Adaptive Algorithms for Personalized Diabetes Treatment

Dynamic systems, especially in real-life applications, are often determined by inter-/intra-variability, uncertainties and time-varying components. Physiological systems are probably the most representative example in which population variability, vital signal measurement noise and uncertain dynamics render their explicit representation and optimization a rather difficult task. Systems characterized by such challenges often require the use of adaptive algorithmic solutions able to perform an iterative structural and/or parametrical update process towards optimized behavior. Adaptive optimization presents the advantages of (i) individualization through learning of basic system characteristics, (ii) ability to follow time-varying dynamics and (iii) low computational cost. In this chapter, the use of online adaptive algorithms is investigated in two basic research areas related to diabetes management: (i) real-time glucose regulation and (ii) real-time prediction of hypo-/hyperglycemia. The applicability of these methods is illustrated through the design and development of an adaptive glucose control algorithm based on reinforcement learning and optimal control and an adaptive, personalized early-warning system for the recognition and alarm generation against hypo- and hyperglycemic events.

Modeling Techniques for the High-Resolution Interpretation of Cryo-Electron Microscopy Reconstructions

Essential biological processes are governed by organized, dynamic interactions between multiple biomolecular systems. Complexes are thus formed to enable the biological function and get dissembled as the process is completed. Examples of such processes include the translation of the messenger RNA into protein by the ribosome, the folding of proteins by chaperonins or the entry of viruses in host cells. Understanding these fundamental processes by characterizing the molecular mechanisms that enable then, would allow the (better) design of therapies and drugs. Such molecular mechanisms may be revealed trough the structural elucidation of the biomolecular assemblies at the core of these processes. Various experimental techniques may be applied to investigate the molecular architecture of biomolecular assemblies. High-resolution techniques, such as X-ray crystallography, may solve the atomic structure of the system, but are typically constrained to biomolecules of reduced flexibility and dimensions. In particular, X-ray crystallography requires the sample to form a three dimensional (3D) crystal lattice which is technically di‑cult, if not impossible, to obtain, especially for large, dynamic systems. Often these techniques solve the structure of the different constituent components within the assembly, but encounter difficulties when investigating the entire system. On the other hand, imaging techniques, such as cryo-electron microscopy (cryo-EM), are able to depict large systems in near-native environment, without requiring the formation of crystals. The structures solved by cryo-EM cover a wide range of resolutions, from very low level of detail where only the overall shape of the system is visible, to high-resolution that approach, but not yet reach, atomic level of detail. In this dissertation, several modeling methods are introduced to either integrate cryo-EM datasets with structural data from X-ray crystallography, or to directly interpret the cryo-EM reconstruction. Such computational techniques were developed with the goal of creating an atomic model for the cryo-EM data. The low-resolution reconstructions lack the level of detail to permit a direct atomic interpretation, i.e. one cannot reliably locate the atoms or amino-acid residues within the structure obtained by cryo-EM. Thereby one needs to consider additional information, for example, structural data from other sources such as X-ray crystallography, in order to enable such a high-resolution interpretation. Modeling techniques are thus developed to integrate the structural data from the different biophysical sources, examples including the work described in the manuscript I and II of this dissertation. At intermediate and high-resolution, cryo-EM reconstructions depict consistent 3D folds such as tubular features which in general correspond to alpha-helices. Such features can be annotated and later on used to build the atomic model of the system, see manuscript III as alternative. Three manuscripts are presented as part of the PhD dissertation, each introducing a computational technique that facilitates the interpretation of cryo-EM reconstructions. The first manuscript is an application paper that describes a heuristics to generate the atomic model for the protein envelope of the Rift Valley fever virus. The second manuscript introduces the evolutionary tabu search strategies to enable the integration of multiple component atomic structures