Serverless or Function-as-a-Service (FaaS) is a recent architectural style that is based on the principles of abstracting infrastructure management and scaling to zero, meaning application instances are dynamically started and shut down to accommodate load. This concept of no idling servers and inherent autoscaling comes with benefits but also drawbacks. This study presents an evaluation of the performance and implications of the serverless architecture and contrasts it with the so-called monolith architectures. Three distinct architectures are implemented and deployed on the FaaS platform Microsoft Azure Functions as well as the PaaS platform Azure Web App. Results were produced through experiments measuring cold starts, response times, and scaling for the tested architectures as well as observations of traits such as cost and vendor lock-in. The results indicate that serverless architectures, while it is subjected to drawbacks such as vendor lock-in and cold starts, provides several benefits to a system such as reliability and cost reduction.

Since the popularisation of media streaming, a number of video streaming services are continually buying new video content to mine the potential profit. As such, newly added content has to be handled appropriately to be recommended to suitable users. In this dissertation, the new item cold-start problem is addressed by exploring the potential of various deep learning features to provide video recommendations. The deep learning features investigated include features that capture the visual-appearance, as well as audio and motion information from video content. Different fusion methods are also explored to evaluate how well these feature modalities can be combined to fully exploit the complementary information captured by them. Experiments on a real-world video dataset for movie recommendations show that deep learning features outperform hand crafted features. In particular, it is found that recommendations generated with deep learning audio features and action-centric deep learning features are superior to Mel-frequency cepstral coefficients (MFCC) and state-of-the-art improved dense trajectory (iDT) features. It was also found that the combination of various deep learning features with textual metadata and hand-crafted features provide significant improvement in recommendations, as compared to combining only deep learning and hand-crafted features. / Dissertation (MEng (Computer Engineering))--University of Pretoria, 2021. / The MultiChoice Research Chair of Machine Learning at the University of Pretoria / UP Postgraduate Masters Research bursary / Electrical, Electronic and Computer Engineering / MEng (Computer Engineering) / Unrestricted

E-commerce is an ever growing industry as the internet infrastructure continues to evolve. The benefits from a recommendation system to any online retail store are several. It can help customers to find what they need as well as increase sales by enabling accurate targeted promotions. Among many techniques that can form recommendation systems, this thesis compares Collaborative Filtering against Association Rule Mining, both implemented in combination with clustering. The suggested implementations are designed with the cold start problem in mind and are evaluated with a data set from an online retail store which sells clothing. The results indicate that Collaborative Filtering is the preferable technique while associated rules may still offer business value to stakeholders. However, the strength of the results is undermined by the fact that only a single data set was used. / E-handel är en växande marknad i takt med att Internet utvecklas samtidigt som antalet användare ständigt ökar. Antalet fördelar från rekommendationssytem som e-butiker kan dra nytta av är flera. Samtidigt som det kan hjälpa kunder att hitta vad de letar efter kan det utgöra underlag för riktade kampanjer, något som kan öka försäljning. Det finns många olika tekniker som rekommendationssystem kan vara byggda utifrån. Detta examensarbete ställer fokus på de två teknikerna Collborative Filtering samt Association Rule Mining och jämför dessa sinsemellan. Båda metoderna kombinerades med klustring och utformades för att råda bot på kallstartsproblemet. De två föreslagna implementationerna testades sedan mot en riktig datamängd från en e-butik med kläder i sitt sortiment. Resultaten tyder på att Collborative Filtering är den överlägsna tekniken samtidigt som det fortfarande finns ett värde i associeringsregler. Att dra generella slutsatser försvåras dock av att enbart en datamängd användes.

