LLM-based In-situ Thought Exchanges for Critical Paper ReadingCritical reading is a primary way through which researchers develop their critical thinking skills. While exchanging thoughts and opinions with peers can strengthen critical reading, junior researchers often lack access to peers who can offer diverse perspectives. To address this gap, we designed an in-situ thought exchange interface informed by peer feedback from a formative study (N=8) to support junior researchers’ critical paper reading. We evaluated the effects of thought exchanges under three conditions (no-agent, single-agent, and multi-agent) with 46 junior researchers over two weeks. Our results showed that incorporating agent-mediated thought exchanges during paper reading significantly improved participants’ critical thinking scores compared to the no-agent condition. In the single-agent condition, participants more frequently made reflective annotations on the paper content. In the multi-agent condition, participants engaged more actively with agents’ responses. Our qualitative analysis further revealed that participants compared and analyzed multiple perspectives in the multi-agent condition. This work contributes to understanding in-situ AI-based support for critical paper reading through thought exchanges and offers design implications for future research.2026XFXinrui Fang et al.The University of TokyoHuman-LLM CollaborationUser Research Methods (Interviews, Surveys, Observation)Prototyping & User TestingIUI
Does It Matter Which Finger You Use? Investigating Finger Identity and Haptic Pattern Recognition for Stationary and Moving FingersHaptic perception on touchscreens varies across fingers, yet little is known about how finger identity and multi-finger use shape tactile discrimination and user experience. We conducted two experiments with four haptic feedback. In Experiment 1, right-handed participants explored each of the ten fingers individually under stationary and moving conditions. Experiment 2 examined two-finger sequences with same participants. Results showed that moving exploration enhanced accuracy, confidence, and enjoyment, while stationary touch increased cognitive and physical load, especially for weaker fingers such as the left ring and pinky. The right thumb and index consistently performed best. In dual-finger trials, moving exploration improved second-finger performance, and adjacent same-hand pairs (e.g., Left Index–Left Thumb, Right Thumb–Right Index) yielded higher synergy. These findings highlight the role of finger anatomy, motion, and coordination, and provide concrete guidelines on which fingers (or combinations) and exploration modes to assign for haptic surfaces that optimize accuracy, comfort, and engagement.2026MJMilad Jamalzadeh et al.University Polytechnic Hauts-De-FranceVibrotactile Feedback & Skin StimulationIn-Vehicle Haptic, Audio & Multimodal FeedbackHaptic WearablesCHI
Every Move You Make: Visualizing Near-Future Motion Under Delay for TeleroboticsDelays in direct teleoperation decouple operator input from robot feedback. We frame this not as a unitary problem but as three facets of operator uncertainty: (1) communication, when commands take effect, (2) trajectory, how inputs map to motion, and (3) environmental, how external factors alter outcomes. We externalized each facet through predictive visualizations: Network, Path, and Envelope. In a controlled study with 24 participants (novices in telerobotics) navigating a simulated robot under a fixed 2.56 s round-trip delay, we compared these visualizations against a delayed-video baseline. Path significantly shortened task time, lowered perceived cognitive load, and reduced reliance on reactive "move-and-wait" behavior. Envelope lowered cognitive load but did not significantly reduce reactive behavior or improve performance, while Network had no measurable effect. These results indicate that predictive support is effective only when trajectory uncertainty is externalized, enabling operators to move from reactive to more proactive control.2026DCDries Cardinaels et al.UHasselt - Flanders MakeTeleoperated DrivingTeleoperation & TelepresenceCHI
TFTune: Creation and Personalization of Pointing Transfer Functions Using Reinforcement LearningPointing transfer functions define the mapping between input devices and onscreen cursor movement. Despite being used by millions daily, only marginal improvements in pointing performance have been achieved by tuning transfer functions since the introduction of acceleration-based gains. We present TFTune, a reinforcement learning-based approach for improving pointing by automatically tuning personalized transfer functions. We show that TFTune-generated functions outperform operating system defaults, improving movement times by 7% on macOS when using a trackpad (7 minutes of tuning) and 8% on participants' personal Windows computers with hardware (i.e., mice and monitors) of varying characteristics (after just 1 minute of tuning). Further, we show that TFTune generalizes beyond traditional pointing devices, providing 16% improvement for a muscle-computer interface (2 minutes of tuning). TFTune demonstrates an initial approach for scalable and meaningful performance improvements in input–output mappings, opening a new direction for exploring the use of machine learning for improving fundamental computer inputs.2026EEEthan Eddy et al.University of New BrunswickForce Feedback & Pseudo-Haptic WeightGenerative AI (Text, Image, Music, Video)AI-Assisted Decision-Making & AutomationCHI