with the cryo-EM map of their assembly. Finally, the third manuscript develops further the latter technique and apply it to annotate consistent 3D patterns in intermediate-resolution cryo-EM reconstructions. The first manuscript, titled An assembly model for Rift Valley fever virus, was submitted for publication in the Journal of Molecular Biology. The cryo-EM structure of the Rift Valley fever virus was previously solved at 27Å-resolution by Dr. Freiberg and collaborators. Such reconstruction shows the overall shape of the virus envelope, yet the reduced level of detail prevents the direct atomic interpretation. High-resolution structures are not yet available for the entire virus nor for the two different component glycoproteins that form its envelope. However, homology models may be generated for these glycoproteins based on similar structures that are available at atomic resolutions. The manuscript presents the steps required to identify an atomic model of the entire virus envelope, based on the low-resolution cryo-EM map of the envelope and the homology models of the two glycoproteins. Starting with the results of the exhaustive search to place the two glycoproteins, the model is built iterative by running multiple multi-body refinements to hierarchically generate models for the different regions of the envelope. The generated atomic model is supported by prior knowledge regarding virus biology and contains valuable information about the molecular architecture of the system. It provides the basis for further investigations seeking to reveal different processes in which the virus is involved such as assembly or fusion. The second manuscript was recently published in the of Journal of Structural Biology (doi:10.1016/j.jsb.2009.12.028) under the title Evolutionary tabu search strategies for the simultaneous registration of multiple atomic structures in cryo-EM reconstructions. This manuscript introduces the evolutionary tabu search strategies applied to enable a multi-body registration. This technique is a hybrid approach that combines a genetic algorithm with a tabu search strategy to promote the proper exploration of the high-dimensional search space. Similar to the Rift Valley fever virus, it is common that the structure of a large multi-component assembly is available at low-resolution from cryo-EM, while high-resolution structures are solved for the different components but lack for the entire system. Evolutionary tabu search strategies enable the building of an atomic model for the entire system by considering simultaneously the different components. Such registration indirectly introduces spatial constrains as all components need to be placed within the assembly, enabling the proper docked in the low-resolution map of the entire assembly. Along with the method description, the manuscript covers the validation, presenting the benefit of the technique in both synthetic and experimental test cases. Such approach successfully docked multiple components up to resolutions of 40Å. The third manuscript is entitled Evolutionary Bidirectional Expansion for the Annotation of Alpha Helices in Electron Cryo-Microscopy Reconstructions and was submitted for publication in the Journal of Structural Biology. The modeling approach described in this manuscript applies the evolutionary tabu search strategies in combination with the bidirectional expansion to annotate secondary structure elements in intermediate resolution cryo-EM reconstructions. In particular, secondary structure elements such as alpha helices show consistent patterns in cryo-EM data, and are visible as rod-like patterns of high density. The evolutionary tabu search strategy is applied to identify the placement of the different alpha helices, while the bidirectional expansion characterizes their length and curvature. The manuscript presents the validation of the approach at resolutions ranging between 6 and 14Å, a level of detail where alpha helices are visible. Up to resolution of 12 Å, the method measures sensitivities between 70-100% as estimated in experimental test cases, i.e. 70-100% of the alpha-helices were correctly predicted in an automatic manner in the experimental data. The three manuscripts presented in this PhD dissertation cover different computation methods for the integration and interpretation of cryo-EM reconstructions. The methods were developed in the molecular modeling software Sculptor (http://sculptor.biomachina.org) and are available for the scientific community interested in the multi-resolution modeling of cryo-EM data. The work spans a wide range of resolution covering multi-body refinement and registration at low-resolution along with annotation of consistent patterns at high-resolution. Such methods are essential for the modeling of cryo-EM data, and may be applied in other fields where similar spatial problems are encountered, such as medical imaging.

