and WRITE operations for smaller cell area and leakage. Finally, design and testing of a complete TCAM chip are presented, and compared with other published designs.

Ternary content addressable memories (TCAMs) are hardware-based parallel lookup tables with bit-level masking capability. They are attractive for applications such as packet forwarding and classification in network routers. Despite the attractive features of TCAMs, high power consumption is one of the most critical challenges faced by TCAM designers. This work proposes circuit techniques for reducing TCAM power consumption. The main contribution of this work is divided in two parts: (i) reduction in match line (ML) sensing energy, and (ii) static-power reduction techniques. The ML sensing energy is reduced by employing (i) positive-feedback ML sense amplifiers (MLSAs), (ii) low-capacitance comparison logic, and (iii) low-power ML-segmentation techniques. The positive-feedback MLSAs include both resistive and active feedback to reduce the ML sensing energy. A body-bias technique can further improve the feedback action at the expense of additional area and ML capacitance. The measurement results of the active-feedback MLSA show 50-56% reduction in ML sensing energy. The measurement results of the proposed low-capacitance comparison logic show 25% and 42% reductions in ML sensing energy and time, respectively, which can further be improved by careful layout. The low-power ML-segmentation techniques include dual ML TCAM and charge-shared ML. Simulation results of the dual ML TCAM that connects two sides of the comparison logic to two ML segments for sequential sensing show 43% power savings for a small (4%) trade-off in the search speed. The charge-shared ML scheme achieves power savings by partial recycling of the charge stored in the first ML segment. Chip measurement results show that the charge-shared ML scheme results in 11% and 9% reductions in ML sensing time and energy, respectively, which can be improved to 19-25% by using a digitally controlled charge sharing time-window and a slightly modified MLSA. The static power reduction is achieved by a dual-VDD technique and low-leakage TCAM cells. The dual-VDD technique trades-off the excess noise margin of MLSA for smaller cell leakage by applying a smaller VDD to TCAM cells and a larger VDD to the peripheral circuits. The low-leakage TCAM cells trade off the speed of READ and WRITE operations for smaller cell area and leakage. Finally, design and testing of a complete TCAM chip are presented, and compared with other published designs.

The overall goal of this thesis is to better understand mechanisms governing associative phase separation within admixtures of plant proteins (e.g., pea and canola) and anionic polysaccharides (e.g., gum Arabic, alginate or é-carrageenan). The process involves the electrostatic attraction between two biopolymers of opposing charges, and typically results in the formation of both soluble and insoluble complexes during an acidic pH titration. If successful, polysaccharides could be triggered to coat the proteins surface to give novel, and hopefully improved functionality as ingredients for food and biomaterials.<p> In the first study, the effect of protein enrichment and pH on the formation of soluble and insoluble complexes in admixtures of pea legumin (Lg) and vicilin (Vn) isolates with gum Arabic (GA) was investigated by turbidimetric, surface charge and fluorometric measurements. The solubility of the protein isolates and mixed biopolymer systems was also studied as a function of pH. Enrichment of the crude Lg and Vn isolates by low pressure liquid chromatography led to a shift towards higher pHs at the onset of soluble complex formation in the presence of GA for both protein isolates, whereas the onset of insoluble complex formation was unaffected. Complexation of the Lg (or Vn) isolates with GA resulted in a shift in the pH where neutrality (zeta potential = 0 mV) occurred to lower pH values, relative to the Lg (or Vn) isolates alone. In the case of the enriched Vn isloate, changes to its tertiary structure were observed by fluorometry upon complexation with GA, whereas no changes were found for the enriched Lg isolate. Complexation of Lg and Vn isolates with GA also had little effect on their solubilities relative to protein alone solutions.<p> In the second study, the formation of soluble and insoluble complexes, and the nature of their interactions as determined by optical density analysis, were investigated in admixtures of canola protein isolate (CPI) and anionic polysaccharides (alginate and é-carrageenan) as a function of pH and biopolymer weight mixing ratio. The solubilities of formed complexes were also investigated versus protein alone. In both CPI-polysaccharide systems, critical pH associated with the onset of soluble and insoluble complexes shifted to higher pHs as the mixing ratios increased from 1:1 to 20:1 (CPI:polysaccharide), and then became constant. There complexes formed primarily through electrostatic attractive forces with secondary stabilization by hydrogen bonding. The solubilities of the CPI-alginate complexes were significantly enhanced relative to CPI alone or CPI-é-carrageenan, which were similar.

