Correctly processing accumulated information is beneficial for our survival. Berghman and Hekkert (2017) argue that this is why we humans derive pleasure from having a sense of aesthetics. These aesthetic experiences can be seen as our brain’s reward system for correctly perceiving and interpret- ing the world around us. While our senses have evolved to perceive and organise the physical world, these very mechanisms also come into play when we interact with the digital realm. Aesthetics in visual training datasets are of importance as it allows us to derive a sense of aesthetic pleasure from digital media. Integrating aesthetics into artificial intelligence, especially in text-to-image generators, becomes important to cater to humans psychological reward systems and to engage them at a deeper level.

This thesis is focused on investigating the annotation method used in the development of the LAION- Aesthetics V2 datasets and comparing it to other annotation methods for measuring aesthetics. The purpose is to explore whether there are more suitable alternatives to the current annotation method (where people are asked to annotate images with the instruction ”how much do you like this image on a scale from 1 to 10?”, (Schuhmann, 2022) which is not backed by literature to actually measure aesthetics), and to evaluate the alignment between the LAION Aesthetics Predictor scores and human ratings.

This thesis explores different distinct levels of inquiry: one focuses on the design of instructions for image annotation tasks (alternative task design), while the other centers around measuring aesthetics during the annotation process (alternative metrics). Both lines of inquiry are supported by relevant literature, indicating their potential capacity to capture aesthetics. In addition to comparing alternative annotation methods, this thesis investigates three hypotheses related to the annotation of aesthetics within the project’s context.

Four experiments are conducted using crowdsourcing to compare alternative task design and alternative metrics. The experiments include semantic concept activation, different phrasing of the annotation instruction, and alternative modalities (such as ranking and two-alternative forced choice). Next to these four experiments, a separate fifth experiment is deployed which looks into the evaluation of image content versus overall image liking. Two post hoc analyses are performed, one which compares scores that the LAION Aesthetics predictor assigns to the stimulus set to human image liking ratings, and one examining the influence of region on image liking ratings.

The LAION aesthetics approach performed equal to the alternatives with scientific backing. The ranking treatment even performed worse. For this data, region did not impact image liking ratings. No significant difference was found between participants’ overall image liking and content liking. The LAION Aesthetics predictor scores partially aligned with human liking ratings but showed some disparities, par- ticularly in extreme ratings. Qualitative analysis suggests that more research is necessary to make a judgement on whether ”liking” is a relevant and appropriate approach for capturing aesthetics.

The limitations of the experiments include small sample sizes and the focus on a specific image class (buildings). Recommendations for future research include exploring different image classes, investigating other ranking modalities, and considering n-alternative forced choice experiments. It is also suggested to examine the influence of regions on aesthetic experiences in more detail, explore Gibbs Sampling with People for measuring image aesthetics, and explore different demographic groups and contexts.

Orthopedic surgeries are identified as one of the most noisy surgeries inside the OR (operation room). The highest noise sound level could reach up to around 130 dB, which is harmful to patients’ well-being and are likely to evoke negative feelings during surgery. In the previous study investigating the emotional experience of patients who receive spinal or general anesthesia in orthopedic surgeries, results showed that different kinds of sounds inside the OR could lead to fluctuating emotions for patients, such as nervousness, curiosity, and relaxation. Therefore, this project is mainly focusing on this user group and aims to improve their experience, especially regarding sounds inside the OR.

The research phase of the report includes two parts. The research in LUMC mapped out the patient journey, for instance, when, and how the design should intervene. Questionnaires for surgeons provided strong support for the rationale and effectiveness of musical interventions in this design project. Later on, the second part of the research, the one on AI music alignment proved that to some extent, AI generative model could understand the textual input with emotional descriptors, and output desired emotionally effective music that could influence the listeners’ perception. With this in mind, the design could refer to the circumplex model that Hevner and Schubert developed on music emotions to position the descriptors of specific musical emotions in two dimensions and achieve control of musical emotions in spatial terms.