Energy and power aware scheduling algorithm modeling for battery powered embedded systems

Energy management has always been recognized as a challenge in mobile systems, especially in modern OS-based mobile systems where multi-functioning are widely supported. Nowadays, it is common for a mobile system user to run multiple applications simultaneously while having a target battery lifetime in mind for a specific application. Traditional OS-level power management (PM) policies make their best effort to save energy under performance constraint, but fail to guarantee a target lifetime, leaving the painful trading off between the total performance of applications and the target lifetime to the user itself. This thesis provides a new way to deal with the problem. It is advocated that a strong energy-aware PM scheme should first guarantee a user-specified battery lifetime to a target application by restricting the average power of those less important applications, and in addition to that, maximize the total performance of applications without harming the lifetime guarantee. As a support, energy, instead of CPU or transmission bandwidth, should be globally managed as the first-class resource by the OS. As the first-stage work of a complete PM scheme, this thesis presents the energy-based fair queuing scheduling, a novel class of energy-aware scheduling algorithms which, in combination with a mechanism of battery discharge rate restricting, systematically manage energy as the first-class resource with the objective of guaranteeing a user-specified battery lifetime for a target application in OS-based mobile systems. Energy-based fair queuing is a cross-application of the traditional fair queuing in the energy management domain. It assigns a power share to each task, and manages energy by proportionally serving energy to tasks according to their assigned power shares. The proportional energy use establishes proportional share of the system power among tasks, which guarantees a minimum power for each task and thus, avoids energy starvation on any task. Energy-based fair queuing treats all tasks equally as one type and supports periodical time-sensitive tasks by allocating each of them a share of system power that is adequate to meet the highest energy demand in all periods. However, an overly conservative power share is usually required to guarantee the meeting of all time constraints. To provide more effective and flexible support for various types of time-sensitive tasks in general purpose operating systems, an extra real-time friendly mechanism is introduced to combine priority-based scheduling into the energy-based fair queuing. Since a method is available to control the maximum time one time-sensitive task can run with priority, the power control and time-constraint meeting can be flexibly traded off. A SystemC-based test-bench is designed to assess the algorithms. Simulation results show the success of the energy-based fair queuing in achieving proportional energy use, time-constraint meeting, and a proper trading off between them. La gestión de energía en los sistema móviles está considerada hoy en día como un reto fundamental, notándose, especialmente, en aquellos terminales que utilizando un sistema operativo implementan múltiples funciones. Es común en los sistemas móviles actuales ejecutar simultaneamente diferentes aplicaciones y tener, para una de ellas, un objetivo de tiempo de uso de la batería. Tradicionalmente, las políticas de gestión de consumo de potencia de los sistemas operativos hacen lo que está en sus manos para ahorrar energía y satisfacer sus requisitos de prestaciones, pero no son capaces de proporcionar un objetivo de tiempo de utilización del sistema, dejando al usuario la difícil tarea de buscar un compromiso entre prestaciones y tiempo de utilización del sistema. Esta tesis, como contribución, proporciona una nueva manera de afrontar el problema. En ella se establece que un esquema de gestión de consumo de energía debería, en primer lugar, garantizar, para una aplicación dada, un tiempo mínimo de utilización de la batería que estuviera especificado por el usuario, restringiendo la potencia media consumida por las aplicaciones que se puedan considerar menos importantes y, en segundo lugar, maximizar las prestaciones globales sin comprometer la garantía de utilización de la batería. Como soporte de lo anterior, la energía, en lugar del tiempo de CPU o el ancho de banda, debería gestionarse globalmente por el sistema operativo como recurso de primera clase. Como primera fase en el desarrollo completo de un esquema de gestión de consumo, esta tesis presenta un algoritmo de planificación de encolado equitativo (fair queueing) basado en el consumo de energía, es decir, una nueva clase de algoritmos de planificación que, en combinación con mecanismos que restrinjan la tasa de descarga de una batería, gestionen de forma sistemática la energía como recurso de primera clase, con el objetivo de garantizar, para una aplicación dada, un tiempo de uso de la batería, definido por el usuario, en sistemas móviles empotrados. El encolado equitativo de energía es una extensión al dominio de la energía del encolado equitativo tradicional. Esta clase de algoritmos asigna una reserva de potencia a cada tarea y gestiona la energía sirviéndola de manera proporcional a su reserva. Este uso proporcional de la energía garantiza que cada tarea reciba una porción de potencia y evita que haya tareas que se vean privadas de recibir energía por otras con un comportamiento más ambicioso. Esta clase de algoritmos trata a todas las tareas por igual y puede planificar tareas periódicas en tiempo real asignando a cada una de ellas una reserva de potencia que es adecuada para proporcionar la mayor de las cantidades de energía demandadas por período. Sin embargo, es posible demostrar que sólo se consigue cumplir con los requisitos impuestos por todos los plazos temporales con reservas de potencia extremadamente conservadoras. En esta tesis, para proporcionar un soporte más flexible y eficiente para diferentes tipos de tareas de tiempo real junto con el resto de tareas, se combina un mecanismo de planificación basado en prioridades con el encolado equitativo basado en energía. En esta clase de algoritmos, gracias al método introducido, que controla el tiempo que se ejecuta con prioridad una tarea de tiempo real, se puede establecer un compromiso entre el cumplimiento de los requisitos de tiempo real y el consumo de potencia. Para evaluar los algoritmos, se ha diseñado en SystemC un banco de pruebas. Los resultados muestran que el algoritmo de encolado equitativo basado en el consumo de energía consigue el balance entre el uso proporcional a la energía reservada y el cumplimiento de los requisitos de tiempo real.