Recommendation systems have been widely utilized in e-commerce settings to aid users through their shopping experiences. The principal advantage of these systems is their ability to narrow down the purchase options in addition to marketing items to customers. However, a number of challenges remain, notably those related to obtaining a clearer understanding of users, their profiles, and their preferences in terms of purchased items. Specifically, recommender systems based on collaborative filtering recommend items that have been rated by other users with preferences similar to those of the targeted users. Intuitively, the more information and ratings collected about the user, the more accurate are the recommendations such systems suggest. In a typical recommender systems database, the data are sparse. Sparsity occurs when the number of ratings obtained by the users is much lower than the number required to build a prediction model. This usually occurs because of the users’ reluctance to share their reviews, either due to privacy issues or an unwillingness to make the extra effort. Grey-sheep users pose another challenge. These are users who shared their reviews and ratings yet disagree with the majority in the systems. The current state-of-the-art typically treats these users as outliers and removes them from the system. Our goal is to determine whether keeping these users in the system may benefit learning. Thirdly, cold-start problems refer to the scenario whereby a new item or user enters the system and is another area of active research. In this case, the system will have no information about the new user or item, making it problematic to find a correlation with others in the system. This thesis addresses the three above-mentioned research challenges through the development of machine learning methods for use within the recommendation system setting. First, we focus on the label and data sparsity though the development of the Hybrid Cluster analysis and Classification learning (HCC-Learn) framework, combining supervised and unsupervised learning methods. We show that combining classification algorithms such as k-nearest neighbors and ensembles based on feature subspaces with cluster analysis algorithms such as expectation maximization, hierarchical clustering, canopy, k-means, and cascade k-means methods, generally produces high-quality results when applied to benchmark datasets. That is, cluster analysis clearly benefits the learning process, leading to high predictive accuracies for existing users. Second, to address the cold-start problem, we present the Popular Users Personalized Predictions (PUPP-DA) framework. This framework combines cluster analysis and active learning, or so-called user-in-the-loop, to assign new customers to the most appropriate groups in our framework. Based on our findings from the HCC-Learn framework, we employ the expectation maximization soft clustering technique to create our user segmentations in the PUPP-DA framework, and we further incorporate Convolutional Neural Networks into our design. Our results show the benefits of user segmentation based on soft clustering and the use of active learning to improve predictions for new users. Furthermore, our findings show that focusing on frequent or popular users clearly improves classification accuracy. In addition, we demonstrate that deep learning outperforms machine learning techniques, notably resulting in more accurate predictions for individual users. Thirdly, we address the grey-sheep problem in our Grey-sheep One-class Recommendations (GSOR) framework. The existence of grey-sheep users in the system results in a class imbalance whereby the majority of users will belong to one class and a small portion (grey-sheep users) will fall into the minority class. In this framework, we use one-class classification to provide a class structure for the training examples. As a pre-assessment stage, we assess the characteristics of grey-sheep users and study their impact on model accuracy. Next, as mentioned above, we utilize one-class learning, whereby we focus on the majority class to first learn the decision boundary in order to generate prediction lists for the grey-sheep (minority class). Our results indicate that including grey-sheep users in the training step, as opposed to treating them as outliers and removing them prior to learning, has a positive impact on the general predictive accuracy.

This report targets a specific problem for recommender algorithms which is the new item problem and propose a method with sentiment analysis as the main tool. Collaborative filtering algorithms base their predictions on a database with users and their corresponding ratings to items. The new item problem occurs when a new item is introduced in the database because the item has no ratings. The item will therefore be unavailable as a recommendation for the users until it has gathered some ratings. Products that can be rated by users in the online community often has experts that get access to these products before its release date for the consumers, this can be taken advantage of in recommender systems. The experts can be used as initial guides for predictions. The method that is used in this report relies on sentiment analysis to translate written reviews by experts into a rating based on the sentiment of the text. This way when a new item is added it is also added with the ratings of experts in the field. The result from this study shows that the recommender algorithm slope one can generate more reliable recommendations with a group of expert users than without when a new item is added to the database. The expert users that is added must have ratings for other items as well as the ratings for the new item to get more accurate recommendations. / Denna rapport studerar påverkan av problemet med nya objekt i rekommendationsalgoritmen Slope One och en metod föreslås i rapporten för att lösa det specifika problemet. Problemet uppstår när ett nytt objekt läggs till i en databas då det inte finns några betyg som getts till objektet/produkten. Då rekommendationsalgoritmer som Slope One baserar sina rekommendationer på relationerna mellan användares betyg av filmer så blir träffsäkerheten låg för en rekommendation av en film med få betyg. Metoden som föreslås i rapporten involverar attitydanalys som det huvudsakliga verktyget för att få information som kan ersätta faktiska betyg som användare gett en produkt. När produkter kan bli betygsatta av användare på olika forum på internet så finns det ofta experter får tillgång till produkten innan den släpps till omvärlden, den information som dessa experter har kan användas för att fylla det informationsgap som finns när ett nytt objekt läggs till. Dessa experter kommer då initiellt att användas som guide för rekomendationssystemet. Så när ett nytt objekt läggs till så görs det tillsammans med betyg från experter för att få mer träffsäkra rekomendationer. Resultatet från denna studie visar att Slope One genererar mer träffsäkra rekommendationer då en ny produkt läggs till i databasen med ett antal betyg som genererats genom attitydanalysanalys på experters textrecensioner. Det är värt att notera att ett betyg enbart för dessa expertanvändare inte håller utan experterna måste ha betyg av andra produkter inom samma område för kunna influera rekommendationer för den nya produkten.

