Master’s thesis describes theoretical basics, solution design, and implementation of constituency (phrasal) parser for Czech language, which is based on a part of speech association into phrases. Created program works with manually built and annotated Czech sample corpus to generate probabilistic context free grammar within runtime machine learning. Parser implementation, based on extended CKY algorithm, then for the input Czech sentence decides if the sentence can be generated by the created grammar and for the positive cases constructs the most probable derivation tree. This result is then compared with the expected parse to evaluate constituency parser success rate.

This thesis focuses on finding an end-to-end unsupervised solution to solve a two-step problem of extracting semantically meaningful topics and trend analysis of these topics from a large temporal text corpus. To achieve this, the focus is on using the latest develop- ments in Natural Language Processing (NLP) related to pre-trained language models like Google’s Bidirectional Encoder Representations for Transformers (BERT) and other BERT based models. These transformer-based pre-trained language models provide word and sentence embeddings based on the context of the words. The results are then compared with traditional machine learning techniques for topic modeling. This is done to evalu- ate if the quality of topic models has improved and how dependent the techniques are on manually defined model hyperparameters and data preprocessing. These topic models provide a good mechanism for summarizing and organizing a large text corpus and give an overview of how the topics evolve with time. In the context of research publications or scientific journals, such analysis of the corpus can give an overview of research/scientific interest areas and how these interests have evolved over the years. The dataset used for this thesis is research articles and papers from a journal, namely ’Journal of Cleaner Productions’. This journal has more than 24000 research articles at the time of working on this project. We started with implementing Latent Dirichlet Allocation (LDA) topic modeling. In the next step, we implemented LDA along with document clus- tering to get topics within these clusters. This gave us an idea of the dataset and also gave us a benchmark. After having some base results, we explored transformer-based contextual word and sentence embeddings to evaluate if this leads to more meaningful, contextual, and semantic topics. For document clustering, we have used K-means clustering. In this thesis, we also discuss methods to optimally visualize the topics and the trend changes of these topics over the years. Finally, we conclude with a method for leveraging contextual embeddings using BERT and Sentence-BERT to solve this problem and achieve semantically meaningful topics. We also discuss the results from traditional machine learning techniques and their limitations.

News value assessment has been done forever in the news media industry and is today often done in real-time without any documentation. Editors take a lot of different qualitative aspects into consideration when deciding what news stories will make it to the first page. This thesis explores how the complex news value assessment process can be translated into a quantitative model, and also how those news values can be predicted in an effective way using machine learning and NLP. Two models for news value were constructed, for which the correlation between modeled and manual news values was measured, and the results show that the more complex model gives a higher correlation. For prediction, different types of features are extracted, Random Forest and SVM are used, and the predictions are evaluated with accuracy, F1-score, RMSE, and MAE. Random Forest shows the best results for all metrics on all datasets, the best result being on the largest dataset, probably due to the smaller datasets having a less even distribution between classes.

The research of open-domain, knowledge-grounded dialogue systems has been advancing rapidly due to the paradigm shift introduced by large language models (LLMs). While the strides have improved the performance of the dialogue systems, the scope is mostly monolingual and English-centric. The lack of multilingual in-task dialogue data further discourages research in this direction. This thesis explores the use of transfer learning techniques to extend the English-centric dialogue systems to multiple languages. In particular, this work focuses on five typologically diverse languages, of which well-performing models could generalize to the languages that are part of the language family as the target languages, hence widening the accessibility of the systems to speakers of various languages. I propose two approaches: Multilingual Retrieval-Augmented Dialogue Model (xRAD) and Multilingual Generative Dialogue Model (xGenD). xRAD is adopted from a pre-trained multilingual question answering (QA) system and comprises a neural retriever and a multilingual generation model. Prior to the response generation, the retriever fetches relevant knowledge and conditions the retrievals to the generator as part of the dialogue context. This approach can incorporate knowledge into conversational agents, thus improving the factual accuracy of a dialogue model. In addition, xRAD has advantages over xGenD because of its modularity, which allows the fusion of QA and dialogue systems so long as appropriate pre-trained models are employed. On the other hand, xGenD takes advantage of an existing English dialogue model and performs a zero-shot cross-lingual transfer by training sequentially on English dialogue and multilingual QA datasets. Both automated and human evaluation were carried out to measure the models' performance against the machine translation baseline. The result showed that xRAD outperformed xGenD significantly and surpassed the baseline in most metrics, particularly in terms of relevance and engagingness. Whilst xRAD performance was promising to some extent, a detailed analysis revealed that the generated responses were not actually grounded in the retrieved paragraphs. Suggestions were offered to mitigate the issue, which hopefully could lead to significant progress of multilingual knowledge-grounded dialogue systems in the future.

