| 1. |
- Aagerup, Ulf, 1969-
(författare)
-
The Impact of User Weight on Brands and Business Practices in Mass Market Fashion
- 2010
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Overweight people claim to be mistreated by the fashion industry. If they were, it would be in line with branding theory supporting the idea of rejecting fat consumers to improve user imagery for fashion brands. However, fashion companies do not confess to such practices.To shed some light on the subject, I have conducted two studies.The first attempts to illustrate what effect, if any, user imagery has on fashion brands. It is an experiment designed to show how the weight of users affects consumers’ perceptions of mass market fashion brands. The findings show that consumers’ impressions of mass market fashion brands are significantly affected by the weight of its users. The effect of male user imagery is ambiguous. For women’s fashion on the other hand, slender users are to be preferred.In the second study I examine what effects these effects have on assortments. I compare the sizes of mass market clothes to the body sizes of the population. No evidence of discrimination of overweight or obese consumers was found -quite the contrary.The reasons for these unexpected findings may be explained by the requirements a brand must fulfil to make management of the customer base for user imagery purposes viable. The brand must be sensitive to user imagery; a requirement that mass market fashion fulfils. However, it must also be feasible for a company to exclude customers, and while garment sizes can be restricted to achieve this, the high volume sales strategy of mass market fashion apparently cannot.
|
|
| 2. |
- Agelis, Sacki
(författare)
-
Reconfigurable Optical Interconnection Networks for High-Performance Embedded
- 2005
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- In embedded computer and communication system the capacity demand for interconnection networks is increasing continuously in order to achieve high-performance systems. Recent breakthroughs show that by using reconfigurability inside a single chip substantial performance gains can be added. However, in this thesis the focus is on system level reconfigurability (between chips or modules) and the performance gains that potentially can be achieved by having support for runtime reconfigurability on the system level.This thesis addresses the field of runtime system level reconfigurability with the use of optics in switches and routers for data- and telecommunications, and in multi-processor systems used for embedded signal processing. Several reconfigurable systems for switching and routing with support to adapt for asymmetric traffic patterns are proposed and compared to identify how design choices affect flexibility, performance etc. The proposed solutions are characterized by their multistage optical interconnection networks with reconfigurable shuffle patterns, where the reconfigurability is provided by micro-optical-electrical mechanical systems. More specifically, application-specific bottlenecks can be resolved by reconfiguring the interconnection network according to the current application demands. The benefits of the architectural solutions are confirmed by simulations that clearly show that the architectures can achieve high performance for both symmetric application characteristics and for several classes of asymmetric application characteristics. The final architectural solution is characterized by electronic packet-switches interconnected through an optical backplane, which is reconfigurable. Moreover, the thesis presents how several signal processing applications can be mapped to run concurrently in a time-shared scheme on a single reconfigurable multi-processor system that has high flexibility to adapt for the application currently at hand. The interconnection network is then adapted (reconfigured) according to the demands of the currently executed application in each time instance. The analysis shows that it is feasible to build such a system with today’s components.