This dissertation develops a framework for modeling vehicle emissions microscopically. In addition, the framework is utilized to develop the VT-Micro model using a number of data sources. Key input variables to the VT-Micro model include instantaneous vehicle speed and acceleration levels. Estimating accurate mobile source emissions is becoming more and more critical as a result of increasing environmental problems in large metropolitan urban areas. Current emission inventory models, such as MOBILE and EMPAC, are designed for developing large scale inventories, but are unable to estimate emissions from specific corridors and intersections. Alternatively, microscopic emission models are capable of assessing the impact of transportation scenarios and performing project-level analyses. The VT-Micro model was developed using data collected at the Oak Ridge National Laboratory (ORNL) that included fuel consumption and emission rate measurements (CO, HC, and NOx) for five light-duty vehicles (LDVs) and three light-duty trucks (LDTs) as a function of the vehicle's instantaneous speed and acceleration levels. The hybrid regression models predict hot stabilized vehicle fuel consumption and emission rates for LDVs and LDTs. The model is found to be highly accurate compared to the ORNL data with coefficients of determination ranging from 0.92 to 0.99. The study compares fuel consumption and emission results from MOBILE5a, VT-Micro, and CMEM models. The dissertation presents that the proposed VT-Micro model appears to be good enough in terms of absolute light-duty hot stabilized normal vehicle tailpipe emissions. Specifically, the emission estimates were found to be within the 95 percent confidence limits of field data and within the same level of magnitude as the MOBILE5a model estimates. Furthermore, the proposed VT-Micro model was found to reflect differences in drive cycles in a fashion that was consistent with field observations. Specifically, the model accurately captures the increase in emissions for aggressive acceleration drive cycles in comparison with other drive cycles. The dissertation also presents a framework for developing microscopic emission models. The framework develops emission models by aggregating data using vehicle and operational variables. Specifically, statistical techniques for aggregating vehicles into homogenous categories are utilized as part of the framework. In addition, the framework accounts for temporal lags between vehicle operational variables and vehicle emissions. Finally, the framework is utilized to develop the VT-Micro model version 2.0 utilizing second-by-second chassis dynamometer emission data for a total of 60 light duty vehicles and trucks. Also, the dissertation introduces a procedure for estimating second-by-second high emitter emissions. This research initially investigates high emitter emission cut-points to verify clear definitions of high emitter vehicles (HEVs) and derives multiplicative factors for newly developed EPA driving cycles. Same model structure with the VT-Micro model is utilized to estimate instantaneous emissions for a total of 36 light duty vehicles and trucks. Finally, the dissertation develops a microscopic framework for estimating instantaneous vehicle start emissions for LDVs and LDTs. The framework assumes a linear decay in instantaneous start emissions over a 200-second time horizon. The initial vehicle start emission rate is computed based on MOBILE6's soak time function assuming a 200-second decay time interval. The validity of the model was demonstrated using independent trips that involved cold start and hot start impacts with vehicle emissions estimated to within 10 percent of the field data. The ultimate expansion of this model is its implementation within a microscopic traffic simulation environment in order to evaluate the environmental impacts of alternative ITS and non-ITS strategies. Also, the model can be applied to estimate vehicle emissions using instantaneous GPS speed measurements. Currently, the VT-Micro model has been implemented in the INTEGRATION software for the environmental assessment of operational-level transportation projects. / Ph. D.