The domain of rhetoric has changed dramatically since its inception as the art of persuasion. It has adapted to encompass many forms of digital media, including, for example, data visualization and coding as a form of literature, but the approach has frequently been that of an outsider looking in. The use of comprehensive computational tools as a part of rhetorical analysis has largely been lacking. In this report, we attempt to address this lack by means of three case studies in natural language processing tasks, all of which can be used as part of a computational approach to rhetoric. At this same moment in time, it is becoming all the more important to transition to renewable energy in order to keep global warming under 1.5 degrees Celsius and ensure that countries meet the conditions of the Paris Agreement. Thus, we make use of speech data on climate policy from the Swedish parliament to ground these three analyses in semantic textual similarity, topic modeling, and political party attribution. We find that speeches are, to a certain extent, consistent within parties, given that a slight majority of most semantically similar speeches come from the same party. We also find that some of the most common topics discussed in these speeches are nuclear energy and the Swedish Green party, purported environmental risks due to renewable energy sources, and the job market. Finally, we find that though pairs of speeches are semantically similar, party rhetoric on the whole is generally not unique enough for speeches to be distinguishable by party. These results then open the door for a broader exploration of computational rhetoric for Swedish political science in the future.

The goal of this project is to investigate any correlation between marketing emails and their receivers using machine learning and only a limited amount of initial data. The data consists of roughly 1200 emails and 98.000 receivers of these. Initially, the emails are grouped together based on their content using text clustering. They contain no information regarding prior labeling or categorization which creates a need for an unsupervised learning approach using solely the raw text based content as data. The project investigates state-of-the-art concepts like bag-of-words for calculating term importance and the gap statistic for determining an optimal number of clusters. The data is vectorized using term frequency - inverse document frequency to determine the importance of terms relative to the document and to all documents combined. An inherit problem of this approach is high dimensionality which is reduced using latent semantic analysis in conjunction with singular value decomposition. Once the resulting clusters have been obtained, the most frequently occurring terms for each cluster are analyzed and compared. Due to the absence of initial labeling an alternative approach is required to evaluate the clusters validity. To do this, the receivers of all emails in each cluster who actively opened an email is collected and investigated. Each receiver have different attributes regarding their purpose of using the service and some personal information. Once gathered and analyzed, conclusions could be drawn that it is possible to find distinguishable connections between the resulting email clusters and their receivers but to a limited extent. The receivers from the same cluster did show similar attributes as each other which were distinguishable from the receivers of other clusters. Hence, the resulting email clusters and their receivers are specific enough to distinguish themselves from each other but too general to handle more detailed information. With more data, this could become a useful tool for determining which users of a service should receive a particular email to increase the conversion rate and thereby reach out to more relevant people based on previous trends. / Målet med detta projekt att undersöka eventuella samband mellan marknadsföringsemail och dess mottagare med hjälp av oövervakad maskininlärning på en brgränsad mängd data. Datan består av ca 1200 email meddelanden med 98.000 mottagare. Initialt så gruperas alla meddelanden baserat på innehåll via text klustering. Meddelandena innehåller ingen information angående tidigare gruppering eller kategorisering vilket skapar ett behov för ett oövervakat tillvägagångssätt för inlärning där enbart det råa textbaserade meddelandet används som indata. Projektet undersöker moderna tekniker så som bag-of-words för att avgöra termers relevans och the gap statistic för att finna ett optimalt antal kluster. Datan vektoriseras med hjälp av term frequency - inverse document frequency för att avgöra relevansen av termer relativt dokumentet samt alla dokument kombinerat. Ett fundamentalt problem som uppstår via detta tillvägagångssätt är hög dimensionalitet, vilket reduceras med latent semantic analysis tillsammans med singular value decomposition. Då alla kluster har erhållits så analyseras de mest förekommande termerna i vardera kluster och jämförs. Eftersom en initial kategorisering av meddelandena saknas så krävs ett alternativt tillvägagångssätt för evaluering av klustrens validitet. För att göra detta så hämtas och analyseras alla mottagare för vardera kluster som öppnat något av dess meddelanden. Mottagarna har olika attribut angående deras syfte med att använda produkten samt personlig information. När de har hämtats och undersökts kan slutsatser dras kring hurvida samband kan hittas. Det finns ett klart samband mellan vardera kluster och dess mottagare, men till viss utsträckning. Mottagarna från samma kluster visade likartade attribut som var urskiljbara gentemot mottagare från andra kluster. Därav kan det sägas att de resulterande klustren samt dess mottagare är specifika nog att urskilja sig från varandra men för generella för att kunna handera mer detaljerad information. Med mer data kan detta bli ett användbart verktyg för att bestämma mottagare av specifika emailutskick för att på sikt kunna öka öppningsfrekvensen och därmed nå ut till mer relevanta mottagare baserat på tidigare resultat.

