Regression analysis is one of the most extensively used statistical tools applied across different fields of science, with linear regression being its most well-known method. How- ever, the traditional procedure to obtain the linear model estimates, the least squares approach, is highly sensitive to even slight departures from the assumed modelling frame- work. This is especially pronounced when atypical values occur in the observed data. This lack of stability of the least squares approach is a serious problem in applications. Thus, the focus of this thesis lies in assessing the available robust alternatives to least squares estimation, which are not so easily affected by any outlying values. First, we introduce the linear regression model theory and derive the least squares method. Then, we char- acterise different types of unusual observations and outline some fundamental robustness measures. Next, we define and examine the robust alternatives to the classical estimation in the linear regression models. Finally, we conduct a comprehensive simulation study comparing the performance of robust methods under different scenarios. 1

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第18405号 / 情博第520号 / 新制||情||92(附属図書館) / 31263 / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 阿久津 達也, 教授 山本 章博, 教授 永持 仁 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

Chapter 4 of this thesis is a minor revision of the work published in "Daisuke Kotani and Yasuo Okabe, Fast Failure Detection of OpenFlow Channels, The 11th Asian Internet Engineering Conference (AINTEC 2015), pp.32-39, November 2015. http://dx.doi.org/10.1145/2837030.2837035" © ACM 2015. / 京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第19847号 / 情博第598号 / 新制||情||104(附属図書館) / 32883 / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 岡部 寿男, 教授 美濃 導彦, 教授 喜多 一 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

No description available.

Indiana University-Purdue University Indianapolis (IUPUI) / Temporal point processes (TPP) are mathematical approaches for modeling asynchronous event sequences by considering the temporal dependency of each event on past events and its instantaneous rate. Temporal point processes can model various problems, from earthquake aftershocks, trade orders, gang violence, and reported crime patterns, to network analysis, infectious disease transmissions, and virus spread forecasting. In each of these cases, the entity’s behavior with the corresponding information is noted over time as an asynchronous event sequence, and the analysis is done using temporal point processes, which provides a means to define the generative mechanism of the sequence of events and ultimately predict events and investigate causality. Among point processes, Hawkes process as a stochastic point process is able to model a wide range of contagious and self-exciting patterns. One of Hawkes process’s well-known applications is predicting the evolution of viral processes on networks, which is an important problem in biology, the social sciences, and the study of the Internet. In existing works, mean-field analysis based upon degree distribution is used to predict viral spreading across networks of different types. However, it has been shown that degree distribution alone fails to predict the behavior of viruses on some real-world networks. Recent attempts have been made to use assortativity to address this shortcoming. This thesis illustrates how the evolution of such a viral process is sensitive to the underlying network’s structure. In Chapter 3 , we show that adding assortativity does not fully explain the variance in the spread of viruses for a number of real-world networks. We propose using the graphlet frequency distribution combined with assortativity to explain variations in the evolution of viral processes across networks with identical degree distribution. Using a data-driven approach, by coupling predictive modeling with viral process simulation on real-world networks, we show that simple regression models based on graphlet frequency distribution can explain over 95% of the variance in virality on networks with the same degree distribution but different network topologies. Our results highlight the importance of graphlets and identify a small collection of graphlets that may have the most significant influence over the viral processes on a network. Due to the flexibility and expressiveness of deep learning techniques, several neural network-based approaches have recently shown promise for modeling point process intensities. However, there is a lack of research on the possible adversarial attacks and the robustness of such models regarding adversarial attacks and natural shocks to systems. Furthermore, while neural point processes may outperform simpler parametric models on in-sample tests, how these models perform when encountering adversarial examples or sharp non-stationary trends remains unknown. In Chapter 4 , we propose several white-box and black-box adversarial attacks against deep temporal point processes. Additionally, we investigate the transferability of whitebox adversarial attacks against point processes modeled by deep neural networks, which are considered a more elevated risk. Extensive experiments confirm that neural point processes are vulnerable to adversarial attacks. Such a vulnerability is illustrated both in terms of predictive metrics and the effect of attacks on the underlying point process’s parameters. Expressly, adversarial attacks successfully transform the temporal Hawkes process regime from sub-critical to into a super-critical and manipulate the modeled parameters that is considered a risk against parametric modeling approaches. Additionally, we evaluate the vulnerability and performance of these models in the presence of non-stationary abrupt changes, using the crimes and Covid-19 pandemic dataset as an example. Considering the security vulnerability of deep-learning models, including deep temporal point processes, to adversarial attacks, it is essential to ensure the robustness of the deployed algorithms that is despite the success of deep learning techniques in modeling temporal point processes. In Chapter 5 , we study the robustness of deep temporal point processes against several proposed adversarial attacks from the adversarial defense viewpoint. Specifically, we investigate the effectiveness of adversarial training using universal adversarial samples in improving the robustness of the deep point processes. Additionally, we propose a general point process domain-adopted (GPDA) regularization, which is strictly applicable to temporal point processes, to reduce the effect of adversarial attacks and acquire an empirically robust model. In this approach, unlike other computationally expensive approaches, there is no need for additional back-propagation in the training step, and no further network isrequired. Ultimately, we propose an adversarial detection framework that has been trained in the Generative Adversarial Network (GAN) manner and solely on clean training data. Finally, in Chapter 6 , we discuss implications of the research and future research directions.