No description available.

Sistemas de recomendação fazem parte do nosso dia-a-dia. Os métodos usados nesses sistemas tem como objetivo principal predizer as preferências por novos itens baseado no perĄl do usuário. As pesquisas relacionadas a esse tópico procuram entre outras coisas tratar o problema do cold-start do usuário, que é o desaĄo de recomendar itens para usuários que possuem poucos ou nenhum registro de preferências no sistema. Uma forma de tratar o cold-start do usuário é buscar inferir as preferências dos usuários a partir de informações adicionais. Dessa forma, informações adicionais de diferentes tipos podem ser exploradas nas pesquisas. Alguns estudos usam informação social combinada com preferências dos usuários, outros se baseiam nos clicks ao navegar por sites Web, informação de localização geográĄca, percepção visual, informação de contexto, etc. A abordagem típica desses sistemas é usar informação adicional para construir um modelo de predição para cada usuário. Além desse processo ser mais complexo, para usuários full cold-start (sem preferências identiĄcadas pelo sistema) em particular, a maioria dos sistemas de recomendação apresentam um baixo desempenho. O trabalho aqui apresentado, por outro lado, propõe que novos usuários receberão recomendações mais acuradas de modelos de predição que já existem no sistema. Nesta tese foram propostas 4 abordagens para lidar com o problema de cold-start do usuário usando modelos existentes nos sistemas de recomendação. As abordagens apresentadas trataram os seguintes aspectos: o Inclusão de informação social em sistemas de recomendação tradicional: foram investigados os papéis de várias métricas sociais em um sistema de recomendação de preferências pairwise fornecendo subsidíos para a deĄnição de um framework geral para incluir informação social em abordagens tradicionais. o Uso de similaridade por percepção visual: usando a similaridade por percepção visual foram inferidas redes, conectando usuários similares, para serem usadas na seleção de modelos de predição para novos usuários. o Análise dos benefícios de um framework geral para incluir informação de redes de usuários em sistemas de recomendação: representando diferentes tipos de informação adicional como uma rede de usuários, foi investigado como as redes de usuários podem ser incluídas nos sistemas de recomendação de maneira a beneĄciar a recomendação para usuários cold-start. o Análise do impacto da seleção de modelos de predição para usuários cold-start: a última abordagem proposta considerou que sem a informação adicional o sistema poderia recomendar para novos usuários fazendo a troca entre os modelos já existentes no sistema e procurando aprender qual seria o mais adequado para a recomendação. As abordagens propostas foram avaliadas em termos da qualidade da predição e da qualidade do ranking em banco de dados reais e de diferentes domínios. Os resultados obtidos demonstraram que as abordagens propostas atingiram melhores resultados que os métodos do estado da arte. / Recommender systems are in our everyday life. The recommendation methods have as main purpose to predict preferences for new items based on userŠs past preferences. The research related to this topic seeks among other things to discuss user cold-start problem, which is the challenge of recommending to users with few or no preferences records. One way to address cold-start issues is to infer the missing data relying on side information. Side information of different types has been explored in researches. Some studies use social information combined with usersŠ preferences, others user click behavior, location-based information, userŠs visual perception, contextual information, etc. The typical approach is to use side information to build one prediction model for each cold user. Due to the inherent complexity of this prediction process, for full cold-start user in particular, the performance of most recommender systems falls a great deal. We, rather, propose that cold users are best served by models already built in system. In this thesis we propose 4 approaches to deal with user cold-start problem using existing models available for analysis in the recommender systems. We cover the follow aspects: o Embedding social information into traditional recommender systems: We investigate the role of several social metrics on pairwise preference recommendations and provide the Ąrst steps towards a general framework to incorporate social information in traditional approaches. o Improving recommendation with visual perception similarities: We extract networks connecting users with similar visual perception and use them to come up with prediction models that maximize the information gained from cold users. o Analyzing the beneĄts of general framework to incorporate networked information into recommender systems: Representing different types of side information as a user network, we investigated how to incorporate networked information into recommender systems to understand the beneĄts of it in the context of cold user recommendation. o Analyzing the impact of prediction model selection for cold users: The last proposal consider that without side information the system will recommend to cold users based on the switch of models already built in system. We evaluated the proposed approaches in terms of prediction quality and ranking quality in real-world datasets under different recommendation domains. The experiments showed that our approaches achieve better results than the comparison methods. / Tese (Doutorado)