DesignTrace: Exploring, Iterating and Tracking Design Alternatives with GenAICreativity support tools have begun to incorporate GenAI for exploring ideas. However, our preliminary study with nine designers showed that current GenAI tools lack explicit support for iteratively evolving, reflecting upon and tracking design alternatives. We developed DesignTrace, an early-stage GenAI design tool that allows designers to experiment with semantically relevant visual variations in an interactive design space. Its representation captures the progression of designers’ visual and semantic ideas through command histories, state tracking, and an interactive branching structure. A study of twelve professional designers shows that DesignTrace’s palette helps express, explore, and reflect on design intentions. Its interactive branching structure helps them maintain visual consistency across design iterations; remember and revisit earlier design decisions; and see connections across ideas. Our work shows how re-envisioning GenAI-based interfaces around explicit design traces enable designers to benefit from generative capabilities while maintaining control as they explore design variants.2026XPXiaohan Peng et al.Université Paris-Saclay, CNRS, InriaGenerative AI (Text, Image, Music, Video)Creative Collaboration & Feedback SystemsGraphic Design & Typography ToolsCHI
How are Vibrotactile Experiences Visually Represented? A Taxonomy of Illustration CharacteristicsVibrotactile experiences (VTX) consist of a multitude of design parameters and experiential dimensions that can be challenging to communicate visually. To understand how this is commonly done in scientific communication, we systematically reviewed VTX illustrations in academic publications. Using inductive and deductive methods, we built a taxonomy detailing characteristics of VTX illustrations that focuses on what is illustrated and how it is depicted. Using the taxonomy, we coded a total of 768 figures spanning 409 publications. These results indicate that (1) half of the illustrations communicate on the timing of vibrotactile feedback with regards to users’ actions, (2) illustrations depict stimuli rather than experiences and infrequently communicate multimodal aspects of the experiences, and (3) contextual information of vibrotactile displays and experiential aspects are often distributed across several complementary figures. We conclude by discussing the benefits and limitations of this taxonomy to support the design process.2026BFBruno Fruchard et al.Univ. Lille, Inria, CNRS, Centrale LilleVibrotactile Feedback & Skin StimulationCHI
Interface Dis/Similarities: Investigating Characteristics Influencing Perceived Differences Between GUIsWhile HCI research acknowledges that prior knowledge can be transferred from familiar interfaces to unfamiliar ones, we lack an understanding of which interface characteristics support this process. To adress this issue, we conducted three experiments to identify the interface characteristics that influence the perception of \dis between software interfaces. The first, which involves a card-sorting activity, identifies seven intrinsic characteristics of interface. The second, conducted via an online pairwise comparison survey, identifies three characteristics inherent to the interface's display context. Finally, the third experiment contrasts the ten identified characteristics to determine their respective influence on perceived interface \dis. Altogether, our results provide actionable guidance for understanding how users perceive differences between interfaces and how such perceptions may inform or facilitate analogical transfer of knowledge from familiar to unfamiliar interfaces.2026RPRaphaël Perraud et al.Univ. Lille, Inria, CNRS, Centrale Lille, UMR 9189 CRIStALUser Research Methods (Interviews, Surveys, Observation)Prototyping & User TestingInteractive Data VisualizationCHI
Investigating Single-Handed Microgesture Scrolling TechniquesScrolling is ubiquitous in our daily computing experience. We explore how single-handed microgestures can be used for scrolling. Based on an analysis of the basic components necessary for scrolling, we selected 3 microgestures: Tap, Hold and Drag. Considering both rate and position controls, we designed 4 microgesture-based scrolling techniques adapted to these 3 microgestures. We contrasted these 4 techniques in a laboratory experiment with 24 participants who performed 2 tasks: a reciprocal selection task, where participants scrolled the view to reach and select a target; and a counting task, where participants scrolled the view to count image occurrences. Our results suggest that the technique based on Drag microgestures with rate control is the most effective for scrolling operations, regardless of the task. This work demonstrates that microgestures, with their advantages for frequent everyday tasks, offer a promising approach to continuous and efficient scrolling control.2026SLSuliac Lavenant et al.Univ. Lille, Inria, CNRS, Centrale Lille, UMR 9189 CRIStALHand Gesture RecognitionOne-Handed Operation & Mobile GesturesCHI