|
|
| 3. |
- Alabdallah, Abdallah, 1979-
(författare)
-
Machine Learning Survival Models : Performance and Explainability
- 2023
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Survival analysis is an essential statistics and machine learning field in various critical applications like medical research and predictive maintenance. In these domains understanding models' predictions is paramount. While machine learning techniques are increasingly applied to enhance the predictive performance of survival models, they simultaneously sacrifice transparency and explainability. Survival models, in contrast to regular machine learning models, predict functions rather than point estimates like regression and classification models. This creates a challenge regarding explaining such models using the known off-the-shelf machine learning explanation techniques, like Shapley Values, Counterfactual examples, and others. Censoring is also a major issue in survival analysis where the target time variable is not fully observed for all subjects. Moreover, in predictive maintenance settings, recorded events do not always map to actual failures, where some components could be replaced because it is considered faulty or about to fail in the future based on an expert's opinion. Censoring and noisy labels create problems in terms of modeling and evaluation that require to be addressed during the development and evaluation of the survival models.Considering the challenges in survival modeling and the differences from regular machine learning models, this thesis aims to bridge this gap by facilitating the use of machine learning explanation methods to produce plausible and actionable explanations for survival models. It also aims to enhance survival modeling and evaluation revealing a better insight into the differences among the compared survival models.In this thesis, we propose two methods for explaining survival models which rely on discovering survival patterns in the model's predictions that group the studied subjects into significantly different survival groups. Each pattern reflects a specific survival behavior common to all the subjects in their respective group. We utilize these patterns to explain the predictions of the studied model in two ways. In the first, we employ a classification proxy model that can capture the relationship between the descriptive features of subjects and the learned survival patterns. Explaining such a proxy model using Shapley Values provides insights into the feature attribution of belonging to a specific survival pattern. In the second method, we addressed the "what if?" question by generating plausible and actionable counterfactual examples that would change the predicted pattern of the studied subject. Such counterfactual examples provide insights into actionable changes required to enhance the survivability of subjects.We also propose a variational-inference-based generative model for estimating the time-to-event distribution. The model relies on a regression-based loss function with the ability to handle censored cases. It also relies on sampling for estimating the conditional probability of event times. Moreover, we propose a decomposition of the C-index into a weighted harmonic average of two quantities, the concordance among the observed events and the concordance between observed and censored cases. These two quantities, weighted by a factor representing the balance between the two, can reveal differences between survival models previously unseen using only the total Concordance index. This can give insight into the performances of different models and their relation to the characteristics of the studied data.Finally, as part of enhancing survival modeling, we propose an algorithm that can correct erroneous event labels in predictive maintenance time-to-event data. we adopt an expectation-maximization-like approach utilizing a genetic algorithm to find better labels that would maximize the survival model's performance. Over iteration, the algorithm builds confidence about events' assignments which improves the search in the following iterations until convergence.We performed experiments on real and synthetic data showing that our proposed methods enhance the performance in survival modeling and can reveal the underlying factors contributing to the explainability of survival models' behavior and performance.
|
|
| 4. |
- Ali Hamad, Rebeen, 1989-
(författare)
-
Towards Reliable, Stable and Fast Learning for Smart Home Activity Recognition
- 2022
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The current population age grows increasingly in industrialized societies and calls for more intelligent tools to monitor human activities. The aims of these intelligent tools are often to support senior people in their homes, to keep track of their daily activities, and to early detect potential health problems to facilitate a long and independent life. The recent advancements of smart environments using miniaturized sensors and wireless communications have facilitated unobtrusively human activity recognition. Human activity recognition has been an active field of research due to its broad applications in different areas such as healthcare and smart home monitoring. This thesis project develops work on machine learning systems to improve the understanding of human activity patterns in smart home environments. One of the contributions of this research is to process and share information across multiple smart homes to reduce the learning time, reduce the need and effort to recollect the training data, as well as increase the accuracy for applications such as activity recognition. To achieve that, several contributions are presented to pave the way to transfer knowledge among smart homes that includes the following studies. Firstly, a method to align manifolds is proposed to facilitate transfer learning. Secondly, we propose a method to further improve