Multi-objective Estimation of Distribution Algorithm Based on Joint Modeling of Objectives and Variables

This paper proposes a new multi-objective estimation of distribution algorithm (EDA) based on joint modeling of objectives and variables. This EDA uses the multi-dimensional Bayesian network as its probabilistic model. In this way it can capture the dependencies between objectives, variables and objectives, as well as the dependencies learnt between variables in other Bayesian network-based EDAs. This model leads to a problem decomposition that helps the proposed algorithm to find better trade-off solutions to the multi-objective problem. In addition to Pareto set approximation, the algorithm is also able to estimate the structure of the multi-objective problem. To apply the algorithm to many-objective problems, the algorithm includes four different ranking methods proposed in the literature for this purpose. The algorithm is applied to the set of walking fish group (WFG) problems, and its optimization performance is compared with an evolutionary algorithm and another multi-objective EDA. The experimental results show that the proposed algorithm performs significantly better on many of the problems and for different objective space dimensions, and achieves comparable results on some compared with the other algorithms.

Adapting autonomously classification data mining algorithms for ubiquitous devices

Abstract Due to recent scientific and technological advances in information sys¬tems, it is now possible to perform almost every application on a mobile device. The need to make sense of such devices more intelligent opens an opportunity to design data mining algorithm that are able to autonomous execute in local devices to provide the device with knowledge. The problem behind autonomous mining deals with the proper configuration of the algorithm to produce the most appropriate results. Contextual information together with resource information of the device have a strong impact on both the feasibility of a particu¬lar execution and on the production of the proper patterns. On the other hand, performance of the algorithm expressed in terms of efficacy and efficiency highly depends on the features of the dataset to be analyzed together with values of the parameters of a particular implementation of an algorithm. However, few existing approaches deal with autonomous configuration of data mining algorithms and in any case they do not deal with contextual or resources information. Both issues are of particular significance, in particular for social net¬works application. In fact, the widespread use of social networks and consequently the amount of information shared have made the need of modeling context in social application a priority. Also the resource consumption has a crucial role in such platforms as the users are using social networks mainly on their mobile devices. This PhD thesis addresses the aforementioned open issues, focusing on i) Analyzing the behavior of algorithms, ii) mapping contextual and resources information to find the most appropriate configuration iii) applying the model for the case of a social recommender. Four main contributions are presented: - The EE-Model: is able to predict the behavior of a data mining algorithm in terms of resource consumed and accuracy of the mining model it will obtain. - The SC-Mapper: maps a situation defined by the context and resource state to a data mining configuration. - SOMAR: is a social activity (event and informal ongoings) recommender for mobile devices. - D-SOMAR: is an evolution of SOMAR which incorporates the configurator in order to provide updated recommendations. Finally, the experimental validation of the proposed contributions using synthetic and real datasets allows us to achieve the objectives and answer the research questions proposed for this dissertation.