Strategies for the Reconciliation of Artistic Intent and Technical Constraints in Mixed Reality PerformancesAs immersive technologies advance, Mixed Reality Performances (MRP) increasingly integrate them but often face technical challenges that must balance artistic vision and practical constraints. These constraints sometimes lead to unavoidable limitations. The diversity of technologies and artistic goals in MRP prevents a one-size-fits-all solution to such dilemmas. This paper presents a model unifying strategies for addressing compromises between artistic intent and technological feasibility. Using reflexive thematic analysis of a performance-led case study, interviews, and independent case studies, we identify recurring strategies in Mixed Reality experiences with varying constraints. These strategies fall on an axis based on the audience's awareness of limitations and are categorized into five approaches: Avoid, Disguise, Tolerate, Integrate, or Leverage. We argue that this framework helps designers better navigate the limitations inherent in creating MRPs, offering practical pathways to align technological capabilities with creative objectives.2025PUPierrick Uro et al.Digital Art Installations & Interactive PerformanceInteractive Narrative & Immersive StorytellingDIS
Facilitating the Parametric Definition of Geometric Properties in Programming-Based CADParametric Computer-aided design (CAD) enables the creation of reusable models by integrating variables into geometric properties, facilitating customization without a complete redesign. However, creating parametric designs in programming-based CAD presents significant challenges. Users define models in a code editor using a programming language, with the application generating a visual representation in a viewport. This process involves complex programming and arithmetic expressions to describe geometric properties, linking various object properties to create parametric designs. Unfortunately, these applications lack assistance, making the process unnecessarily demanding. We propose a solution that allows users to retrieve parametric expressions from the visual representation for reuse in the code, streamlining the design process. We demonstrated this concept through a proof-of-concept implemented in the programming-based CAD application, OpenSCAD, and conducted an experiment with 11 users. Our findings suggest that this solution could significantly reduce design errors, improve interactivity and engagement in the design process, and lower the entry barrier for newcomers by reducing the mathematical skills typically required in programming-based CAD applications2024JAJ Felipe Gonzalez Avila et al.Desktop 3D Printing & Personal FabricationCircuit Making & Hardware PrototypingUIST
Tutorial mismatches: understanding the frictions due to interface differences when following software video tutorialsVideo tutorials are the main medium to learn novel software skills. However, the User Interface (UI) presented in a video tutorial may differ from the learner's UI because of customizations or differences in software versions. We investigate the frictions resulting from such differences on a learners' ability to reproduce a task demonstrated in a video tutorial. Through a morphological analysis, we first identify 13 types of "interface differences" that differ in terms of availability, reachability and spatial location of features in the interface. To better assess the frictions resulting from each of these differences, we then conduct a laboratory study with 26 participants instructed to reproduce a vector graphics editing task. Our results highlight interesting UI comparison behaviors, and illustrate various approaches employed to visually locate features.2024RPRaphaël Perraud et al.Programming Education & Computational ThinkingUser Research Methods (Interviews, Surveys, Observation)Prototyping & User TestingDIS
Understanding the Challenges of OpenSCAD Users for 3D PrintingDirect manipulation has been established as the main interaction paradigm for Computer-Aided Design (CAD) for decades. It provides fast, incremental, and reversible actions that allow for an iterative process on a visual representation of the result. Despite its numerous advantages, some users prefer a programming-based approach where they describe the 3D model they design with a specific programming language, such as OpenSCAD. It allows users to create complex structured geometries and facilitates abstraction. Unfortunately, most current knowledge about CAD practices only focuses on direct manipulation programs. In this study, we interviewed 20 programming-based CAD users to understand their motivations and challenges. Our findings reveal that this programming-oriented population presents difficulties in the design process in tasks such as 3D spatial understanding, validation and code debugging, creation of organic shapes, and code-view navigation.2024JAJ Felipe Gonzalez Avila et al.Université de Lille, Carleton UniversityDesktop 3D Printing & Personal FabricationCircuit Making & Hardware PrototypingCHI
