Remini: Leveraging Chatbot-Mediated Mutual Reminiscence for Promoting Positive Affect and Feeling of Connectedness among Loved OnesMutual reminiscence, defined as revisiting shared positive memories through reciprocal self-disclosure, strengthens emotional bonds, enhances well-being, and deepens intimacy. However, most technology-mediated reminiscence tools emphasize individual reflection or one-way storytelling, which overlooks the dynamic, interactive dialogue essential for meaningful mutual reminiscence. To address this limitation, we introduce Remini, a chatbot designed to support reciprocal self-disclosure between close partners such as couples, friends, or family members. Grounded in the Social Functions of Autobiographical Memory (SFAM) framework, Remini uses conversational AI to guide emotionally rich exchanges through five narrative phases: rapport building, memory narration, elaboration, reflection, and summary. In a mixed-method, both between- and within-subjects study (N = 48, 24 dyads), we compare Remini to a baseline chatbot that offers minimal memory-trigger prompts. Our findings show that structured guidance from Remini significantly improves positive affect, feeling of connection, and engagement. It also fosters more detailed narrative co-construction and greater reciprocal self-disclosure. Participant feedback highlights the practical value, perceived benefits, and design considerations of chatbot-mediated reminiscence. We contribute empirically grounded design implications for conversational agents that strengthen human connection through mutual reminiscence.2025ZJZhuoqun Jiang et al.Connecting FamiliesCSCW
Enhancing Deliberativeness: Evaluating the Impact of Multimodal Reflection NudgesNudging participants with text-based reflective nudges enhances deliberation quality on online deliberation platforms. The effectiveness of multimodal reflective nudges, however, remains largely unexplored. Given the multi-sensory nature of human perception, incorporating diverse modalities into self-reflection mechanisms has the potential to better support various reflective styles. This paper explores how presenting reflective nudges of different types (direct: persona and indirect: storytelling) in different modalities (text, image, video and audio) affects deliberation quality. We conducted two user studies with 20 and 200 participants respectively. The first study identifies the preferred modality for each type of reflective nudges, revealing that text is most preferred for persona and video is most preferred for storytelling. The second study assesses the impact of these modalities on deliberation quality. Our findings reveal distinct effects associated with each modality, providing valuable insights for developing more inclusive and effective online deliberation platforms.2025SYShunYi Yeo et al.Singapore University of Technology and DesignParticipatory DesignInteractive Narrative & Immersive StorytellingCHI
Understanding End-User Perception of Transfer Risks in Smart ContractsBlockchain smart contracts are increasingly used in critical use cases (e.g., financial transactions). Thus, it is pertinent to ensure that their end-users understand risks in attempting token transfers. Addressing this, we investigate end-user comprehension of five transfer risks (e.g. the end-user being blacklisted) in the most popular Ethereum contract, USD Tether (USDT), and their prevalence in other top ERC-20 contracts. First, we conducted a user study investigating end-user comprehension of transfer risks in USDT with 110 participants. Second, we performed source code analysis of the next top (78) ERC-20 smart contracts to identify the prevalence of these risks. Study results show that the majority of end-users do not comprehend some real risks, and confuse real and fictitious risks. This holds regardless of participants’ self-rated programming and Web3 proficiency. Source code analysis demonstrates that examined risks are prevalent in up to 19.2% of the top ERC-20 contracts.2025YPYustynn Panicker et al.Singapore University of Technology and DesignPrivacy by Design & User ControlPrivacy Perception & Decision-MakingCHI
Help Me Reflect: Leveraging Self-Reflection Interface Nudges to Enhance Deliberativeness on Online Deliberation PlatformsThe deliberative potential of online platforms has been widely examined. However, little is known about how various interface-based reflection nudges impact the quality of deliberation. This paper presents two user studies with 12 and 120 participants, respectively, to investigate the impacts of different reflective nudges on the quality of deliberation. In the first study, we examined five distinct reflective nudges: persona, temporal prompts, analogies and metaphors, cultural prompts and storytelling. Persona, temporal prompts, and storytelling emerged as the preferred nudges for implementation on online deliberation platforms. In the second study, we assess the impacts of these preferred reflectors more thoroughly. Results revealed a significant positive impact of these reflectors on deliberative quality. Specifically, persona promotes a deliberative environment for balanced and opinionated viewpoints while temporal prompts promote more individualised viewpoints. Our findings suggest that the choice of reflectors can significantly influence the dynamics and shape the nature of online discussions.2024SYShunYi Yeo et al.Singapore University of Technology and DesignSocial Platform Design & User BehaviorParticipatory DesignCHI