A fitting algorithm for random modeling the PLC channel

The characteristics of the power-line communication (PLC) channel are difficult to model due to the heterogeneity of the networks and the lack of common wiring practices. To obtain the full variability of the PLC channel, random channel generators are of great importance for the design and testing of communication algorithms. In this respect, we propose a random channel generator that is based on the top-down approach. Basically, we describe the multipath propagation and the coupling effects with an analytical model. We introduce the variability into a restricted set of parameters and, finally, we fit the model to a set of measured channels. The proposed model enables a closed-form description of both the mean path-loss profile and the statistical correlation function of the channel frequency response. As an example of application, we apply the procedure to a set of in-home measured channels in the band 2-100 MHz whose statistics are available in the literature. The measured channels are divided into nine classes according to their channel capacity. We provide the parameters for the random generation of channels for all nine classes, and we show that the results are consistent with the experimental ones. Finally, we merge the classes to capture the entire heterogeneity of in-home PLC channels. In detail, we introduce the class occurrence probability, and we present a random channel generator that targets the ensemble of all nine classes. The statistics of the composite set of channels are also studied, and they are compared to the results of experimental measurement campaigns in the literature.

Modeling the evolution of item rating networks using time-domain preferential attachment

The understanding of the structure and dynamics of the intricate network of connections among people that consumes products through Internet appears as an extremely useful asset in order to study emergent properties related to social behavior. This knowledge could be useful, for example, to improve the performance of personal recommendation algorithms. In this contribution, we analyzed five-year records of movie-rating transactions provided by Netflix, a movie rental platform where users rate movies from an online catalog. This dataset can be studied as a bipartite user-item network whose structure evolves in time. Even though several topological properties from subsets of this bipartite network have been reported with a model that combines random and preferential attachment mechanisms [Beguerisse Díaz et al., 2010], there are still many aspects worth to be explored, as they are connected to relevant phenomena underlying the evolution of the network. In this work, we test the hypothesis that bursty human behavior is essential in order to describe how a bipartite user-item network evolves in time. To that end, we propose a novel model that combines, for user nodes, a network growth prescription based on a preferential attachment mechanism acting not only in the topological domain (i.e. based on node degrees) but also in time domain. In the case of items, the model mixes degree preferential attachment and random selection. With these ingredients, the model is not only able to reproduce the asymptotic degree distribution, but also shows an excellent agreement with the Netflix data in several time-dependent topological properties.

Advanced background modeling with RGB-D sensors through classifiers combination and inter-frame foreground prediction

An innovative background modeling technique that is able to accurately segment foreground regions in RGB-D imagery (RGB plus depth) has been presented in this paper. The technique is based on a Bayesian framework that efficiently fuses different sources of information to segment the foreground. In particular, the final segmentation is obtained by considering a prediction of the foreground regions, carried out by a novel Bayesian Network with a depth-based dynamic model, and, by considering two independent depth and color-based mixture of Gaussians background models. The efficient Bayesian combination of all these data reduces the noise and uncertainties introduced by the color and depth features and the corresponding models. As a result, more compact segmentations, and refined foreground object silhouettes are obtained. Experimental results with different databases suggest that the proposed technique outperforms existing state-of-the-art algorithms.

Atomistic modeling of Ge damage accumulation, amorphization and solid phase epitaxial regrowth