The design concept mainly has two phases, the preference input phase, which allows patients to choose their favorite prompt list for the AI model to generate later on adding the emotional descriptors together as prompts; the music experience phase on the other hand, mainly happens during surgery when inside the OR, patients will be able to freely explore the music emotion variations during surgery with their favorite music styles (prompts). The design goal for the project would be to design a music companion product for orthopedic surgery patients who are undergoing spinal anesthesia. The product interaction aims to improve their emotional experience during the surgery period. The interaction should be distracting, desirable, and immersive for patients.

Once the general direction of the design concept is determined, the iteration phase began. The prompt input part is based on the brainstorming session and the co-creation session to develop an ideal interface for the experience. While for the music experience part, one of the three interaction ways will be chosen through the concept testing and evaluation procedure. The chosen concept will be further developed into the final concept - Music · Wave.

In the future, I imagine that the OR would be an immersive sensing space for the patients as well, so adding visuals into the whole design would be another vision, that maybe will be further designed in later projects. This project, however, would only give a rough glance into what the whole experience could be like in the future, and how the surgery soundscape could be reframed.

Vehicle automation will increasingly release car drivers from driving tasks, allowing them to engage in previously inconceivable activities. Experiential components are therefore expected to become central in human automotive mobility. In this context, the design of future in-vehicle experiences is a research gap to still discover, that concerns both academia and industry alike.

This graduation project explored the opportunities of using robotics and AI technology for the provision of meaningful autonomous in-vehicle experiences. In that aim, different user interfaces (UIs) were analyzed, as a way to visualize and study different user-technology interactions. The outcomes of the research comprise recommendations about promising scenarios that could be included in autonomous vehicles as well as insights into how different UIs shape user experiences differently.

The project first reviewed prior academic work on the topics of fundamental needs (from positive design), vehicle automation, user experience design, robotics, AI, and user interfaces. A research approach was proposed based on that analysis; because robotics and AI present ample capabilities, in-vehicle scenarios should be designed first, to later define (through UIs) the role that technology should have in those scenarios. These are the research questions that were proposed:

1. What in-vehicle scenarios can be designed to support
fundamental needs?
• How do different user interfaces enable those use cases?
• How do different types of user interfaces affect the in-vehicle user experiences?

2. What are the most promising scenario and user interface
combinations?

To answer those questions, a co-creative workshop was designed, with the aim of collecting user needs and perceptions as data. The participants were asked to envision future needs in AVs, based on the typology of 13 fundamental needs, as well as to design meaningful scenarios that would fulfill them. Through a questionnaire, they voted for the most promising scenarios (i.e., most attractive and most innovative scenarios). The workshop was conducted three times and a total of 18 participants were recruited.

The preliminary results from the workshop were 13 meaningful scenarios, each of them aimed at fulfilling one fundamental need; additionally, those scenarios were adapted to ambient, graphical, and tangible UIs. Besides, promising UI and scenario combinations were identified, according to the participant‘s perspective. Finally, insights were clustered on how different UIs shape the in-vehicle user experience. Those outcomes were contrasted and enriched in validation sessions, where a total of six field experts contributed. The experts brought in insights from the HCI, future mobility, and positive design fields.

This research aims to contribute to both industrial and academic practices. First, fundamental needs and sub-needs have been explored in the context of autonomous vehicles. Second, design directions and examples are offered for the development of meaningful in-vehicle experiences. Additionally, conclusions on how users perceive tangible, graphical, and ambient UIs are given, which could be used as guidelines for designing interactions. Finally, insights are offered about how end-users perceived innovations, as well as how innovative and attractive solutions are differently framed.

Finally, additional research paths were revealed through the study, that future research may consider. For instance, activity-based fundamental need hierarchies could vary across different demographic groups. Apart from that, further work could be done in the classification of the UIs, as well as in mapping out the interactions that they enabled. Finally, in-vehicle scenarios could be related to the concept of ’innovation adoption’ to study what solutions to develop further in the coming years.