Based on an initiative to increase the punctuality on Swedish railways in the short term, this study on potential for increased punctuality by adjustments in the train time schedule on The Western Main Line was initiated. This was done by examining alternatives with different adjustments based on today’s timetable. The goal with the study was to increase the punctuality for the highspeed trains on The Western Main Line by five percentage points. Additionally, the study had a purpose of testing if the Swiss concept of Taktfahrplan could be applicable on a Swedish railway. The adjustments of the timetable were based on three parameters that have impact on the robustness of a timetable; dwell time, allowance and headway between trains. Those were the parameters that were tested. Based on a literature study and data on today’s delays, punctuality and dwell times for the highspeed trains, the current situation could be analyzed. The analysis showed that the trains of today’s timetable departed densely in the mornings, the biggest amounts of trains were near the big cities, punctuality had improved from 2019 to 2020, problematic points with large delays existed around Hallsberg and Falköping and the performed dwell times were longer than the planned ones for every station. Based on the analysis of today’s situation, three case alternatives as well as an alternative for comparison were created. In case alternative 1, dwell times were extended for the highspeed trains stops, in alternative 2, allowance was added and in alternative 3, headway was extended at the start stations. Time schedules for the alternatives were created in RailSys and were then simulated on a selected stretch with a disturbance filter that was based on real distributions of delay. The results did not show considerable difference in punctuality for the case alternatives. However, alternative 1 and 2 showed improvement, while the punctuality was reduced in alternative 3. A fourth alternative was created, where the two cases that had led to improvement were combined. After simulation of alternative 4, it could be established that it was the case that provided the highest improvement in punctuality. A Taktfahrplan (regular departures and station meetings) was made in RailSys that showed how well such a timetable was applicable for the highspeed trains on The Western Main Line. The results turned out well because meeting spots for these trains, at half hour traffic and the selected stopping patterns, occurred most evidently at Katrineholm C, but also at Södertälje syd, Skövde C and Herrljunga. The conclusions of the study were that none of the alternatives led to a 5-percentage increase in punctuality. However, three cases led to increased punctuality, while one case produced decreased punctuality. The parameters that produced the highest increase in punctuality was the combination of dwell time and allowance adjustments. The consequences of the adjustments were longer planned driving times for the trains and lower capacity utilization on the track. Lastly, it was established that Taktfahrplan was applicable for the highspeed trains on The Western Main Line, but further analysis would be required before implementation, especially with consideration to planning for practical introduction.