Work email is one of the organisations most critical tool today. It`s have become a standard way to communicate internally and externally. It can also affect our well-being. Email overload has become a well-known issue for many people. With interviews, follow up interviews, and a workshop, three persons from an accounting firm prioritise pre-define emails needs. And identified several other email needs that were added to the priority list. A thematic analysis and summarizing of a Likert scale was conducted to identify underlying work email needs and work email needs that are not apparent. Three work email needs were selected and using scenario-based methods and the elements of PACT to investigating how the characteristics of a chatbot can help solve the identified work email overload issue? The result shows that email overload is percept different from individual to individual. The choice of how email is handled and email activities indicate how email overload feeling is experienced. The result shows a need to get a sense of the email content quickly, fast collect financial information and information from Swedish authorities, and repetitive, time-consuming tasks. Suggestions on how this problem can be solved have been put forward for many years, and how to use machine learning to help reduce email overload. However, many of these proposed solutions have not yet been implemented on a full scale. One conclusion may be that since email overload is not experienced in the same way, individuals have different needs - One solution does not fit all. With the help of the character of a chatbot, many problems can be solved. And with a technological character of a chatbot that can learn individuals' email patterns, suggest email task to the user and performing tasks to reducing the email overload perception. Using keyword for email intents to get a sense of the email content faster and produce quick links where to find information about the identified subject. And to work preventive give the user remainder and perform repetitive tasks on specific dates.

In the last few years the World Wide Web has dramatically changed the way people are communicating with each other. The growing availability of Social Media Systems like Internet fora, weblogs and social networks ensure that the Internet is today, what it was originally designed for: A technical platform in which all users are able to interact with each other. Nowadays, there are billions of user comments available discussing all aspects of life and the data source is still growing. This thesis investigates, whether it is possible to use this growing amount of freely provided user comments to extract quality related information. The concept is based on the observation that customers are not only posting marketing relevant information. They also publish product oriented content including positive and negative experiences. It is assumed that this information represents a valuable data source for quality analyses: The original voices of the customers promise to specify a more exact and more concrete definition of \"quality\" than the one that is available to manufacturers or market researchers today. However, the huge amount of unstructured user comments makes their evaluation very complex. It is impossible for an analysis protagonist to manually investigate the provided customer feedback. Therefore, Social Media specific algorithms have to be developed to collect, pre-process and finally analyze the data. This has been done by the Social Media monitoring system AIM (Automotive Internet Mining) that is the subject of this thesis. It investigates how manufacturers, products, product features and related opinions are discussed in order to estimate the overall product quality from the customers\\\'' point of view. AIM is able to track different types of data sources using a flexible multi-agent based crawler architecture. In contrast to classical web crawlers, the multi-agent based crawler supports individual crawling policies to minimize the download of irrelevant web pages. In addition, an unsupervised wrapper induction algorithm is introduced to automatically generate content extraction parameters which are specific for the crawled Social Media systems. The extracted user comments are analyzed by different content analysis algorithms to gain a deeper insight into the discussed topics and opinions. Hereby, three different topic types are supported depending on the analysis needs. * The creation of highly reliable analysis results is realized by using a special context-aware taxonomy-based classification system. * Fast ad-hoc analyses are applied on top of classical fulltext search capabilities. * Finally, AIM supports the detection of blind-spots by using a new fuzzified hierarchical clustering algorithm. It generates topical clusters while supporting multiple topics within each user comment. All three topic types are treated in a unified way to enable an analysis protagonist to apply all methods simultaneously and in exchange. The systematically processed user comments are visualized within an easy and flexible interactive analysis frontend. Special abstraction techniques support the investigation of thousands of user comments with minimal time efforts. Hereby, specifically created indices show the relevancy and customer satisfaction of a given topic.:1 Introduction 1.1 Chapter Overview 2 Problem Definition and Data Environment 2.1 Commonly Applied Quality Sensors 2.2 The Growing Importance of Social Media 2.3 Social Media based Quality Experience 2.4 Change to the Holistic Concept of Quality 2.5 Definition of User Generated Content and Social Media 2.6 Social Media Software Architectures 3 Data Collection 3.1 Related Work 3.2 Requirement Analysis 3.3 A Blackboard Crawler Architecture 3.4 Semi-supervised Wrapper Generation 3.5 Structure Modifification Detection 3.6 Conclusion 4 Hierarchical Fuzzy Clustering 4.1 Related Work 4.2 Generalization of Agglomerative Crisp Clustering Algorithms 4.3 Topic Groups Generation 4.4 Evaluation 4.5 Conclusion 5 A Social Media Monitoring System for Quality Analyses 5.1 Related Work 5.2 Pre-Processing Workflow 5.3 Quality Indices 5.4 AIM Architecture 5.5 Evaluation 5.6 Conclusion 6 Conclusion and Perspectives 6.1 Contributions and Conclusions 6.2 Perspectives Bibliography / In den letzten Jahren hat sich das World Wide Web dramatisch verändert. War es vor einigen Jahren noch primär eine Informationsquelle, in der ein kleiner Anteil der Nutzer Inhalte veröffentlichen konnte, so hat sich daraus eine Kommunikationsplattform entwickelt, in der jeder Nutzer aktiv teilnehmen kann. Die dadurch enstehende Datenmenge behandelt jeden Aspekt des täglichen Lebens. So auch Qualitätsthemen. Die Analyse der Daten verspricht Qualitätssicherungsmaßnahmen deutlich zu verbessern. Es können dadurch Themen behandelt werden, die mit klassischen Sensoren schwer zu messen sind. Die systematische und reproduzierbare Analyse von benutzergenerierten Daten erfordert jedoch die Anpassung bestehender Tools sowie die Entwicklung neuer Social-Media spezifischer Algorithmen. Diese Arbeit schafft hierfür ein völlig neues Social Media Monitoring-System, mit dessen Hilfe ein Analyst tausende Benutzerbeiträge mit minimaler Zeitanforderung analysieren kann. Die Anwendung des Systems hat einige Vorteile aufgezeigt, die es ermöglichen, die kundengetriebene Definition von \"Qualität\" zu erkennen.:1 Introduction 1.1 Chapter Overview 2 Problem Definition and Data Environment 2.1 Commonly Applied Quality Sensors 2.2 The Growing Importance of Social Media 2.3 Social Media based Quality Experience 2.4 Change to the Holistic Concept of Quality 2.5 Definition of User Generated Content and Social Media 2.6 Social Media Software Architectures 3 Data Collection 3.1 Related Work 3.2 Requirement Analysis 3.3 A Blackboard Crawler Architecture 3.4 Semi-supervised Wrapper Generation 3.5 Structure Modifification Detection 3.6 Conclusion 4 Hierarchical Fuzzy Clustering 4.1 Related Work 4.2 Generalization of Agglomerative Crisp Clustering Algorithms 4.3 Topic Groups Generation 4.4 Evaluation 4.5 Conclusion 5 A Social Media Monitoring System for Quality Analyses 5.1 Related Work 5.2 Pre-Processing Workflow 5.3 Quality Indices 5.4 AIM Architecture 5.5 Evaluation 5.6 Conclusion 6 Conclusion and Perspectives 6.1 Contributions and Conclusions 6.2 Perspectives Bibliography

