arXiv subjects

Changkun Ou

Publications and source records attributed to Changkun Ou.

10 recordsLinked to original sources

arXiv2026

Progressive Autonomy as Preference Learning: A Formalization of Trust Calibration for Agentic Tool Use

We formalize trust calibration for agentic tool use (deciding when an automated agent's proposed action may execute autonomously versus require human approval) as a preference-learning problem. A policy gateway maintains a Gaussian-process posterior over a latent human risk-tolerance function, observed through a probit likelihood on binary approve/deny feedback, and escalates to the human exactly where the approval outcome is most uncertain. We show this is structurally an instance of Preferential Bayesian Optimization, inheriting its inference machinery (approximate Gaussian-process classification) and its sample-efficiency argument (uncertainty-targeted querying), while differing in objective: classifying an action space into allow/block/ask regions rather than optimizing a design.

cs.AI↗

arXiv2025

At the Speed of the Heart: Evaluating Physiologically-Adaptive Visualizations for Supporting Engagement in Biking Exergaming in Virtual Reality

Many exergames face challenges in keeping users within safe and effective intensity levels during exercise. Meanwhile, although wearable devices continuously collect physiological data, this information is seldom leveraged for real-time adaptation or to encourage user reflection. We designed and evaluated a VR cycling simulator that dynamically adapts based on users' heart rate zones. First, we conducted a user study (N=50) comparing eight visualization designs to enhance engagement and exertion control, finding that gamified elements like non-player characters (NPCs) were promising for feedback delivery. Based on these findings, we implemented a physiology-adaptive exergame that adjusts visual feedback to keep users within their target heart rate zones. A lab study (N=18) showed that our system has potential to help users maintain their target heart rate zones. Subjective ratings of exertion, enjoyment, and motivation remained largely unchanged between conditions. Our findings suggest that real-time physiological adaptation through NPC visualizations can improve workout regulation in exergaming.

cs.HC↗

arXiv2023

Designing and Evaluating an Adaptive Virtual Reality System using EEG Frequencies to Balance Internal and External Attention States

Virtual reality finds various applications in productivity, entertainment, and training scenarios requiring working memory and attentional resources. Working memory relies on prioritizing relevant information and suppressing irrelevant information through internal attention, which is fundamental for successful task performance and training. Today, virtual reality systems do not account for the impact of working memory loads resulting in over or under-stimulation. In this work, we designed an adaptive system based on EEG correlates of external and internal attention to support working memory task performance. Here, participants engaged in a visual working memory N-Back task, and we adapted the visual complexity of distracting surrounding elements. Our study first demonstrated the feasibility of EEG frontal theta and parietal alpha frequency bands for dynamic visual complexity adjustments. Second, our adaptive system showed improved task performance and diminished perceived workload compared to a reverse adaptation. Our results show the effectiveness of the proposed adaptive system, allowing for the optimization of distracting elements in high-demanding conditions. Adaptive systems based on alpha and theta frequency bands allow for the regulation of attentional and executive resources to keep users engaged in a task without resulting in cognitive overload.

cs.HC↗

arXiv2023

The Impact of Expertise in the Loop for Exploring Machine Rationality

Human-in-the-loop optimization utilizes human expertise to guide machine optimizers iteratively and search for an optimal solution in a solution space. While prior empirical studies mainly investigated novices, we analyzed the impact of the levels of expertise on the outcome quality and corresponding subjective satisfaction. We conducted a study (N=60) in text, photo, and 3D mesh optimization contexts. We found that novices can achieve an expert level of quality performance, but participants with higher expertise led to more optimization iteration with more explicit preference while keeping satisfaction low. In contrast, novices were more easily satisfied and terminated faster. Therefore, we identified that experts seek more diverse outcomes while the machine reaches optimal results, and the observed behavior can be used as a performance indicator for human-in-the-loop system designers to improve underlying models. We inform future research to be cautious about the impact of user expertise when designing human-in-the-loop systems.

cs.HC↗

arXiv2023

Short-Form Videos Degrade Our Capacity to Retain Intentions: Effect of Context Switching On Prospective Memory