CollabCoder: A Lower-barrier, Rigorous Workflow for Inductive Collaborative Qualitative Analysis with Large Language ModelsCollaborative Qualitative Analysis (CQA) can enhance qualitative analysis rigor and depth by incorporating varied viewpoints. Nevertheless, ensuring a rigorous CQA procedure itself can be both complex and costly. To lower this bar, we take a theoretical perspective to design a one-stop, end-to-end workflow, CollabCoder, that integrates Large Language Models (LLMs) into key inductive CQA stages. In the independent open coding phase, CollabCoder offers AI-generated code suggestions and records decision-making data. During the iterative discussion phase, it promotes mutual understanding by sharing this data within the coding team and using quantitative metrics to identify coding (dis)agreements, aiding in consensus-building. In the codebook development phase, CollabCoder provides primary code group suggestions, lightening the workload of developing a codebook from scratch. A 16-user evaluation confirmed the effectiveness of CollabCoder, demonstrating its advantages over the existing CQA platform. All related materials of CollabCoder, including code and further extensions, will be included in: https://gaojie058.github.io/CollabCoder/.2024JGJie Gao et al.Singapore University of Technology and DesignHuman-LLM CollaborationUser Research Methods (Interviews, Surveys, Observation)Prototyping & User TestingCHI
Understanding Pedestrians’ Perception of Safety and Safe Mobility PracticesWalking is one of the greenest and most common travel modes. However, evidence shows a trend of decreased walking, and safety is a key barrier preventing many people from walking. Additionally, there is a limited understanding of pedestrians’ safe mobility practices and safety perception. Drawing on 449 survey responses from a representative sample in the United Kingdom, our work highlights how identities and walking situations intersect with individuals’ safety perceptions and diverse practices of pedestrians’ safe mobility. The role of technology used for negotiating safety and current challenges in both safe route planning and walking are also highlighted. Our work extends existing insights into pedestrians’ perception of safety and practices by adding empirical evidence and more nuanced contexts. This paper proposes two implications for design in response to design opportunities that surfaced from our mixed-method data analysis. Both the contributions and limitations of our work are also discussed.2024MZHaimo Zhang et al.The Open UniversityV2X (Vehicle-to-Everything) Communication DesignPedestrian & Cyclist SafetyCHI
DataDive: Supporting Readers' Contextualization of Statistical Statements with Data ExplorationStatistical statements that refer to data to support narratives or claims are commonly used to inform readers about the magnitude of social issues. While contextualizing statistical statements with relevant data supports readers in building their own interpretation of statements, the complexity of finding contextual information on the web and linking statistical statements with it impedes readers' efforts to do so. We present DataDive, an interactive tool for contextualizing statistical statements for the readers of online texts. Based on users' selections of statistical statements, our tool uses an LLM-powered pipeline to generate candidates of relevant contexts and poses them as guiding questions to the user as potential contexts for exploration. When the user selects a question, DataDive employs visualizations to further help the user compare and explore contextually relevant data. A technical evaluation shows that DataDive generates important and diverse questions that facilitate exploration around statistical statements and retrieves relevant data for comparison. Moreover, a user study with 21 participants suggests that DataDive facilitates users to explore diverse contexts and to be more aware of how statistical data could relate to the text.2024HKHyunwoo Kim et al.Interactive Data VisualizationData StorytellingVisualization Perception & CognitionIUI
GlassMessaging: Towards Ubiquitous Messaging Using OHMDshttps://doi.org/10.1145/36109312023NJNuwan Janaka et al.Context-Aware ComputingUbiquitous ComputingUbiComp
