Designing structured visuals such as presentation slides is essential for communicative needs, necessitating both content creation and visual planning skills. In this work, we tackle the challenge of automated slide generation, where models produce slide presentations from natural language (NL) instructions. We first introduce the SlidesBench benchmark, the first benchmark for slide generation with 7k training and 585 testing examples derived from 310 slide decks across 10 domains. SlidesBench supports evaluations that are (i) reference-based to measure similarity to a target slide, and (ii) reference-free to measure the design quality of generated slides alone. We benchmark end-to-end image generation and program generation methods with a variety of models, and find that programmatic methods produce higher-quality slides in user-interactable formats. Built on the success of program generation, we create AutoPresent, an 8B LlaMa-based model trained on 7k pairs of instructions paired with code for slide generation, and achieve results comparable to the closed-source model GPT-4O. We further explore iterative design refinement where the model is tasked to self-refine its own output, and we found that this process improves the slide’s quality. We hope that our work will provide a basis for future work on generating structured visuals. Our code, data, demo, and video demonstrations are publicly available at https: //github.com/para-lost/AutoPresent

Recent advances in general-purpose AI underscore the urgent need to align AI systems with human goals and values. Yet, the lack of a clear, shared understanding of what constitutes "alignment" limits meaningful progress and cross-disciplinary collaboration. In this position paper, we argue that the research community should explicitly define and critically reflect on "alignment" to account for the bidirectional and dynamic relationship between humans and AI. Through a systematic review of over 400 papers spanning HCI, NLP, ML, and more, we examine how alignment is currently defined and operationalized. Building on this analysis, we introduce the Bidirectional Human-AI Alignment framework, which not only incorporates traditional efforts to align AI with human values but also introduces the critical, underexplored dimension of aligning humans with AI – supporting cognitive, behavioral, and societal adaptation to rapidly advancing AI technologies. Our findings reveal significant gaps in current literature, especially in long-term interaction design, human value modeling, and mutual understanding. We conclude with three central challenges and actionable recommendations to guide future research toward more nuanced, reciprocal, and human-AI alignment approaches.

In this paper, we explore the design and use of conversational telepresence robots to help homebound older adults interact with the external world. An initial needfinding study (N=8) using video vignettes revealed older adults’ experiential needs for robot-mediated remote experiences such as exploration, reminiscence and social participation. We then designed a prototype system to support these goals and conducted a technology probe study (N=11) to garner a deeper understanding of user preferences for remote experiences. The study revealed user interactive patterns in each desired experience, highlighting the need of robot guidance, social engagements with the robot and the remote bystanders. Our work identifies a novel design space where conversational telepresence robots can be used to foster meaningful interactions in the remote physical environment. We offer design insights into the robot’s proactive role in providing guidance and using dialogue to create personalized, contextualized and meaningful experiences.

A long-standing challenge in accessible computing has been to get developers to produce the accessible UI code necessary for assistive technologies to work properly. AI coding assistants (e.g., Github Copilot) potentially offer a new opportunity to make UI code more accessible automatically, but it is unclear how their use impacts code accessibility and what developers need to know in order to use them effectively. In this paper, we report on a study where developers untrained in accessibility were tasked with building web UI components with and without an AI coding assistant. Our findings suggest that while current AI coding assistants show potential for creating more accessible UIs, they currently require accessibility awareness and expertise, limiting their expected impact.

User interface (UI) design is a difficult yet important task for ensuring the usability, accessibility, and aesthetic qualities of applications. In our paper, we develop a machine-learned model, UIClip, for assessing the design quality and visual relevance of a UI given its screenshot and natural language description. To train UIClip, we used a combination of automated crawling, synthetic augmentation, and human ratings to construct a large-scale dataset of UIs, collated by description and ranked by design quality. Through training on the dataset, UIClip implicitly learns properties of good and bad designs by i) assigning a numerical score that represents a UI design’s relevance and quality and ii) providing design suggestions. In an evaluation that compared the outputs of UIClip and other baselines to UIs rated by 12 human designers, we found that UIClip achieved the highest agreement with ground-truth rankings. Finally, we present three example applications that demonstrate how UIClip can facilitate downstream applications that rely on instantaneous assessment of UI design quality: i) UI code generation, ii) UI design tips generation, and iii) quality-aware UI example search.