the performance of activity recognition over the existing methods. Moreover, we explore imbalanced class problems in human activity recognition and propose a method to handle imbalanced human activities. The summary of these studies are provided below. In our work, it is hypothesized that aligning learned low-dimensional manifolds from disparate datasets could be used to transfer knowledge between different but related datasets. The t-distributed Stochastic Neighbor Embedding(t-SNE) is used to project the high-dimensional input dataset into low-dimensional manifolds. However, since t-SNE is a stochastic algorithm and there is a large variance of t-SNE maps, a thorough analysis of the stability is required before applying Transfer learning. In response to this, an extension to Local Procrustes Analysis called Normalized Local Procrustes Analysis (NLPA) is proposed to non-linearly align manifolds by using locally linear mappings to test the stability of t-SNE low-dimensional manifolds. Experiments show that the disparity from using NLPA to align low-dimensional manifolds decreases by order of magnitude compared to the disparity obtained by Procrustes Analysis (PA). NLPA outperforms PA and provides much better alignments for the low-dimensional manifolds. This indicates that t-SNE low-dimensional manifolds are locally stable, which is the part of the contribution in this thesis.Human activity recognition in smart homes shows satisfying recognition results using existing methods. Often these methods process sensor readings that precede the evaluation time (where the decision is made) to evaluate and deliver real-time human activity recognition. However, there are several critical situations, such as diagnosing people with dementia where "preceding sensor activations" are not always sufficient to accurately recognize the resident's daily activities in each evaluated time. To improve performance, we propose a method that delays the recognition process to include some sensor activations that occur after the point in time where the decision needs to be made. For this, the proposed method uses multiple incremental fuzzy temporal windows to extract features from both preceding and some oncoming sensor activations. The proposed method is evaluated with two temporal deep learning models: one-dimensional convolutional neural network (1D CNN) and long short-term memory (LSTM) on a binary sensor dataset of real daily living activities. The experimental evaluation shows that the proposed method achieves significantly better results than the previous state-of-the-art. Further, one of the main problems of activity recognition in a smart home setting is that the frequency and duration of human activities are intrinsically imbalanced. The huge difference in the number of observations for the categories means that many machine learning algorithms focus on the classification of the majority examples due to their increased prior probability while ignoring or misclassifying minority examples. This thesis explores well-known class imbalance approaches (synthetic minority over-sampling technique, cost-sensitive learning and ensemble learning) applied to activity recognition data with two temporal data pre-processing for the deep learning models LSTM and 1D CNN. This thesis proposes a data level perspective combined with a temporal window technique to handle imbalanced human activities from smart homes in order to make the learning algorithms more sensitive to the minority class. The experimental results indicate that handling imbalanced human activities from the data-level outperforms algorithm level and improved the classification performance.
|
|
| 5. |
- Aramrattana, Maytheewat
(författare)
-
Modelling and Simulation for Evaluation of Cooperative Intelligent Transport System Functions
- 2016
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Future vehicles are expected to be equipped with wireless communication technology, that enables them to be “connected” to each others and road infrastructures. Complementing current autonomous vehicles and automated driving systems, the wireless communication allows the vehicles to interact, cooperate, and be aware of its surroundings beyond their own sensors’ range. Such sys- tems are often referred to as Cooperative Intelligent Transport Systems (C-ITS), which aims to provide extra safety, efficiency, and sustainability to transporta- tion systems. Several C-ITS applications are under development and will require thorough testing and evaluation before their deployment in the real-world. C- ITS depend on several sub-systems, which increase their complexity, and makes them difficult to evaluate.Simulations are often used to evaluate many different automotive applications, including C-ITS. Although they have been used extensively, simulation tools dedicated to determine all aspects of C-ITS are rare, especially human factors aspects, which are often ignored. The majority of the simulation tools for C-ITS rely heavily on different combinations of network and traffic simulators. The human factors issues have been covered in only a few C-ITS simulation tools, that involve a driving simulator. Therefore, in this thesis, a C-ITS simulation framework that combines driving, network, and traffic simulators is presented. The simulation framework is able to evaluate C-ITS applications from three perspectives; a) human driver; b) wireless communication; and c) traffic systems.Cooperative Adaptive Cruise Control (CACC) and its applications are chosen as the first set of C-ITS functions to be evaluated. Example scenarios from CACC and platoon merging applications are presented, and used as test cases for the simulation framework, as well as to elaborate potential usages of it. Moreover, approaches, results, and challenges from composing the simulation framework are presented and discussed. The results shows the usefulness of the proposed simulation framework.