DirectGPT: A Direct Manipulation Interface to Interact with Large Language ModelsWe characterize and demonstrate how the principles of direct manipulation can improve interaction with large language models. This includes: continuous representation of generated objects of interest; reuse of prompt syntax in a toolbar of commands; manipulable outputs to compose or control the effect of prompts; and undo mechanisms. This idea is exemplified in DirectGPT, a user interface layer on top of ChatGPT that works by transforming direct manipulation actions to engineered prompts. A study shows participants were 50% faster and relied on 50% fewer and 72% shorter prompts to edit text, code, and vector images compared to baseline ChatGPT. Our work contributes a validated approach to integrate LLMs into traditional software using direct manipulation. Data, code, and demo available at https://osf.io/3wt6s.2024DMDamien Masson et al.University of WaterlooHuman-LLM CollaborationExplainable AI (XAI)AI-Assisted Decision-Making & AutomationCHI
Statslator: Interactive Translation of NHST and Estimation Statistics Reporting Styles in Scientific DocumentsInferential statistics are typically reported using p-values (NHST) or confidence intervals on effect sizes (estimation). This is done using a range of styles, but some readers have preferences about how statistics should be presented and others have limited familiarity with alternatives. We propose a system to interactively translate statistical reporting styles in existing documents, allowing readers to switch between interval estimates, p-values, and standardized effect sizes, all using textual and graphical reports that are dynamic and user customizable. Forty years of CHI papers are examined. Using only the information reported in scientific documents, equations are derived and validated on simulated datasets to show that conversions between p-values and confidence intervals are accurate. The system helps readers interpret statistics in a familiar style, compare reports that use different styles, and even validate the correctness of reports. Code and data: https://osf.io/x4ue72023DMDamien Masson et al.Interactive Data VisualizationTime-Series & Network Graph VisualizationUIST
Typing Efficiency and Suggestion Accuracy Influence the Benefits and Adoption of Word SuggestionsSuggesting words to complete a given sequence of characters is a common feature of typing interfaces. Yet, previous studies have not found a clear benefit, some even finding it detrimental. We report on the first study to control for two important factors, word suggestion accuracy and typing efficiency. Our accuracy factor is enabled by a new methodology that builds on standard metrics of word suggestions. Typing efficiency is based on device type. Results show word suggestions are used less often in a desktop condition, with little difference between tablet and phone conditions. Very accurate suggestions do not improve entry speed on desktop, but do on tablet and phone. Based on our findings, we discuss implications for the design of automation features in typing systems.2021QRQuentin Roy et al.University of WaterlooHuman-LLM CollaborationRecommender System UXCHI
Modeling and Reducing Spatial Jitter caused by Asynchronous Input and Output RatesJitter in interactive systems occurs when visual feedback is perceived as unstable or trembling even though the input signal is smooth or stationary. It can have multiple causes such as sensing noise, or feedback calculations introducing or exacerbating sensing imprecisions. Jitter can however occur even when each individual component of the pipeline works perfectly, as a result of the differences between the input frequency and the display refresh rate. This asynchronicity can introduce rapidly-shifting latencies between the rendered feedbacks and their display on screen, which can result in trembling cursors or viewports. This paper contributes a better understanding of this particular type of jitter. We first detail the problem from a mathematical standpoint, from which we develop a predictive model of jitter amplitude as a function of input and output frequencies, and a new metric to measure this spatial jitter. Using touch input data gathered in a study, we developed a simulator to validate this model and to assess the effects of different techniques and settings with any output frequency. The most promising approach, when the time of the next display refresh is known, is to estimate (interpolate or extrapolate) the user’s position at a fixed time interval before that refresh. When input events occur at 125 Hz, as is common in touch screens, we show that an interval of 4 to 6 ms works well for a wide range of display refresh rates. This method effectively cancels most of the jitter introduced by input/output asynchronicity, while introducing minimal imprecision or latency.2020AAAxel Antoine et al.Eye Tracking & Gaze InteractionVisualization Perception & CognitionNotification & Interruption ManagementUIST