VISAR: A Human-AI Argumentative Writing Assistant with Visual Programming and Rapid Draft PrototypingIn argumentative writing, writers must brainstorm hierarchical writing goals, ensure the persuasiveness of their arguments, and revise and organize their plans through drafting. Recent advances in large language models (LLMs) have made interactive text generation through a chat interface (e.g., ChatGPT) possible. However, this approach often neglects implicit writing context and user intent, lacks support for user control and autonomy, and provides limited assistance for sensemaking and revising writing plans. To address these challenges, we introduce VISAR, an AI-enabled writing assistant system designed to help writers brainstorm and revise hierarchical goals within their writing context, organize argument structures through synchronized text editing and visual programming, and enhance persuasiveness with argumentation spark recommendations. VISAR allows users to explore, experiment with, and validate their writing plans using automatic draft prototyping. A controlled lab study confirmed the usability and effectiveness of VISAR in facilitating the argumentative writing planning process.2023ZZZheng Zhang et al.Human-LLM CollaborationAI-Assisted Creative WritingUIST
Magical Brush: A Symbol-Based Modern Chinese Painting System for NovicesModern Chinese painting is a new type of painting inherited from ancient Chinese painting. Drawing modern Chinese painting is time-consuming and laborious, which is difficult for novices to start. Symbols are fundamental components of Chinese cultural works both materially and mentally. We introduce a symbol-based modern Chinese painting system termed Magical Brush. Magical Brush combines symbolic cultural factors with AI generative models, with the attempt to help novices create a complete modern Chinese painting, learn basic ideas of Chinese paintings and obtain co-creation engagement. In user study, we compare Magical Brush to other AI and non-AI digital painting tools. Results indicate that by combining cultural factors, Magical Brush can help novices easily create modern Chinese paintings and experience the cultural connotations in the process.2023XHXu Haoran et al.Computer Science And Technology, Computer Science And TechnologyGenerative AI (Text, Image, Music, Video)Interactive Narrative & Immersive StorytellingCHI
De-Stijl: Facilitating Graphics Design with Interactive 2D Color Palette RecommendationSelecting a proper color palette is critical in crafting a high-quality graphic design to gain visibility and communicate ideas effectively. To facilitate this process, we propose De-Stijl, an intelligent and interactive color authoring tool to assist novice designers in crafting harmonic color palettes, achieving quick design iterations, and fulfilling design constraints. Through De-Stijl, we contribute a novel 2D color palette concept that allows users to intuitively perceive color designs in context with their proportions and proximities. Further, De-Stijl implements a holistic color authoring system that supports 2D palette extraction, theme-aware and spatial-sensitive color recommendation, and automatic graphical elements (re)colorization. We evaluated De-Stijl through an in-lab user study by comparing the system with existing industry standard tools, followed by in-depth user interviews. Quantitative and qualitative results demonstrate that De-Stijl is effective in assisting novice design practitioners to quickly colorize graphic designs and easily deliver several alternatives.2023XSAlark Joshi et al.University of Waterloo360° Video & Panoramic ContentGraphic Design & Typography ToolsPrototyping & User TestingCHI
FingerX: Rendering Haptic Shape of Virtual Objects Augmented from Real Objects using Extendable and Withdrawable Supports on FingersInteracting with not only virtual but also real objects, or even virtual objects augmented by real objects becomes a trend of virtual reality (VR) interactions and is common in augmented reality (AR). However, current haptic shape rendering devices generally focus on feedback of virtual objects, and require the users to put down or take off those devices to perceive real objects. Therefore, we propose FingerX to render haptic shapes and enable users to touch, grasp and interact with virtual and real objects simultaneously. An extender on the fingertip extends to a corresponding height to support between the fingertip and the real objects or the hand, to render virtual shapes. A ring rotates and withdraws the extender behind the fingertip when touching real objects. By independently controlling four extenders and rings on each finger with the exception of the pinky finger, FingerX renders feedback in three common scenarios, including touching virtual objects augmented by real environments (e.g., a desk), grasping virtual objects augmented by real objects (e.g., a bottle) and grasping virtual objects in the hand. We conducted a shape recognition study to evaluate the recognition rates for these three scenarios and obtained an average recognition rate of 76.59% with shape visual feedback. We then performed a VR study to observe how users interact with virtual and real objects simultaneously and verify that FingerX significantly enhances VR realism, compared to current vibrotactile methods.2022HTLynn Tsai et al.National Chengchi UniversityVibrotactile Feedback & Skin StimulationForce Feedback & Pseudo-Haptic WeightHand Gesture RecognitionCHI