|
|
| 6. |
- Ashfaq, Awais, 1990-
(författare)
-
Predicting clinical outcomes via machine learning on electronic health records
- 2019
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The rising complexity in healthcare, exacerbated by an ageing population, results in ineffective decision-making leading to detrimental effects on care quality and escalates care costs. Consequently, there is a need for smart decision support systems that can empower clinician's to make better informed care decisions. Decisions, which are not only based on general clinical knowledge and personal experience, but also rest on personalised and precise insights about future patient outcomes. A promising approach is to leverage the ongoing digitization of healthcare that generates unprecedented amounts of clinical data stored in Electronic Health Records (EHRs) and couple it with modern Machine Learning (ML) toolset for clinical decision support, and simultaneously, expand the evidence base of medicine. As promising as it sounds, assimilating complete clinical data that provides a rich perspective of the patient's health state comes with a multitude of data-science challenges that impede efficient learning of ML models. This thesis primarily focuses on learning comprehensive patient representations from EHRs. The key challenges of heterogeneity and temporality in EHR data are addressed using human-derived features appended to contextual embeddings of clinical concepts and Long-Short-Term-Memory networks, respectively. The developed models are empirically evaluated in the context of predicting adverse clinical outcomes such as mortality or hospital readmissions. We also present evidence that, surprisingly, different ML models primarily designed for non-EHR analysis (like language processing and time-series prediction) can be combined and adapted into a single framework to efficiently represent EHR data and predict patient outcomes.
|
|
| 7. |
- Assabie Lake, Yaregal, 1975-
(författare)
-
Multifont recognition System for Ethiopic Script
- 2006
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- In this thesis, we present a general framework for multi-font, multi-size and multi-style Ethiopic character recognition system. We propose structural and syntactic techniques for recognition of Ethiopic characters where the graphically comnplex characters are represented by less complex primitive structures and their spatial interrelationships. For each Ethiopic character, the primitive structures and their spatial interrelationships form a unique set of patterns.The interrelationships of primitives are represented by a special tree structure which resembles a binary search tree in the sense that it groups child nodes as left and right, and keeps the spatial position of primitives in orderly manner. For a better computational efficiency, the primitive tree is converted into string pattern using in-order traversal, which generates a base of the alphabet that stores possibly occuring string patterns for each character. The recognition of characters is then achieved by matching the generated patterns with each pattern in a stored knowledge base of characters.Structural features are extracted using direction field tensor, which is also used for character segmentation. In general, the recognition system does not need size normalization, thinning or other preprocessing procedures. The only parameter that needs to be adjusted during the recognition process is the size of Gaussian window which should be chosen optimally in relation to font sizes. We also constructed an Ethiopic Document Image Database (EDIDB) from real life documents and the recognition system is tested with respect to variations in font type, size, style, document skewness and document type. Experimental results are reported.
|
|
| 8. |
- Averfalk, Helge, 1988-
(författare)
-
Enhanced District Heating Technology : Maintaining Future System Feasibility
- 2017
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- When heat demand and high temperature heat supply gradually decreases in the future, then it will effect district heating systems ability to compete on the heat market. A good way to mitigate less district heating feasibility is to operate systems with lower temperature levels and the most conceivable way to achieve lower temperature levels is to decrease return temperatures.Thus, this thesis emphasise temperature errors embedded in district heating systems. Only a selection of temperature errors are analysed in this thesis. First, the temperature error that occurs due to recirculation in distribution networks at low heat demands. Second, the temperature error that occurs due to hot water circulation in multi-family buildings. Third, the temperature error that occurs due to less than possible heat transfer in heat exchangers, i.e. too short thermal lengths.In order to address these temperature errors three technology changes have been proposed (i) three-pipe distribution network to separate the recirculation return flow from the delivery return flow, (ii) apartment substations to eliminate hot water circulation utilisation, and (iii) improved heat exchangers for lower return temperatures at a constant scenario. Analysis of proposed changes has resulted in annual average return temperatures between 17-21 °C.Furthermore, rapid introduction of intermittent renewable electricity supply in the energy system has prompted an increased necessity of power system balancing capacities. Large-scale conversion of power-to-heat in electric boilers and heat pumps is a feasible alternative to achieve such balancing capacities. Analysis of the unique Swedish experience with utilisation of large heat pumps installations connected to district heating systems show that since the 1980s 1527 MW of heat power has been installed, about 80 % of the capacity was still in use by 2013. Thus, a cumulative value of over three decades of operation and maintenance exists within Swedish district heating systems.The two papers presented in this thesis are related to future district heating systems through the five abilities of fourth generation district heating (4GDH), which are documented in the definition paper of 4GDH.