Recent developments in the field of parallel and distributed computing has led to a proliferation of solving large and computationally intensive mathematical, science, or engineering problems, that consist of several parallelizable parts and several non-parallelizable (sequential) parts. In a parallel and distributed computing environment, the performance goal is to optimize the execution of parallelizable parts of an application on concurrent processors. This requires efficient application scheduling and resource allocation for mapping applications to a set of suitable parallel processors such that the overall performance goal is achieved. However, such computational environments are often prone to unpredictable variations in application (problem and algorithm) and system characteristics. Therefore, a robustness study is required to guarantee a desired level of performance. Given an initial workload, a mapping of applications to resources is considered to be robust if that mapping optimizes execution performance and guarantees a desired level of performance in the presence of unpredictable perturbations at runtime. In this research, a stochastic process algebra, Performance Evaluation Process Algebra (PEPA), is used for obtaining resource allocations via a numerical analysis of performance modeling of the parallel execution of applications on parallel computing resources. The PEPA performance model is translated into an underlying mathematical Markov chain model for obtaining performance measures. Further, a robustness analysis of the allocation techniques is performed for finding a robustmapping from a set of initial mapping schemes. The numerical analysis of the performance models have confirmed similarity with the simulation results of earlier research available in existing literature. When compared to direct experiments and simulations, numerical models and the corresponding analyses are easier to reproduce, do not incur any setup or installation costs, do not impose any prerequisites for learning a simulation framework, and are not limited by the complexity of the underlying infrastructure or simulation libraries.

3D circuit integration is becoming increasingly important as one of the remaining techniques for staying on Moore’s law trajectory. 3D Integrated Circuits (ICs) can be realized using the Through Silicon Via (TSV) approach. In order to extract the full benefits of 3D and for better yield, it has been suggested that the TSVs should be arranged as bundles rather than parallel TSVs. TSVs are required to route the signals through different dies in a multi-tier 3D IC. TSVs are excellent but scarce electrical conductors. Hence, it is important to utilize these resources very efficiently. In high performance 3D ICs, signaling techniques play a crucial role in determining the overall performance of the system. In this work, 3x3 and 4x4 TSV bundles are considered. Electrical parasitics of TSV bundles are extracted using Ansoft Q3D Extractor. Various techniques for signaling over TSV bundles are analyzed in this work. Performance, energy and robustness are the crucial aspects to be considered for analyzing a signaling technique. For performance analysis, maximum data rate for each of the signaling techniques is obtained and the dominant factors that determine these values are identified. 3D integration is fairly a new field and does not have common standards. Different research groups (both academic and industry) across the globe have different manufacturing technologies to suit their needs. In this work, we obtain the electrical parasitics of TSV bundles for different TSV radii ranging from 1mm to 15mm. The TSV radius for most of the 3D integration technologies falls within this range. Maximum data rates are determined for different TSV radii ranging from 1mm to 15mm. This study across different TSV radii helps in choosing a better signaling technique for a particular TSV radius depending on the design goals. Energy/bit for each of the signaling techniques is obtained for a common data rate of 10Gbps Pseudo Random Bit Sequence (PRBS) input. For robustness analysis, the impact of process, voltage and temperature variations between driver and receiver circuits is analyzed. Ansoft Q3D extractor, NCSU 45nm PDK and HSPICE simulation tool are used. From the simulation results, it is observed that a differential technique is beneficial for smaller radii in terms of maximum data rate that can be obtained. For a radius above 7mm, single ended current mode signaling gives a better data rate. Low swing single ended signaling techniques consume less energy but suffer slightly more due to process variations compared to full swing voltage mode signaling. In terms of robustness to supply noise, differential signaling is more robust compared to single ended techniques. An increase in the temperature reduces the data rates of both single ended and differential signaling techniques. Hence, depending on the TSV radius of target technology and process and environment variations, an optimum signaling technique can be chosen.

In the field of image analysis in pattern recognition, shape is an important attribute to characterize graphical objects. It provides important information about an image. In the thesis, I proposed a new descriptor for image identification and classification, named Average Degree Descriptor. We did some experiments and compared its performance with Degree Descriptor. We also analyzed the Average Degree Descriptor theoretically, by comparing the data of distorted shapes and shapes of Kiki/Bouba. Since we also need to classify or identify some 3-dimension shapes in practical application, we proposed an approach to transform 3-dimension shapes to 2-dimension shapes. Moreover, we also studied the robustness of the proposed Average Degree Descriptor in random degradation. Results show that the proposed Average Degree Descriptor has good performance in image identification, even with random degradation.