This thesis evaluates the performance of different machine learning approaches to log classification based on a dataset derived from simulating intrusive behavior towards an enterprise web application. The first experiment consists of performing attacks towards the web app in correlation with the logs to create a labeled dataset. The second experiment consists of one unsupervised model based on a variational autoencoder and four super- vised models based on both conventional feature-engineering techniques with deep neural networks and embedding-based feature techniques followed by long-short-term memory architectures and convolutional neural networks. With this dataset, the embedding-based approaches performed much better than the conventional one. The autoencoder did not perform well compared to the supervised models. To conclude, embedding-based ap- proaches show promise even on datasets with different characteristics compared to natural language.

Using AI to automatically describe images is a challenging task. The aim of this study has been to compare the use of character-based language models with one of the current state-of-the-art token-based language models, im2txt, to generate image captions, with focus on morphological correctness. Previous work has shown that character-based language models are able to outperform token-based language models in morphologically rich languages. Other studies show that simple multi-layered LSTM-blocks are able to learn to replicate the syntax of its training data. To study the usability of character-based language models an alternative model based on TensorFlow im2txt has been created. The model changes the token-generation architecture into handling character-sized tokens instead of word-sized tokens. The results suggest that a character-based language model could outperform the current token-based language models, although due to time and computing power constraints this study fails to draw a clear conclusion. A problem with one of the methods, subsampling, is discussed. When using the original method on character-sized tokens this method removes characters (including special characters) instead of full words. To solve this issue, a two-phase approach is suggested, where training data first is separated into word-sized tokens where subsampling is performed. The remaining tokens are then separated into character-sized tokens. Future work where the modified subsampling and fine-tuning of the hyperparameters are performed is suggested to gain a clearer conclusion of the performance of character-based language models.