|
|
| 9. |
- Bengtsson, Jerker
(författare)
-
Efficient implementation of stream applications on processor arrays
- 2006
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This thesis concludes work conducted on exploring the usage of parallel and reconfigurable processor architectures in industrial high-performance embedded systems. This kind of systems has by tradition been built using a mix of digital signal processors and custom made hardware. Digital signal processors provide full functional felxibility, but at the cost of lower performance. Custom made hardware can be optimized for specific functions for high performance, but at the cost of inflexibility and high development costs. A desire is to combine flexibility and performance using commercial hardware, without trading too much of performance for flexibility.Parallel and reconfigurable architectures provide a flexible computing space constituting processing elements that are coupled through configurable communication structures. Architectures designed with less complex processing elements render a high degree of utilizable parallelism at the cost of having to use a portion of the pocessing elements for control functions. In the thesis it is shown that it is possible to utilize this kind of architecture to achieve high performance efficiency, despite the fact that a large fraction of PEs are required to implement control-oriented portions in a fairly complex algorithm.A major problem is that architectures of this kind expose a very complex programming abstraction for compilers and programmers. The approach taken in this work is a domain-specific stream processing model which provides means to express application-specific dataflows and computations in terms of streams. An extensive application study comprising the baseband processing in radio base stations has been used to define sufficient data types, operators and language construct. Furthermore, to support industrial requirements on portability to different architectures, it must be possible to express parallelism and characteristic computations without exposing of hardware details in the source code.To be able to prototype and set up experiments with stream processing languages an experimental programming framework has been developed. A first prototype language with specific primitive types, operators and stream constructs has been implemented in order to elaborate with baseband programming. It is demonstrated how these types and operators can be used to express machine-independent bit field and other fine-grained data parallel computations. Furthermore, the language has been designed with constructs for efficient and flexible programming of reconfiguration of distributed function parameters.
|
|
| 10. |
- Bergenhem, Carl
(författare)
-
Protocols with Heterogeneous Real-Time Services for High-Performance Embedded Networks
- 2002
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Network protocols for applications that demand high performance and heterogeneous real-time services are presented. These protocols control the medium access to the network and offer additional features to the user, both different user services for traffic and services for parallel and distributed real-time processing. The network architecture assumed is a unidirectional pipelined optical ring.Radar Signal Processing (RSP) is a typical application area. Such a system contains many computation nodes that are interconnected in order to co-operate and thereby achieve higher performance. The computing performance of the whole system is greatly affected by the choice of network. Computing nodes in a parallel computer for RSP should be tightly coupled, i.e., communications cost (e.g. latency) between nodes should be small, so that the whole system can be perceived as a single unit. This is possible if a suitable network with an efficient protocol is used.There is an industrial need for new high-performance networks with support for the, often heterogeneous, real-time requirements found in (often embedded) applications such as RSP and other areas such as multimedia. The traffic this kind of network can be classified according to its requirements. The proposed protocols partition the traffic into three classes with distinctly different qualities. These classes are traffic with hard real-time demands, such as mission critical commands, traffic with soft real-time demands, such as process data (a deadline miss here only leads to decreased performance) and, finally, traffic with no real-time constraints at all. The contributions of the present thesis are protocols that integrate heterogeneous real-time services for the three traffic classes.The performance of the proposed protocols is evaluated through simulations and analysis. It is shown that the protocol is an efficient choice for RSP systems. A brief survey of related technologies is included in the thesis. These are studied from the perspectives of application, architecture and user service.
|
|