En los últimos años, el Ge ha ganado de nuevo atención con la finalidad de ser integrado en el seno de las existentes tecnologías de microelectrónica. Aunque no se le considera como un canddato capaz de reemplazar completamente al Si en el futuro próximo, probalemente servirá como un excelente complemento para aumentar las propiedades eléctricas en dispositivos futuros, especialmente debido a su alta movilidad de portadores. Esta integración requiere de un avance significativo del estado del arte en los procesos de fabricado. Técnicas de simulación, como los algoritmos de Monte Carlo cinético (KMC), proporcionan un ambiente atractivo para llevar a cabo investigación y desarrollo en este campo, especialmente en términos de costes en tiempo y financiación. En este estudio se han usado, por primera vez, técnicas de KMC con el fin entender el procesado “front-end” de Ge en su fabricación, específicamente la acumulación de dañado y amorfización producidas por implantación iónica y el crecimiento epitaxial en fase sólida (SPER) de las capas amorfizadas. Primero, simulaciones de aproximación de clisiones binarias (BCA) son usadas para calcular el dañado causado por cada ión. La evolución de este dañado en el tiempo se simula usando KMC sin red, o de objetos (OKMC) en el que sólamente se consideran los defectos. El SPER se simula a través de una aproximación KMC de red (LKMC), siendo capaz de seguir la evolución de los átomos de la red que forman la intercara amorfo/cristalina. Con el modelo de amorfización desarrollado a lo largo de este trabajo, implementado en un simulador multi-material, se pueden simular todos estos procesos. Ha sido posible entender la acumulación de dañado, desde la generación de defectos puntuales hasta la formación completa de capas amorfas. Esta acumulación ocurre en tres regímenes bien diferenciados, empezando con un ritmo lento de formación de regiones de dañado, seguido por una rápida relajación local de ciertas áreas en la fase amorfa donde ambas fases, amorfa y cristalina, coexisten, para terminar en la amorfización completa de capas extensas, donde satura el ritmo de acumulación. Dicha transición ocurre cuando la concentración de dañado supera cierto valor límite, el cual es independiente de las condiciones de implantación. Cuando se implantan los iones a temperaturas relativamente altas, el recocido dinámico cura el dañado previamente introducido y se establece una competición entre la generación de dañado y su disolución. Estos efectos se vuelven especialmente importantes para iones ligeros, como el B, el cual crea dañado más diluido, pequeño y distribuido de manera diferente que el causado por la implantación de iones más pesados, como el Ge. Esta descripción reproduce satisfactoriamente la cantidad de dañado y la extensión de las capas amorfas causadas por implantación iónica reportadas en la bibliografía. La velocidad de recristalización de la muestra previamente amorfizada depende fuertemente de la orientación del sustrato. El modelo LKMC presentado ha sido capaz de explicar estas diferencias entre orientaciones a través de un simple modelo, dominado por una única energía de activación y diferentes prefactores en las frecuencias de SPER dependiendo de las configuraciones de vecinos de los átomos que recristalizan. La formación de maclas aparece como una consecuencia de esta descripción, y es predominante en sustratos crecidos en la orientación (111)Ge. Este modelo es capaz de reproducir resultados experimentales para diferentes orientaciones, temperaturas y tiempos de evolución de la intercara amorfo/cristalina reportados por diferentes autores. Las parametrizaciones preliminares realizadas de los tensores de activación de tensiones son también capaces de proveer una buena correlación entre las simulaciones y los resultados experimentales de velocidad de SPER a diferentes temperaturas bajo una presión hidrostática aplicada. Los estudios presentados en esta tesis han ayudado a alcanzar un mejor entendimiento de los mecanismos de producción de dañado, su evolución, amorfización y SPER para Ge, además de servir como una útil herramienta para continuar el trabajo en este campo. In the recent years, Ge has regained attention to be integrated into existing microelectronic technologies. Even though it is not thought to be a feasible full replacement to Si in the near future, it will likely serve as an excellent complement to enhance electrical properties in future devices, specially due to its high carrier mobilities. This integration requires a significant upgrade of the state-of-the-art of regular manufacturing processes. Simulation techniques, such as kinetic Monte Carlo (KMC) algorithms, provide an appealing environment to research and innovation in the field, specially in terms of time and funding costs. In the present study, KMC techniques are used, for the first time, to understand Ge front-end processing, specifically damage accumulation and amorphization produced by ion implantation and Solid Phase Epitaxial Regrowth (SPER) of the amorphized layers. First, Binary Collision Approximation (BCA) simulations are used to calculate the damage caused by every ion. The evolution of this damage over time is simulated using non-lattice, or Object, KMC (OKMC) in which only defects are considered. SPER is simulated through a Lattice KMC (LKMC) approach, being able to follow the evolution of the lattice atoms forming the amorphous/crystalline interface. With the amorphization model developed in this work, implemented into a multi-material process simulator, all these processes can be simulated. It has been possible to understand damage accumulation, from point defect generation up to full amorphous layers formation. This accumulation occurs in three differentiated regimes, starting at a slow formation rate of the damage regions, followed by a fast local relaxation of areas into the amorphous phase where both crystalline and amorphous phases coexist, ending in full amorphization of extended layers, where the accumulation rate saturates. This transition occurs when the damage concentration overcomes a certain threshold value, which is independent of the implantation conditions. When implanting ions at relatively high temperatures, dynamic annealing takes place, healing the previously induced damage and establishing a competition between damage generation and its dissolution. These effects become specially important for light ions, as B, for which the created damage is more diluted, smaller and differently distributed than that caused by implanting heavier ions, as Ge. This description successfully reproduces damage quantity and extension of amorphous layers caused by means of ion implantation reported in the literature. Recrystallization velocity of the previously amorphized sample strongly depends on the substrate orientation. The presented LKMC model has been able to explain these differences between orientations through a simple model, dominated by one only activation energy and different prefactors for the SPER rates depending on the neighboring configuration of the recrystallizing atoms. Twin defects formation appears as a consequence of this description, and are predominant for (111)Ge oriented grown substrates. This model is able to reproduce experimental results for different orientations, temperatures and times of evolution of the amorphous/crystalline interface reported by different authors. Preliminary parameterizations for the activation strain tensors are able to also provide a good match between simulations and reported experimental results for SPER velocities at different temperatures under the appliance of hydrostatic pressure. The studies presented in this thesis have helped to achieve a greater understanding of damage generation, evolution, amorphization and SPER mechanisms in Ge, and also provide a useful tool to continue research in this field.