Debiased-CAM to mitigate image perturbations with faithful visual explanations of machine learningModel explanations such as saliency maps can improve user trust in AI by highlighting important features for a prediction. However, these become distorted and misleading when explaining predictions of images that are subject to systematic error (bias) by perturbations and corruptions. Furthermore, the distortions persist despite model fine-tuning on images biased by different factors (blur, color temperature, day/night). We present Debiased-CAM to recover explanation faithfulness across various bias types and levels by training a multi-input, multi-task model with auxiliary tasks for explanation and bias level predictions. In simulation studies, the approach not only enhanced prediction accuracy, but also generated highly faithful explanations about these predictions as if the images were unbiased. In user studies, debiased explanations im- proved user task performance, perceived truthfulness and perceived helpfulness. Debiased training can provide a versatile platform for robust performance and explanation faithfulness for a wide range of applications with data biases.2022WZHaimo Zhang et al.School of Computing, National University of SingaporeExplainable AI (XAI)Algorithmic Transparency & AuditabilityCHI
Does Mode of Digital Contact Tracing Affect User Willingness to Share Information? A Quantitative StudyDigital contact tracing can limit the spread of infectious diseases. Nevertheless, barriers remain to attain sufficient adoption. In this study, we investigate how willingness to participate in contact tracing is affected by two critical factors: the modes of data collection and the type of data collected. We conducted a scenario-based survey study among 220 respondents in the United States (U.S.) to understand their perceptions about contact tracing associated with automated and manual contact tracing methods. The findings indicate a promising use of smartphones and a combination of public health officials and medical health records as information sources. Through a quantitative analysis, we describe how different modalities and individual demographic factors may affect user compliance when participants are asked to provide four key information pieces for contact tracing.2022CZCamellia Zakaria et al.University of Massachusetts AmherstPrivacy by Design & User ControlPrivacy Perception & Decision-MakingCHI
Nudge for Reflection: More than Just a Channel to Political KnowledgeReflection, a process that organizes information into a structure that incorporates both own and others’ perspectives, was previously believed to function mainly as an antecedent of political knowledge. In this paper, we first design a simple interface nudge to encourage users to reflect on their views on political issues. Second, we use an experimental study to show that reflection works in a way more than leading to political knowledge. Results from a between-subjects online experiment (N = 168) covering one crucial public issue in Singapore (i.e., fertility) showed that (a) reflection interacts with information access to influence perceived issue knowledge; (b) reflection enhances perceived attitude certainty, including perceived attitude clarity and perceived attitude correctness; (c) reflection promotes willingness to express opinions in private settings.2021WZWeiyu Zhang et al.National University of SingaporeActivism & Political ParticipationInclusive DesignCHI
"You Cannot Offer Such a Suggestion": Designing for Family Caregiver Input in Home Care SystemsPrevious work has looked closely at the challenges of using patient-generated data to enable remote assessment and monitoring by healthcare professionals. In this paper, we examine family caregivers who act as proxies for patients who may not have the capacity of capturing the necessary data. We worked with occupational therapists to develop an application for remote assessment of the safety of patients' homes by occupational therapists with the assistance of family caregivers. We evaluated the application with family caregivers and found two features unique to communication between family caregivers and healthcare professionals: Caregivers want to be able to direct healthcare professionals' attention to support problem-solving at home, and they include their perspective on how to best meet the patient's health needs. We discuss the importance of these findings for home systems in the domain of long-term chronic care.2020PFPin Sym Foong et al.National University of SingaporeElderly Care & Dementia SupportAging-in-Place Assistance SystemsCHI