Visual(-Inertial) Odometry (VIO) and Simultaneous Localization and Mapping (SLAM) are hot topics in Computer Vision today. These technologies have various applications, including robotics, autonomous driving, and virtual reality. They may also be valuable in studying human behavior and navigation through head-mounted visual systems. A complication to SLAM and VIO systems could potentially be visual degeneration such as motion blur. This thesis attempts to evaluate the robustness to motion blur of two open-source state-of-the-art VIO and SLAM systems, namely Delayed Marginalization Visual-Inertial Odometry (DM-VIO) and ORB-SLAM3. There are no real-world benchmark datasets with varying amounts of motion blur today. Instead, a semi-synthetic dataset was created with a dynamic trajectory-based motion blurring technique on an existing dataset, TUM VI. The systems were evaluated in two sensor configurations, Monocular and Monocular-Inertial. The systems are evaluated using the Root Mean Square (RMS) of the Absolute Trajectory Error (ATE).  Based on the findings, the visual input highly influences DM-VIO, and performance decreases substantially as motion blur increases, regardless of the sensor configuration. In the Monocular setup, the performance decline significantly going from centimeter precision to decimeter. The performance is slightly improved using the Monocular-Inertial configuration. ORB-SLAM3 is unaffected by motion blur performing on centimeter precision, and there is no significant difference between the sensor configurations. Nevertheless, a stochastic behavior can be noted in ORB-SLAM3 that can cause some sequences to deviate from this. In total, ORB-SLAM3 outperforms DM-VIO on the all sequences in the semi-synthetic datasets created for this thesis. The code used in this thesis is available at GitHub https://github.com/cmannila along with forked repositories of DM-VIO and ORB-SLAM3 / Visual(-Inertial) Odometry (VIO) och Simultaneous Localization and Mapping (SLAM) är av stort intresse inom datorseende (Computer Vision). Dessa system har en variation av tillämpningar såsom robotik, själv-körande bilar och VR (Virtual Reality). En ytterligare potentiell tillämpning är att integrera SLAM/VIO i huvudmonterade system, såsom glasögon, för att kunna studera beteenden och navigering hos bäraren. En komplikation till SLAM och VIO skulle kunna vara en visuell degration i det visuella systemet såsom rörelseoskärpa. Detta examensarbete försöker utvärdera robustheten mot rörelseoskärpa i två tillgängliga state-of-the-art system, DM-VIO (Delayed Marginalization Visual-Inertial Odometry) och ORB-SLAM3. Idag finns det inga tillgängliga dataset som innehåller specifikt varierande mängder rörelseoskärpa. Således, skapades ett semisyntetiskt dataset baserat på ett redan existerande, vid namn TUM VI. Detta gjordes med en dynamisk rendering av rörelseoskärpa enligt en känd rörelsebana erhållen från datasetet. Med denna teknik kunde olika mängder exponeringstid simuleras.  DM-VIO och ORB-SLAM3 utvärderades med två sensor konfigurationer, Monocular (en kamera) och Monokulär-Inertial (en kamera med Inertial Measurement Unit). Det objektiva mått som användes för att jämföra systemen var Root Mean Square av Absolute Trajectory Error i meter. Resultaten i detta arbete visar på att DM-VIO är i hög-grad beroende av den visuella signalen som används, och prestandan minskar avsevärt när rörelseoskärpan ökar, oavsett sensorkonfiguration. När enbart en kamera (Monocular) används minskar prestandan från centimeterprecision till diameter. ORB-SLAM3 påverkas inte av rörelseoskärpa och presterar med centimeterprecision för alla sekvenser. Det kan heller inte påvisas någon signifikant skillnad mellan sensorkonfigurationerna. Trots detta kan ett stokastiskt beteende i ORB-SLAM3 noteras, detta kan ha orsakat vissa sekvenser att bete sig avvikande. I helhet, ORB-SLAM3 överträffar DM-VIO på alla sekvenser i det semisyntetiska datasetet som skapats för detta arbete. Koden som använts i detta arbete finns tillgängligt på GitHub https://github.com/cmannila tillsammans med forkade repository för DM-VIO och ORB-SLAM3.