While the car fleet becomes increasingly larger it is important to lower the amounts of pollutants from each individual diesel or gasoline engine to almost zero levels. The pollutants from these engines predominantly originate from high NOx emissions and particulates, in the case when diesel is utilized, and emissions at cold start from gasoline engines. One way of treating the high NOx levels is to introduce ammonia in the diesel exhausts and let it react with the NOx to form nitrogen gas and water, which is called SCR (Selective Catalytic Reduction). However, in order to make this system reduce NOx efficiently enough for meeting future legislations, closed loop control is required. To realize this type of system an NOx or ammonia sensor is needed. The cold start emissions from gasoline vehicles are primarily due to a high light-off time for the catalytic converter. Another reason is the inability to quickly heat the sensor used for controlling the air-to-fuel ratio in the exhausts, also called the lambda value, which is required to be in a particular range for the catalytic converter to work properly. This problem may be solved utilizing another, more robust sensor for this purpose. This thesis presents the efforts made to test the SiC-based field effect transistor (SiC-FET) sensor technology both as an ammonia sensor for SCR systems and as a cold start lambda sensor. The SiC-FET sensor has been shown to be highly sensitive to ammonia both in laboratory and engine measurements. As a lambda sensor it has proven to be both sensitive and selective, and its properties have been studied in lambda stairs both in engine exhausts and in the laboratory. The influence of metal gate restructuring on the linearity of the sensor has also been investigated. The speed of response for both sensor types has been found to be fast enough for closed loop control in each application. / <p>On the day of the public defence of the doctoral thesis, the status of article III was: in press. Report code: LiU-Tek-Lic-2003:50.</p>

<p>While the car fleet becomes increasingly larger it is important to lower the amounts of pollutants from each individual diesel or gasoline engine to almost zero levels. The pollutants from these engines predominantly originate from high NO_x emissions and particulates, in the case when diesel is utilized, and emissions at cold start from gasoline engines. One way of treating the high NO_x levels is to introduce ammonia in the diesel exhausts and let it react with the NO_x to form nitrogen gas and water, which is called SCR (Selective Catalytic Reduction). However, in order to make this system reduce NO_x efficiently enough for meeting future legislations, closed loop control is required. To realize this type of system an NO_x or ammonia sensor is needed. The cold start emissions from gasoline vehicles are primarily due to a high light-off time for the catalytic converter. Another reason is the inability to quickly heat the sensor used for controlling the air-to-fuel ratio in the exhausts, also called the lambda value, which is required to be in a particular range for the catalytic converter to work properly. This problem may be solved utilizing another, more robust sensor for this purpose.</p><p>This thesis presents the efforts made to test the SiC-based field effect transistor (SiC-FET) sensor technology both as an ammonia sensor for SCR systems and as a cold start lambda sensor. The SiC-FET sensor has been shown to be highly sensitive to ammonia both in laboratory and engine measurements. As a lambda sensor it has proven to be both sensitive and selective, and its properties have been studied in lambda stairs both in engine exhausts and in the laboratory. The influence of metal gate restructuring on the linearity of the sensor has also been investigated. The speed of response for both sensor types has been found to be fast enough for closed loop control in each application.</p> / On the day of the public defence of the doctoral thesis, the status of article III was: in press. Report code: LiU-Tek-Lic-2003:50.