Painting therapy has become an important approach to the treatment of depression in adolescents. However, depressed adolescents face low self-confidence, low engagement, and self-absorption when undergoing painting therapy. Informed by need-finding investigations, We developed Realms of the Heart, a human-AI collaborative gamified complementary treatment system for depressed adolescents. Realms of the Heart includes two AI-driven strategies: the transformation of natural objects collected by teens’ exploration into brushes, and the generation of sketches corresponding to natural objects by using brushes based on ControlNet. We conducted a randomized controlled trial. Participants were randomized into an experimental group and a control group. The results showed that Realms of the Heart increased users’ motivation and significantly alleviated their level of depression. These findings not only shed light on the role of human-AI co-created painting therapy in the treatment of adolescent depression but also pave the way for more informed design strategies for art therapy.

Olfactory stimuli have demonstrated the potential to evoke emotional depth and enhance user experiences in HCI. Yet, their role in shaping perceptions of social robots remains largely untapped. This study investigates how olfactory (scent) and auditory (voice) stimuli influence user perceptions of social robots. Using a 2x2 between-subjects design, participants interacted with a social robot under conditions with pleasant/unpleasant scents and friendly/unfriendly voices. The study measured perceived trust, friendliness, competence, and engagement. Our findings show that pleasant scents can enhance the perceptions of friendliness and engagement, while friendly voices can improve trust, friendliness, and engagement. The congruent combination of scents and voices affects friendliness and engagement but does not influence trust and competence. This study contributes to the growing work on multi-sensory Human-Robot Interaction (HRI) design, offering implications for creating more socially interactive robots.

Due to the limited internal space and economic costs, virtual keyboards(VKs) of mobile devices commonly use monotonous global sound and vibration feedback, resulting in unsatisfactory realism and negative experience. To mitigate the monotonous feedback of high-frequency keystrokes on VKs, we propose an interactive sound-vibrotactile feedback design. This system dynamically adjusts sound and vibrotactile re-sponses based on the user's typing speed. We conducted a user study (N=30) and found that the collaborative varied auditory-vibrotactile feedback provided a better experience, created a local feedback illusion, and improved global feedback. Additionally, users preferred subtle variations over distinct variations. Our method can improve global VK feedback and provide a better experience without increasing additional costs and burdens. This study contributes to improving the VK experience on low-cost mobile devices and provides design suggestions for VKs.

Selecting appropriate fonts for sentence-level text, such as slide titles and poster headings, requires pleasing glyphs and semantic congruence with the text. This task is especially challenging for users without design backgrounds. Previous studies established associations between tags and fonts but overlooked the different design intents between sentence-level text and word-level text influencing font selections. Recently, large language models (LLMs) like ChatGPT with extensive knowledge bases have emerged as promising solutions for various tasks. This study evaluated the capabilities of GPT-3.5 in recommending fonts compared to professional designers to explore the application of LLMs in the design field. We conducted comprehensive evaluations, including both quantitative metrics to assess font diversity, accuracy, and consistency, and a user study involving designers and non-designers. The results indicate that while GPT-3.5 performs well in selecting common serif and sans-serif fonts suitable for general use that align with the preferences of users without design backgrounds, it falls short in stylized font selection. Our findings highlight the potential and limitations of using GPT-3.5 for font selection in collaborative design environments and provide insights into integrating such models into design tools to support multidisciplinary teams.

This study aimed to detect stressful hazard perception scenarios subjectively and objectively when using intelligent driving systems. We used electrooculography (EOG), electroencephalography (EEG), subjective ratings, and interviews to identify potential stressful hazard perceptions and record improvements in an intelligent navigation-guided pilot (NGP) system. Moreover, we analyzed electrophysiological data. Our study contributes to the use of engagement, concentration, and phase locking value connectivity based on EEG to support previous research methodologies using beta power, pupil size, fixation ratio, fixation duration, and subjective evaluations for investigating hazard perception. Our analyses showed that stressful hazard perception scenarios occurred mainly when encountering broken and solid lines, frequent lane changes, cars approaching suddenly, several cars driving in parallel, and the decision to change lanes but immediately pulling back upon using the NPG system. Our findings shed light on obtaining accurate results based on subjective and objective evaluations for developing intelligent driving systems.

In the digital era, the significance of design education is on the rise due to its ability to cultivate creativity. However, the disconnect between design practice and theory, coupled with the abundance of design knowledge, poses challenges to learning in this field. Despite the potential of large language models (LLMs) to integrate various data sources for facilitating design knowledge dissemination, they face obstacles such as the scarcity of design-related datasets and limited natural language representations. To overcome these challenges, we introduce DRELM, a design-centric language model that serves as an assistant providing UI design recommendations. We also offer corresponding resources to advance language modeling research in the design domain. Importantly, DesignInstruct stands out as a premier dataset for guiding user interface tasks, while DesignEvaluation significantly contributes to autonomous design evaluation and decision support. In our research, we utilize supervised data from DesignInstruct and DesignEvaluation to fine-tune pre-trained Qwen-7B models for design tasks. Experiments conducted on test data affirm the effectiveness of our dataset in enhancing knowledge comprehension, design execution, and evaluation. We commit to making all training data and DRELM models at https://github.com/sssala/DRELM-A-Language-Model-for-Design-Recommendation-and-Evaluation.

Ovarian cancer presents significant well-being challenges for middle-aged and older women. Recent research underscores the vital role of recovery identity in predicting wellbeing. However, a research gap exists regarding the influence of online support groups (OSPs) on identity synthesis for middle-aged and older cancer patients. This study introduces "Rehab-Diary," a mobile age-friendly OSP grounded in The Social Identity Model of Identity Change, aimed at helping ovarian cancer patients foster recovery identity. A four-week randomized controlled trial involving 68 participants assessed the OSP's impact. The interface was tailored for ease of use by older individuals. The findings demonstrate the feasibility of utilizing Rehab- Diary among older individuals. The intervention effectively enhanced recovery identity. This study offers evidence-based insights for developing future age-friendly online support interventions, ultimately enhancing ovarian cancer patients' quality of care.