Despite the fact that climate change is becoming increasingly dangerous and prevalent, there is still a lack of public engagement. This can be explained by the fact that the media portrays climate change as an abstract concept. The message can be more effectively communicated through visual art because it is more likely to invoke emotional responses in individuals. By including human perception and rating data, the generative adversarial neural network (GAN) produces better image output. Therefore, this paper explores methods for using the human perception of beauty in order to improve StyleGAN2 outputs. In GANAesthetic, UI sliders allow users to explore satellite images interactively, that is, visually appealing satellite images generated from StyleGAN2. The GANAesthetic was determined to be the most appropriate methodology for the study. The choice of GANAesthetic over other approaches will be explained in this paper, as well as its implementation. The paper will also describe an experiment to discover aesthetic latent dimensions.

Generative Adversarial Networks (GANs) brought rapid developments in generating synthetic images by mimicking structures in the training data. With the list of application of GANs growing drastically, it has lately become an exciting technology to explore for designers to communicate their ideas and arts through technology and create engaging experiences for humans. Nevertheless, translating human experiences to artificial intelligence and creating visually pleasant imagery is a challenging task due to complex semantics of human perception. To address this issue, we introduce an iterative training approach in which the generated images are curated by humans and the most pleasing ones are fed back into the network to retrain. Additionally, we do a factorial analysis to investigate how the aesthetic quality and the diversity are affected by the size of training data and training length. In experiments, we validate that this method can significantly improve the aesthetic quality of generated images regardless of the dataset size and training length, however the use of smaller datasets comes with a cost of reduction in the image diversity and novelty in the output images. The aesthetic bias towards certain contexts can also deteriorate the diversity and affect the model evaluations. On the other hand, no significant relationship has been found regarding the training length, however this could possibly be due to instabilities that happen during the model convergence progress.

This paper aims to evaluate which automated measures of aesthetic beauty are the best predictors for human ratings of aesthetics and proposes that typicality and novelty may increase the correlation between the two. To study the correlation between these metrics, a literature study was performed to find a select amount of potentially good predictors, a pipeline was created to extract these values from each image within our dataset, a survey was conducted to vote for which images were considered most aesthetic, and finally regression analysis was performed to see which metrics offered highest correlation with the human rating data. From this we could see there were indeed a number of automated metrics that consistently scored high as predictors for the human aesthetic ratings and there was a slight improvement in the fit of the prediction model upon including novelty as a feature. However, at this moment, the improvement is not significant to conclude these features are better at predicting human ratings.

Generative Adversarial Networks (GANs) can create artwork images and we need effective ways of rating their aesthetic values. This could help us determine the most aesthetic artwork images (and identify the GANs that created them) and train GANs to produce more aesthetic artwork images in the future. In this research, we analyzed the effectiveness of using two different survey formats (binary-choice and four-choice) for displaying GAN-produced artwork images to crowd- sourced workers and gathering their ratings. The artwork images were of different landscapes like the desert, arctic, coastal regions, etc. Additionally, we investigated how the choice of showing different images together (image groupings) per question affects the final rating results. Results demonstrate that the four-choice format is superior to the binary-choice format in producing more consistent, reliable, and accurate results. The effects of the different image groupings were insignif- icant for the results of the four-choice format. In contrast, different image groupings displayed statistically significant changes in the results for the binary-choice format. However, it was found that crowdsourced workers preferred the binary-choice format more as they found it to be less strenuous and more effective in allowing them to express their rating choices.