Have a SEAT on Stage : Restoring Trust with Spectator Experience Augmentation TechniquesWhen the collaboration between humans and machines happens in public, the audience can face difficulties in distinguishing the actual human contribution from the contribution of autonomous processes. In music concerts involving digital interfaces doubts about the performer's contribution can drastically hinder the audience interest. The disappearing of the direct physical link between actions and effects is one of the reasons of this confusion. Consequently both artists and researchers have explored techniques to augment the experience of spectators. However their respective impact on the multiple aspects of audience experience has not yet been formally compared. In this controlled study, we compare two techniques : pre-performance explanations and visual augmentations. Despite contradictory results on comprehension tasks, we show that contrary to pre-performance explanations, visual augmentations improve the audience experience, increase their subjective comprehension and restore the trust in performers by reversing the doubt in their favour.2020OCOlivier Capra et al.Digital Art Installations & Interactive PerformanceInteractive Narrative & Immersive StorytellingDIS
Leveraging Distal Vibrotactile Feedback for Target AcquisitionMany touch based interactions provide limited opportunities for direct tactile feedback; examples include multi-user touch displays, augmented reality based projections on passive surfaces, and mid-air input. In this paper, we consider distal feedback, through vibrotactile stimulation on a smart-watch placed on the user's non-dominant wrist, as an alternative feedback mechanism to interaction location vibrotactile feedback, under the user's finger. We compare the effectiveness of interaction location feedback vs. distal feedback through a Fitts's Law task completed on a smartphone. Results show that distal and interaction location feedback both reduce errors in target acquisition and exhibit statistically comparable performance, suggesting that distal vibrotactile feedback is a suitable alternative when interaction location feedback is not readily available.2019JHJay Henderson et al.University of WaterlooVibrotactile Feedback & Skin StimulationFoot & Wrist InteractionCHI
RayCursor: A 3D Pointing Facilitation Technique based on RaycastingRaycasting is the most common target pointing technique in virtual reality environments. However, performance on small and distant targets is impacted by the accuracy of the pointing device and the user's motor skills. Current pointing facilitation techniques are currently only applied in the context of the virtual hand, i.e. for targets within reach. We propose enhancements to Raycasting: filtering the ray, and adding a controllable cursor on the ray to select the nearest target. We describe a series of studies for the design of the visual feedforward, filtering technique, as well as a comparative study between different 3D pointing techniques. Our results show that highlighting the nearest target is one of the most efficient visual feedforward technique. We also show that filtering the ray reduces error rate in a drastic way. Finally we show the benefits of RayCursor compared to Raycasting and another technique from the literature.2019MBMarc Baloup et al.Université de Lille & InriaFull-Body Interaction & Embodied InputImmersion & Presence ResearchCHI
Using High Frequency Accelerometer and Mouse to Compensate for End-to-end Latency in Indirect InteractionEnd-to-end latency corresponds to the temporal difference between a user input and the corresponding output from a system. It has been shown to degrade user performance in both direct and indirect interaction. If it can be reduced to some extend, latency can also be compensated through software compensation by trying to predict the future position of the cursor based on previous positions, velocities and accelerations. In this paper, we propose a hybrid hardware and software prediction technique specifically designed for partially compensating end-to-end latency in indirect pointing. We combine a computer mouse with a high frequency accelerometer to predict the future location of the pointer using Euler based equations. Our prediction method results in more accurate prediction than previously introduced prediction algorithms for direct touch. A controlled experiment also revealed that it can improve target acquisition time in pointing tasks.2018AAAxel Antoine et al.Université de LillePrototyping & User TestingCHI