To collaborate effectively, humans and AI agents need to trust each other. Communication between teammates is an essential component to achieve this, as it makes the AI system more understandable to humans. However, previous research lacks a focus on ways to build an AI agent's trust in its human teammate and, consequently, on how the AI's beliefs can be communicated to the human. As such, this study explores how real-time visual explanations of the AI agent's trust in its human teammate influence human trust and overall satisfaction. Through a user experiment (n=46) conducted on an Urban Search and Rescue simulation, integrating trust explanations was compared against a baseline containing no such information. Results show a statistically significant increase in both human trust and satisfaction when the explanations are provided, highlighting the need for further exploration into methods of communicating trust.

This study delves into machine learning (ML) education by conducting a comprehensive literature review, a targeted survey of ML lecturers in Dutch universities, and a comparative experiment. These methods aid in addressing the challenges of aligning teaching methods with the evolving nature of ML and the growing demands of the field, and fill in knowledge gaps on the success of different teaching methods in ML education. The paper investigates whether traditional methods are effective in equipping future engineers with the necessary skills for tomorrow's challenges, amidst the rapid advancement of ML and its applications. The literature review explores the range of teaching methods in ML education and not only, emphasizing a shift towards technology-enhanced and active learning approaches when teaching ML. A survey of ML lecturers explores the landscape of ML education in Dutch universities. The study investigates teaching methodologies, tools, and challenges, providing valuable insights into the evolving practices of ML instruction. Findings indicate a predominant trend towards adopting a blended approach, with lectures, projects, and group work forming core instructional methods. Virtual environments, active learning strategies, and staying informed through community engagement are highlighted. Word frequency and thematic analyses reveal key themes, emphasizing student-centric learning, practical application, and the integration of diverse teaching methods. Additionally, an experimental comparison of two teaching methods, lecture and jigsaw, sheds light on their seemingly similar efficacy when applied to the domain of ML education. The research contributes to the optimization of ML education practices, offering comprehensive insights for educators and policymakers.

This research investigates the impact of goal-oriented visualization on machine learning knowl-edge acquisition, particularly exploring its poten-tial to address procrastination in academic settings. By examining participants with no prior machine learning experience, the study employs compara-tive and correlational measures to analyze quiz per-
formance, study times, and visualization practices. While encouraging trends are observed, the small sample size emphasizes the need for further re-search to definitively establish the impact of visu-alization on academic performance. The study ad-heres to ethical guidelines, ensuring participant pri-vacy and obtaining informed consent, contributing to responsible research practices.

Crowd-powered conversational systems (CPCS) solicit the wisdom of crowds to quickly respond to on-demand users' needs. The very factors that make this a viable solution ---such as the availability of diverse crowd workers on-demand--- also lead to great challenges. The ever-changing pool of online workers powering conversations with individual users makes it particularly difficult to generate contextually consistent responses from a single user's standpoint. To tackle this, prior work has employed conversational facts extracted by workers to maintain a global memory, albeit with limited success. Leveraging systematic context in affective crowdsourcing tasks has remained unexplored. Through a controlled experiment, we explored if a conversational agent, dubbed ContextBot, can provide workers with the required context on the fly for successful completion of affective support tasks in CPCS, and explore the impact of ContextBot on the response quality of workers and their interaction experience. To this end, we recruited workers (N=351) from the Prolific crowdsourcing platform and carried out a 3*3 factorial between-subjects study. Experimental conditions varied based on (i) whether or not context was elicited and informed by motivational interviewing techniques (MI, non-MI, and chat history), and (ii) different conversational entry points for workers to produce responses (early, middle, and late). Our findings showed that workers who entered the conversation earliest were more likely to produce highly consistent responses after interacting with ContextBot. Better user experience from workers was expected after they interacted with ContextBot at a late entry. We found that interacting with ContextBot through task completion did not negatively impact workers' cognitive load.