With the COVID-19 pandemic testing humanity worldwide in unforeseen ways, wellbeing assessment has stepped to the foreground of individual health status. Conversational User Interfaces (CUIs) prove promising as an assessment tool, but lacking the means to retain users engaged during the process. This research aims to explore a solution introducing a graphical modality to a CUI, that improves user experience during wellbeing assessment. A psychological projective testing technique was researched, the H-T-P test, which exposes one's personality, mood and feelings through drawing. A system was designed that utilizes a simplified version of the test and image classification to provide a new means of wellbeing assessment. The system's performance and usability show that such an approach is indeed feasible. This paper describes the design process of the system, the resulting prototype and possible improvements for future implementations.

Due to COVID-19 the overall wellbeing worldwide decreased. Assessing and improving wellbeing became a more important subject. This article describes the design research that uses the My Wellness Check survey created by the Delft University for Technology and aims to create a dashboard for wellbeing. That includes a way of authenticating users to very sensitive data. Also, finding ways how to improve personal wellbeing by using a recommendation system based on different types of filtering and the additional elements that are needed for a recommendation system.

The increase in global internet users brings forth a vast amount of social media users and therefore opinions that are shared online. A subset of those users, adolescents, seem to develop some sort of addiction towards social media, which could lead to low life satisfaction. This paper tries to extract trending topics and their relation to well-being in order to help organizations like MyWellnessCheck check in on adolescents and students. The results indicated that this was possible, despite the vast amount of spam that is present online. Unsurprisingly, current events made the list of trending topics with negative sentiment like ”school shootings”, as well as unexpected topics with positive sentiment that could potentially im- prove well-being, like ”safe snacks”.

A decrease in wellbeing worldwide due to the COVID-19 pandemic called for ways to assess wellbeing in a scalable and adequate manner. Conversational User Interfaces (CUIs) seem suitable, however, applying them optimally in certain contexts remains a challenge. This study aims to find ways to make CUIs more engaging and have a better experience by making them adaptive. A 3x2 between-subjects experiment is designed in which the effects of avatar presence, gender, and an empathic conversational style are researched. A chatbot was created in telegram, and the visual design and conversational style were altered to measure the effects on Questionnaire Experience (QX), Enjoyment, and Empathy. In total 30 participants chatted with a randomly assigned chatbot and filled in a survey about their experiences. There is no statistical preference for avatar presence or conversational style. Male gendered chatbots score higher on QX, but female chatbots are perceived as more empathic when comparing gender.

With the surge of mental health issues during COVID-19, more emphasis has turned towards assessing wellbeing. At the same time, recent advances in AI have shown huge potential in a variety of fields. However, few solutions are available at the intersection of those two fields. This research explores how the use of transformer models like GPT-3 could have a positive impact in the domain of wellbeing and proposes a solution to automatize the survey question creation process. After comparing several GPT-3 question creation methods, it was found that through clever prompt engineering and added context, it can be possible to generate syntactically and contextually correct questions about any specific wellbeing context. In addition, the paper discusses potential ways to assess such questions and offers a demo for a question generation web application.

In the Netherlands, the healthcare sector is facing increasing staff shortages and the demand for adequately trained healthcare personnel is expected to increase in the coming years. Shortages in obstetric care mean that not all women and their partners receive the care they need before, during and after pregnancy. To counter these shortages, there are opportunities in technology and continuity in maternity care.

Erasmus MC has set up its own (online) care path for preconception care, Smarter Pregnant. This programme consists of online platforms for preconception screening and the outpatient clinic healthy pregnant (OCHP). OCHP is a counselling session for couples or individuals to make pregnancy as healthy as possible or to obtain the healthiest possible lifestyle to get pregnant. Most of the pain points occur when filling in and processing data in the Slimmer Zwanger's system. The current system of Smarter Pregnant is often not understood by patients and nurses have to process the data manually. This processing of data is perceived by nurses as monotonous - which is why nurses report that the work feels robotic. Moreover, there is less focus on delivering care and more on interacting with the system. As a result, patients have to search for more information on the internet, getting lost in the validated and non-validated e-health for conception. Therefore, the need for an authoritative and trustworthy source of information for the OCHP is great.