Graph-based Semi-Supervised Learning in Acoustic Modeling for Automatic Speech Recognition

Thesis (Ph.D.)--University of Washington, 2016-06

Patterns in complex systems modeling

The design, development, and use of complex systems models raises a unique class of challenges and potential pitfalls, many of which are commonly recurring problems. Over time, researchers gain experience in this form of modeling, choosing algorithms, techniques, and frameworks that improve the quality, confidence level, and speed of development of their models. This increasing collective experience of complex systems modellers is a resource that should be captured. Fields such as software engineering and architecture have benefited from the development of generic solutions to recurring problems, called patterns. Using pattern development techniques from these fields, insights from communities such as learning and information processing, data mining, bioinformatics, and agent-based modeling can be identified and captured. Collections of such 'pattern languages' would allow knowledge gained through experience to be readily accessible to less-experienced practitioners and to other domains. This paper proposes a methodology for capturing the wisdom of computational modelers by introducing example visualization patterns, and a pattern classification system for analyzing the relationship between micro and macro behaviour in complex systems models. We anticipate that a new field of complex systems patterns will provide an invaluable resource for both practicing and future generations of modelers.

Towards a Modeling Environment for Earth Systems Simulations

The developments of models in Earth Sciences, e.g. for earthquake prediction and for the simulation of mantel convection, are fare from being finalized. Therefore there is a need for a modelling environment that allows scientist to implement and test new models in an easy but flexible way. After been verified, the models should be easy to apply within its scope, typically by setting input parameters through a GUI or web services. It should be possible to link certain parameters to external data sources, such as databases and other simulation codes. Moreover, as typically large-scale meshes have to be used to achieve appropriate resolutions, the computational efficiency of the underlying numerical methods is important. Conceptional this leads to a software system with three major layers: the application layer, the mathematical layer, and the numerical algorithm layer. The latter is implemented as a C/C++ library to solve a basic, computational intensive linear problem, such as a linear partial differential equation. The mathematical layer allows the model developer to define his model and to implement high level solution algorithms (e.g. Newton-Raphson scheme, Crank-Nicholson scheme) or choose these algorithms form an algorithm library. The kernels of the model are generic, typically linear, solvers provided through the numerical algorithm layer. Finally, to provide an easy-to-use application environment, a web interface is (semi-automatically) built to edit the XML input file for the modelling code. In the talk, we will discuss the advantages and disadvantages of this concept in more details. We will also present the modelling environment escript which is a prototype implementation toward such a software system in Python (see www.python.org). Key components of escript are the Data class and the PDE class. Objects of the Data class allow generating, holding, accessing, and manipulating data, in such a way that the actual, in the particular context best, representation is transparent to the user. They are also the key to establish connections with external data sources. PDE class objects are describing (linear) partial differential equation objects to be solved by a numerical library. The current implementation of escript has been linked to the finite element code Finley to solve general linear partial differential equations. We will give a few simple examples which will illustrate the usage escript. Moreover, we show the usage of escript together with Finley for the modelling of interacting fault systems and for the simulation of mantel convection.