Reality Rifts are interfaces between the physical and the virtual reality, where incoherent observations of physical behavior lead users to imagine comprehensive and plausible end-to-end dynamics. Reality Rifts emerge in interactive physical systems that lack one or more components that are central to their operation, yet where the physical end-to-end interaction persists with plausible outcomes. Even in the presence of a Reality Rift, users can still interact with a system—much like they would with the unaltered and complete counterpart—leading them to implicitly infer the existence and imagine the behavior of the lacking components from observable phenomena and outcomes. Therefore, dynamic systems with Reality Rifts trigger doubt, curiosity, and rumination—a sense of wonder that users experience when observing a Reality Rift due to their innate curiosity.

Sighted and blind and low vision (BLV) creators alike use videos to communicate with broad audiences. Yet, video editing remains inaccessible to BLV creators. Our formative study revealed that current video editing tools make it difficult to access the visual content, assess the visual quality, and efficiently navigate the timeline. We present AVscript, an accessible text-based video editor. AVscript enables users to edit their video using a script that embeds the video’s visual content, visual errors (e.g., dark or blurred footage), and speech. Users can also efficiently navigate between scenes and visual errors or locate objects in the frame or spoken words of interest. A comparison study (N=12) showed that AVscript significantly lowered BLV creators’ mental demands while increasing confidence and independence in video editing. We further demonstrate the potential of AVscript through an exploratory study (N=3) where BLV creators edited their own footage.

Presentation slides commonly use visual patterns for structural navigation, such as titles, dividers, and build slides. However, screen readers do not capture such intention, making it time-consuming and less accessible for blind and visually impaired (BVI) users to linearly consume slides with repeated content. We present Slide Gestalt, an automatic approach that identifies the hierarchical structure in a slide deck. Slide Gestalt computes the visual and textual correspondences between slides to generate hierarchical groupings. Readers can navigate the slide deck from the higher-level section overview to the lower-level description of a slide group or individual elements interactively with our UI. We derived side consumption and authoring practices from interviews with BVI readers and sighted creators and an analysis of 100 decks. We performed our pipeline with 50 real-world slide decks and a large dataset. Feedback from eight BVI participants showed that Slide Gestalt helped navigate a slide deck by anchoring content more efficiently, compared to using accessible slides.

Modeling user interfaces (UIs) from visual information allows systems to make inferences about the functionality and semantics needed to support use cases in accessibility, app automation, and testing. Current datasets for training machine learning models are limited in size due to the costly and time-consuming process of manually collecting and annotating UIs. We crawled the web to construct WebUI, a large dataset of 400,000 rendered web pages associated with automatically extracted metadata. We analyze the composition of WebUI and show that while automatically extracted data is noisy, most examples meet basic criteria for visual UI modeling. We applied several strategies for incorporating semantics found in web pages to increase the performance of visual UI understanding models in the mobile domain, where less labeled data is available: (i) element detection, (ii) screen classification and (iii) screen similarity.

Blind and low vision (BLV) creators use images to communicate with sighted audiences. However, creating or retrieving images is challenging for BLV creators as it is difficult to use authoring tools or assess image search results. Thus, creators limit the types of images they create or recruit sighted collaborators. While text-to-image generation models let creators generate high-fidelity images based on a text description (i.e. prompt), it is difficult to assess the content and quality of generated images. We present GenAssist, a system to make text-to-image generation accessible. Using our interface, creators can verify whether generated image candidates followed the prompt, access additional details in the image not specified in the prompt, and skim a summary of similarities and differences between image candidates. To power the interface, GenAssist uses a large language model to generate visual questions, vision-language models to extract answers, and a large language model to summarize the results. Our study with 12 BLV creators demonstrated that GenAssist enables and simplifies the process of image selection and generation, making visual authoring more accessible to all.