Social media platforms use short, highly engaging videos to catch users' attention. While the short-form video feeds popularized by TikTok are rapidly spreading to other platforms, we do not yet understand their impact on cognitive functions. We conducted a between-subjects experiment (N=60) investigating the impact of engaging with TikTok, Twitter, and YouTube while performing a Prospective Memory task (i.e., executing a previously planned action). The study required participants to remember intentions over interruptions. We found that the TikTok condition significantly degraded the users' performance in this task. As none of the other conditions (Twitter, YouTube, no activity) had a similar effect, our results indicate that the combination of short videos and rapid context-switching impairs intention recall and execution. We contribute a quantified understanding of the effect of social media feed format on Prospective Memory and outline consequences for media technology designers to not harm the users' memory and wellbeing.

cs.HC↗

arXiv2022

The Human in the Infinite Loop: A Case Study on Revealing and Explaining Human-AI Interaction Loop Failures

Interactive AI systems increasingly employ a human-in-the-loop strategy. This creates new challenges for the HCI community when designing such systems. We reveal and investigate some of these challenges in a case study with an industry partner, and developed a prototype human-in-the-loop system for preference-guided 3D model processing. Two 3D artists used it in their daily work for 3 months. We found that the human-AI loop often did not converge towards a satisfactory result and designed a lab study (N=20) to investigate this further. We analyze interaction data and user feedback through the lens of theories of human judgment to explain the observed human-in-the-loop failures with two key insights: 1) optimization using preferential choices lacks mechanisms to deal with inconsistent and contradictory human judgments; 2) machine outcomes, in turn, influence future user inputs via heuristic biases and loss aversion. To mitigate these problems, we propose descriptive UI design guidelines. Our case study draws attention to challenging and practically relevant imperfections in human-AI loops that need to be considered when designing human-in-the-loop systems.

cs.HC↗

arXiv2022

Generative 3D Animation Pipelines: Automating Facial Retargeting Workflows

Design tools in the 3D industry, while powerful, are still tedious and labor-intensive when it comes to bringing a creative idea for a visual effect to life. In this position paper, we discussed how an infamous generative synthetic media, deepfakes, could be of use and embedded into common sophisticated 3D workflows to reduce user workloads in areas such as 3D model editing, material design, and character animation. As a case discussion, we also prototyped a tool to address the retargeting problem in character animation. Although deepfakes themselves have received a negative public image, the results of our interviews with field experts are unexpectedly positive in regard to our tool that utilizes deepfake algorithms. Lastly, we also discussed our experience and observed design practices to put deepfakes to good use, including how we could avoid potential misuses directly by design, how this design changes user interactions, and subsequent open issues.

cs.HC↗

arXiv2021

Modeling Web Browsing Behavior across Tabs and Websites with Tracking and Prediction on the Client Side

Clickstreams on individual websites have been studied for decades to gain insights into user interests and to improve website experiences. This paper proposes and examines a novel sequence modeling approach for web clickstreams, that also considers multi-tab branching and backtracking actions across websites to capture the full action sequence of a user while browsing. All of this is done using machine learning on the client side to obtain a more comprehensive view and at the same time preserve privacy. We evaluate our formalism with a model trained on data collected in a user study with three different browsing tasks based on different human information seeking strategies from psychological literature. Our results show that the model can successfully distinguish between browsing behaviors and correctly predict future actions. A subsequent qualitative analysis identified five common web browsing patterns from our collected behavior data, which help to interpret the model. More generally, this illustrates the power of overparameterization in ML and offers a new way of modeling, reasoning with, and prediction of observable sequential human interaction behaviors.

cs.HC↗

arXiv2019

Identifying Malicious Players in GWAP-based Disaster Monitoring Crowdsourcing System

Disaster monitoring is challenging due to the lake of infrastructures in monitoring areas. Based on the theory of Game-With-A-Purpose (GWAP), this paper contributes to a novel large-scale crowdsourcing disaster monitoring system. The system analyzes tagged satellite pictures from anonymous players, and then reports aggregated and evaluated monitoring results to its stakeholders. An algorithm based on directed graph centralities is presented to address the core issues of malicious user detection and disaster level calculation. Our method can be easily applied in other human computation systems. In the end, some issues with possible solutions are discussed for our future work.

cs.SI↗

arXiv2019

WatchOut: A Road Safety Extension for Pedestrians on a Public Windshield Display

We conducted a field study to investigate whether public windshield displays are applicable as an additional interactive digital road safety warning sign. We focused on investigating the acceptance and usability of our novel public windshield display and its potential use for future applications. The study has shown that users are open-minded to the idea of an extraverted windshield display regardless the use case, whether it is used for safety purposes or different content. Contrary to our hypothesis most people assumed they would mistrust the system if it were as well established as traffic lights and primarily rely on their own perception.

cs.HC↗