Nudge for Deliberativeness: How Interface Features Influence Online DiscourseCognitive load is a significant challenge to users for being deliberative. Interface design has been used to mitigate this cognitive state. This paper surveys literature on the anchoring effect, partitioning effect and point-of-choice effect, based on which we propose three interface nudges, namely, the word-count anchor, partitioning text fields, and reply choice prompt. We then conducted a 2×2×2 factorial experiment with 80 participants (10 for each condition), testing how these nudges affect deliberativeness. The results showed a significant positive impact of the word-count anchor. There was also a significant positive impact of the partitioning text fields on the word count of response. The reply choice prompt showed a surprisingly negative affect on the quantity of response, hinting at the possibility that the reply choice prompt induces a fear of evaluation, which could in turn dampen the willingness to reply.2020SMSanju Menon et al.National University of SingaporePrivacy by Design & User ControlSocial Platform Design & User BehaviorUser Research Methods (Interviews, Surveys, Observation)CHI
Understanding Digitally-Mediated Empathy: An Exploration of Visual, Narrative, and Biosensory Informational CuesDigitally sharing our experiences engages a process of empathy shaped by available informational cues. Biosensory data is one informative cue, but the relationship to empathy is underexplored. In this study, we investigate this process by showing a video of a "target'' person's visual perspective watching a virtual reality film to sixty "observers''. We vary information available to observers via three experimental conditions: a baseline unmodified video, video with narrative text, or with a graph of electrodermal activity (EDA) of the target. Compared to baseline, narrative text increased empathic accuracy (EA) while EDA had an opposite, negative effect. Qualitatively, observers describe their empathic processes as using their own feelings supplemented with the information presented depending on the interpretability of that information. Both narration and EDA prompted observers to reconsider assumptions about another's experience. Our findings lead to a discussion of digitally-mediated empathy with implications for associated research and product development.2019MCMax T Curran et al.University of California, BerkeleyVisualization Perception & CognitionBiosensors & Physiological MonitoringCHI
Dissonance Between Human and Machine UnderstandingComplex machine learning models are deployed in several critical domains including healthcare and autonomous vehicles nowadays, albeit as functional blackboxes. Consequently, there has been a recent surge in interpreting decisions of such complex models in order to explain their actions to humans. Models which correspond to human interpretation of a task are more desirable in certain contexts and can help attribute liability, build trust, expose biases and in turn build better models. It is therefore crucial to understand how and which models conform to human understanding of tasks. In this paper we present a large-scale crowdsourcing study that reveals and quantifies the dissonance between human and machine understanding, through the lens of an image classification task. In particular, we seek to answer the following questions: Which (well performing) complex ML models are closer to humans in their use of features to make accurate predictions? How does task difficulty affect the feature selection capability of machines in comparison to humans? Are humans consistently better at selecting features that make image recognition more accurate? Our findings have important implications on human-machine collaboration, considering that a long term goal in the field of artificial intelligence is to make machines capable of learning and reasoning like humans.2019ZZHaimo Zhang et al.Human–machine configurationsCSCW
GestAKey: Touch Interaction on Individual KeycapsConventionally, keys on a physical keyboard have only two states: ``released’’ and ``pressed’’. As such, various techniques, such as hotkeys, are designed to enhance the keyboard expressiveness. Realizing that user inevitably perform touch actions during keystrokes, we propose GestAKey, leveraging location and motion of the touch on individual keycaps to augment the functionalities of existing keystrokes. With a log study, we collected touch data for both normal usage (typing and hotkeys) and while performing touch gestures (location and motion), which are analyzed to assess the viability of augmenting keystrokes with simultaneous gestures. A controlled experiment was conducted to compare GestAKey with existing keyboard interaction techniques, in terms of efficiency and learnability. The results show that GestAKey has comparable performance with hotkey. We further discuss the insights of integrating such touch modality into existing keyboard interaction, and demonstrate several usage scenarios.2018YSYilei Shi et al.Singapore University Of Technology and DesignHand Gesture RecognitionHuman-LLM CollaborationCHI