Quantum Information Processing with spin systems: from modeling to possible implementations

The physical implementation of quantum information processing is one of the major challenges of current research. In the last few years, several theoretical proposals and experimental demonstrations on a small number of qubits have been carried out, but a quantum computing architecture that is straightforwardly scalable, universal, and realizable with state-of-the-art technology is still lacking. In particular, a major ultimate objective is the construction of quantum simulators, yielding massively increased computational power in simulating quantum systems. Here we investigate promising routes towards the actual realization of a quantum computer, based on spin systems. The first one employs molecular nanomagnets with a doublet ground state to encode each qubit and exploits the wide chemical tunability of these systems to obtain the proper topology of inter-qubit interactions. Indeed, recent advances in coordination chemistry allow us to arrange these qubits in chains, with tailored interactions mediated by magnetic linkers. These act as switches of the effective qubit-qubit coupling, thus enabling the implementation of one- and two-qubit gates. Molecular qubits can be controlled either by uniform magnetic pulses, either by local electric fields. We introduce here two different schemes for quantum information processing with either global or local control of the inter-qubit interaction and demonstrate the high performance of these platforms by simulating the system time evolution with state-of-the-art parameters. The second architecture we propose is based on a hybrid spin-photon qubit encoding, which exploits the best characteristic of photons, whose mobility is exploited to efficiently establish long-range entanglement, and spin systems, which ensure long coherence times. The setup consists of spin ensembles coherently coupled to single photons within superconducting coplanar waveguide resonators. The tunability of the resonators frequency is exploited as the only manipulation tool to implement a universal set of quantum gates, by bringing the photons into/out of resonance with the spin transition. The time evolution of the system subject to the pulse sequence used to implement complex quantum algorithms has been simulated by numerically integrating the master equation for the system density matrix, thus including the harmful effects of decoherence. Finally a scheme to overcome the leakage of information due to inhomogeneous broadening of the spin ensemble is pointed out. Both the proposed setups are based on state-of-the-art technological achievements. By extensive numerical experiments we show that their performance is remarkably good, even for the implementation of long sequences of gates used to simulate interesting physical models. Therefore, the here examined systems are really promising buildingblocks of future scalable architectures and can be used for proof-of-principle experiments of quantum information processing and quantum simulation.

DVMS 1.5: A user manual (the data visualization and modeling system)

The data available during the drug discovery process is vast in amount and diverse in nature. To gain useful information from such data, an effective visualisation tool is required. To provide better visualisation facilities to the domain experts (screening scientist, biologist, chemist, etc.),we developed a software which is based on recently developed principled visualisation algorithms such as Generative Topographic Mapping (GTM) and Hierarchical Generative Topographic Mapping (HGTM). The software also supports conventional visualisation techniques such as Principal Component Analysis, NeuroScale, PhiVis, and Locally Linear Embedding (LLE). The software also provides global and local regression facilities . It supports regression algorithms such as Multilayer Perceptron (MLP), Radial Basis Functions network (RBF), Generalised Linear Models (GLM), Mixture of Experts (MoE), and newly developed Guided Mixture of Experts (GME). This user manual gives an overview of the purpose of the software tool, highlights some of the issues to be taken care while creating a new model, and provides information about how to install & use the tool. The user manual does not require the readers to have familiarity with the algorithms it implements. Basic computing skills are enough to operate the software.