The GTZAN dataset, a collection of 1000 songsspanning 10 genres, proposed by Tzanetakis hasbeen around for 20 years. In this time hundredsof researches and applications have included thisdatabase. However, there seem to be some seri-ous limitations to this dataset. There are dupli-cates, mislabellings, low audio recordings and nar-row representations of genres. This paper aimsto research the effects of both audio quality andthe content of this dataset on genre classification.A Support Vector Machine (SVM) has been usedto retrain and compare different versions of thedataset. Two experiments have been proposed inthe paper. In the first experiment, a comparison be-tween a lossless dataset of high audio quality andan mp3 version of that same dataset of a loweraudio quality have been investigated. The lowerquality dataset performed worse on the SVM clas-sifier of this size. The second experiment pro-posed a new metal dataset, based on a wider andmore balanced range of metal sub-genres. Thismetal dataset has replaced the original metal partof the GTZAN dataset. Some retrainings done thisway had a higher accuracy than the original, givingconfidence that representing a well-balanced genremight improve classification performance. Finally,it has been found that the original GTZAN classi-fier is inaccurate on audio samples outside of itsdataset, where the new retrainings done on losslessdatasets without much preprocessing seem to per-form substantially better. This last finding has notbeen verified systematically and asks for more ver-ification.

Music Information Retrieval (MIR) is a field of research that focusses on extracting information from music related data. This includes the genre of music and the beats per minute (BPM) of a song. Pipelines that extract this information from music are called feature extractors. Essentia is a library for such feature extraction. Often, the audio codec and quality is not considered in research setups within the field of MIR, while this could have an influence on the results. Therefore the main research question is "How do different audio codecs and audio quality impact genre classification and beats per minute (BPM) recognition in Essentia?". To answer this, the genre has been narrowed down to rock and the chosen audio codecs are FLAC, MP3 LAME and OGG Voribs. In collaboration with Muziekweb, a Dutch music library that collects all music that has been released in The Netherlands, it was possible to gather music files in lossless format. To degrade the audio quality, classify songs and recognize BPM, python pipelines for codec conversion, rock genre classification and BPM recognition were created an ran on this data. It has been concluded that changes in audio codec and quality have an influence on genre classification and BPM recognition in Essentia. It has not been concluded which codec and quality is best to use in the field of MIR. Further research is needed to answer this.

Federated learning (FL) is a new paradigm that allows several parties to train a model together without sharing their proprietary data. This paper investigates vertical federated learning, which addresses scenarios in which collaborating organizations own data from the same set of users but with differing features. The survey provides an overview of how five alternative Vertical Federated
Learning frameworks function, as well as a description of their performance and security assurances. A thorough comparison of how each of the alternatives handles the trade-offs between data privacy, framework performance, and model performance is extracted based on the examined frameworks.
This allows the reader to form an opinion about benefits and disadvantages of various techniques across the vertical federated learning landscape.

Deep learning has made tremendous success in the past decade. As a result, it is becoming widely deployed in various safety and security-critical applications like autonomous driving, malware detection, fingerprint identification, and financial fraud detection. It was recently shown that deep neural networks are susceptible to multiple attacks and pose a serious security threat for such systems. Such an attack is the backdoor attack. In this attack, the model's training is partially or fully outsourced to a malicious party. The adversary designs a secret property, the trigger, and embeds it into a small fraction of the training data. The trained model associates this property with a specific (chosen by the adversary) target label. The backdoored model behaves well on regular inputs, but a misclassification will occur with a very high probability when the trigger is present. The advent of this attack sparked an arms race in the deep learning community resulting in numerous backdoor attacks and defenses during the last years. In this thesis, we conduct a systematic evaluation that can push our knowledge further, aiming at more effective defenses in the future for various applications like image classification, natural language processing, and sound recognition. We show that the trigger's size is positively correlated to the attack success rate in almost all cases. On the contrary, the trigger's position is not always connected to the attack success rate and depends on the used neural network. Furthermore, we are the first to experiment with inaudible triggers in sound recognition and show that they can be used for stealthy real-world attacks. Moreover, we show that backdoor attacks could be a framework that further pushes our knowledge of how deep neural networks learn.
%Most of these attacks focus on computer vision, and only a few experiment with natural language processing and sound recognition. Additionally, most of the defenses are empirical and can be bypassed by slightly altering the attack scenario. Additionally, we exploited global average pooling's properties aiming at a more effective attack. In particular, we created dynamic backdoors that can be effective even if different trigger positions are used during training and inference without poisoning more data or altering the poisoning procedure. The exploration of this layer also showed that when a model generalizes well on new inputs, it is less probable to be susceptible to backdoor attacks.