The specific focus of this study is on creating a trustworthy VA for OCHP. The dimensions of trustworthiness are leading for the interaction and thus the design of the Voice Assistant. There are three layers that were considered when creating the Voice Assistant for the OCHP; the user, the context and the technology. In the current context, trustworthiness is generated by taking responsibility and exuding expertise. Responsibility is something the nurse has to take on the word of the patient and partner. The trustworthiness of the nurse can be measured by his or her expertise. Expertise is considered a dimension of trustworthiness in existing models.

To decide what the VA had to look and sound like, several experimental prototypes were conducted during the context analysis and collection of existing theories and methods. These prototypes served both as validation of research, and to spawn further research. The results of these studies led to a final prototype.

Each test ended with a qualitative interview with the patient based on the dimensions of trustworthiness. To structure patients' responses to the VA, three research questions were formulated regarding; expectations, comparing human-to-human versus human-to-robot interaction and characteristics affecting trustworthiness. Because the context analysis revealed that in the context of OCHP, trustworthiness depends on expertise, this research question was reformulated specifically regarding expertise.

These themes helped accomplish and structure the overarching goal of the project: Forming principles on how trustworthy VAs can be designed for healthcare in the future. These principles form a set of pointers on how trustworthiness can be implemented in a VA for healthcare.

It would be hard to think of someone who isn’t benefiting from their social circles and social interactions within them. The many and varied social interactions we all experience are intrinsic to nurturing and maintaining our social connections, which are arguably one of the core capabilities to which we can attribute our dominance as a species. In parallel, we have our capacity to develop technologies. Which goes from the first tools and deliberately making of fire to antibiotics and the world wide web. All of which undoubtedly elevated our quality of life in general.Still, in spite of all these advancements, we seem to struggle with the pursuit of well-being - to achieve a balance between our capacities and the challenges we face. To be in a state of satisfaction while still driven and stimulated. And to flourish.One can say that today we are more connected to one another than ever before, with hundreds of people one touch away in our powerful smartphones and computers. Yet we seem to feel more isolated than ever (Coombs, 2020). Even more now with the challenges imposed by the Covid-19 pandemic, literally isolating us even more.It shows that simply having the means to connect isn’t enough. It is necessary to understand what motivates us on a behavioural level, what makes us engage and maintain quality social interactions?This graduation project sets out to explore the experience of university students, in the Netherlands, during the Covid-19 pandemic. Aiming to be informed by the unique set of circumstances we face today, in the hope to design better digital solutions for our future. The project has the goal of assisting students in a remote study situation to be able to build and nourish their networks and personal relationships. The project’s approach is focused on qualitative data gathered in context mapping activities (Sanders & Stappers, 2012), combined with literature studies and fitting within a double diamond approach(TU Delft, 2019). In order to design for sustained well-being, the project follows the multi-stage framework for sustained wellbeing promoted by technology proposed by Wiese, Pohlmeyer & Hekkert,(2020).In order to understand the target group’s experiences, we have looked into their routines during the lockdown and collected their recollections of their previous university experiences. With the insights from the research, a problem statement was formed, leading to a design direction.The design direction was focused on stimulating and facilitating the engagement with positive activities related to the nurturing of personal relationships (Wiese, Pohlmeyer & Hekkert, 2020), exploring how to reinstate the drivers of behaviour necessary for the engagement with such activities in a virtual scenario.The ideation phase explored 10 different ways of virtual social interactions, which were later exposed to the target group for feedback and combined into one final concept: Atrium, a close-knit environment to feel connected and interact with your peers in digital university life.Atrium is a well-being informed digital platform. Designed to lower the friction in initiating social interactions between university peers. It creates an environment that conveys belonging and keeps you connected to the community, motivating students in their daily work and enabling moments for spontaneous and more natural online social interactions. Atrium means a communal space to be together and naturally connect, not only for tech-savvy and extroverted personalities but for all students.