The major aim of this study is to evaluate the role of Turkish non-governmental organizations (NGOs) in the consolidation of democracy in Turkey regarding the characteristics defined by the classical pluralist theories, starting with the ideas of Alexis de Tocqueville, and by contemporary pluralist approaches to civil society. To that end, characteristics of two examples, iHD and KiH-Y&Ccedil / V will be examined using three dimensions: effectiveness, enhancing solidarity and responsibility, and the dimension of inner democracy. As the methodology of the case study, in-depth interview techniques were adapted for capturing the qualities of the organizations in a detailed way. At that level, in-depth interviews from each organization were conducted with members having different positions. In addition to this, all written documents obtained from the organizations were scanned and the relevant ones have been examined for the sake of the study. Since the aim of the study is to understand what part Turkish NGOs play in the consolidation of democracy in Turkey considering the assumptions of classical and contemporary pluralist school in their analysis of voluntary associations and interest groups, this study espoused an associational concept of democracy as the method of work, which is a contemporary concept that underlines the democratic role of free and voluntary associations. In this context, this study tries to discuss the following questions: To what extent are Turkish NGOs independent from the state? To what extent can they have leverage on governmental policies? Do they mitigate conflict through overlapping interests? Do they enhance social trust, tolerance, compromise and a sense of solidarity in society and within their organization? To what extent are they democratic, pluralist and horizontally structured inside their organization and to what extent they are democratic regarding their decision making process? Are they open and inclusive enough in their membership recruitment techniques?

No description available.

No description available.

Sequential structure imposed by the forward linear progression of time is omnipresent in all cognitive behaviors. This thesis proposes a novel model to store sequences of any length, scalar or vectorial, in binary neural networks. Particularly, the model that we introduce allows resolving some well known problems in sequential learning, such as error intolerance, catastrophic forgetting and the interference issue while storing complex sequences, etc. The quantity of the total sequential information that the network is able to store grows quadratically with the number of nodes. And the efficiency - the ratio between the capacity and the total amount of information consumed by the storage device - can reach around 30%. This work could be considered as an extension of the non oriented clique-based neural networks previously proposed and studied within our team. Those networks composed of binary neurons and binary connections utilize the concept of graph redundancy and sparsity in order to acquire a quadratic learning diversity. To obtain the ability to store sequences, connections are provided with orientation to form a tournament-based neural network. This is more natural biologically speak- ing, since communication between neurons is unidirectional, from axons to synapses. Any component of the network, a cluster or a node, can be revisited several times within a sequence or by multiple sequences. This allows the network to store se- quences of any length, independent of the number of clusters and only limited by the total available resource of the network. Moreover, in order to allow error correction and provide robustness to the net- work, both spatial assembly redundancy and sequential redundancy, with or without anticipation, may be combined to offer a large amount of redundancy in the activation of a node. Subsequently, a double layered structure is introduced with the purpose of accurate retrieval. The lower layer of tournament-based hetero-associative network stores sequential oriented associations between patterns. An upper auto-associative layer of mirror nodes is superposed to emphasize the co-occurrence of the elements belonging to the same pattern, in the form of a clique. This model is then extended to a hierarchical structure, which helps resolve the interference issue while storing complex sequences. This thesis also contributes in proposing and assessing new decoding rules with respect to sparse messages in order to fully benefit from the theoretical quadratic law of the learning diversity. Besides the performance aspect, the biological plausibility is also a constant concern during this thesis work.