Most of the content creation applications that are currently common in use are regular PC applications with simulated 3D visualisation on a 2D screen and indirect interaction through a mouse and keyboard. Augmented Reality (AR) is a medium that can provide actual 3D visualisation and more hands-on interaction for these applications. This thesis explores how AR can be used for story authoring, a specific type of content creation. Both types of existing AR interfaces, tangible and touch-less were explored and combined in a useful way. The Story ARtist application was developed to evaluate the designed interactions and AR visualisation for story authoring. Its interface combines tangible and touch-less interactions in the form of physical markers and hand tracking input. A tabletop environment is used to visualise the story authoring elements dynamically, using the 3D space that AR provides. Story authoring is kept simple for accessibility, with a plot point structure focused on core story elements like actions, characters and objects. A user study was done with the concept application to evaluate the designed AR interaction and visualisation for story authoring. The results show that AR has considerable potential for story authoring.

University students are expected to study on their own for large amounts of time. However a lot of these hours are not spent effectively by students. Eventually students who have trouble with self-studying in an effective manner may end up failing courses because of this. When students realise they have a (self-study) problem they can seek aid by contacting the academic counsellor. However, there are problems with this workflow: students often need time to acknowledge they are having issues. Then, even when students contact the academic counsellor to get help, it is a difficult task to provide personalised support for each student.
In this thesis we investigate the feasibility of a self-study support agent that can assist students with feedback on their self-study behavior. This agent can support students next to the academic counsellors. An agent can continuously track students and provide immediate feedback, whereas counsellors have a limited amount of time available (per student). As part of this thesis we conduct a focus group with academic counsellors, organise a workshop with first year students to create a design for a prototype self-study support agent. Thereafter we implement this prototype and use this in an experiment where the activities of several first year students are tracked over the period of two weeks. Then we analyse the data collected with our prototype agent. In doing so we show that the concept of the self-study support agent is feasible. We envision that future work can realise actual deployment of a self-study support agent.

This project focuses on finding what defines an appropriate moment to notify in a smart reminder system. Specifically, the goal is to find a way in which smart reminders systems can be extended through the use of user values to ultimately provide more appropriately timed reminders. This is essential in providing software aided support. A system is designed from scratch, combining existing concepts of activity prediction and value-based design. A statistical Markov chain model is made from predictions based on Expectation Maximization and Apriori algorithms. User values are quantified and optimized following the concept of a Socially Adaptive Electronic Partner and added to the model to identify an appropriate moment for a notification. The concept of values is broken down into two aspects. Firstly, the value loss invoked by the nuisance of receiving a notification during a certain activity. Secondly, the expected value gain achieved by actually remembering is simulated through the expected time between the moment of notification and the deadline. The model is implemented in a Node.js web application, following the principles of a RESTful web API. The model is tested for both its success in correct prediction and the moment selection. The basic predictive model shows a 91% success rate but falls short at 73% when considering values. After optimizing the system for user values, up to a 13% improvement in the success rate and an 18% improvement in the score (more appropriate moment) is found for the model considering user values with respect to the basic, predictive model. Overall, a clear and workable approach to value-based smart reminders is shown through a statistical and dynamic approach to incorporate the concept of user values in a smart-reminder system.