Blind users rely on alternative text (alt-text) to understand an image; however, alt-text is often missing. AI-generated captions are a more scalable alternative, but they often miss crucial details or are completely incorrect, which users may still falsely trust. In this work, we sought to determine how additional information could help users better judge the correctness of AI-generated captions. We developed ImageExplorer, a touch-based multi-layered image exploration system that allows users to explore the spatial layout and information hierarchies of images, and compared it with popular text-based (Facebook) and touch-based (Seeing AI) image exploration systems in a study with 12 blind participants. We found that exploration was generally successful in encouraging skepticism towards imperfect captions. Moreover, many participants preferred ImageExplorer for its multi-layered and spatial information presentation, and Facebook for its summary and ease of use. Finally, we identify design improvements for effective and explainable image exploration systems for blind users.

Visual slide-based presentations are ubiquitous, yet slide authoring tools are largely inaccessible to people who are blind or visually impaired (BVI). When authoring presentations, the 9 BVI presenters in our formative study usually work with sighted collaborators to produce visual slides based on the text content they produce. While BVI presenters valued collaborators’ visual design skill, the collaborators often felt they could not fully review and provide feedback on the visual changes that were made. We present Diffscriber, a system that identifies and describes changes to a slide’s content, layout, and style for presentation authoring. Using our system, BVI presentation authors can efficiently review changes to their presentation by navigating either a summary of high-level changes or individual slide elements. To learn more about changes of interest, presenters can use a generated change hierarchy to navigate to lower-level change details and element styles. BVI presenters using Diffscriber were able to identify slide design changes and provide feedback more easily as compared to using only the slides alone. More broadly, Diffscriber illustrates how advances in detecting and describing visual differences can improve mixed-ability collaboration.

Blind or visually impaired (BVI) individuals often rely on alternative text (alt-text) in order to understand an image; however, alt-text is often missing or incomplete. Automatically-generated captions are a more scalable alternative, but they are also often missing crucial details, and, sometimes, are completely incorrect, which may still be falsely trusted by BVI users. We hypothesize that additional information could help BVI users better judge the correctness of an auto-generated caption. To achieve this, we present Image Explorer, a touch-based multi-layered image exploration system that enables users to explore the spatial layout and information hierarchies in an image. Image Explorer leverages several off-the-shelf deep learning models to generate segmentation and labeling results for an image, combines and filters the generated information, and presents the resulted information in hierarchical layers. In a pilot study with three BVI users, participants used Image Explorer, Seeing AI, and Facebook to explore images with auto-generated captions of diverging quality, and judge the correctness of the captions. Preliminary results show that participants made more accurate judgements about the correctness of the captions when using Image Explorer, although they were highly confident about their judgement regardless of the tool used. Overall, Image Explorer is a novel touch exploration system that makes images more accessible for BVI users by potentially encouraging skepticism and enabling users to independently validate auto-generated captions.

We present Slidecho, a system that enables non-visual access of the slide content in a presentation video on-demand. Slidecho automatically extracts slides and their text and image elements from the presentation video and aligns these elements to the presenter’s speech. When listening to the video, Slidecho provides learners with audio notifications about slide changes and slide elements that are not described by the presenter. The learner can pause the video and browse the entire slide, or only the undescribed slide elements, to gain information. A technical evaluation with presentation videos in-the-wild shows that compared to the presenter’s speech alone, Slidecho provides access to an additional 20% of total text elements and 30% of total image elements that were previously not described. Blind and visually impaired participants in our user study reported that it was easier to locate undescribed slide elements with Slidecho’s synchronized interface than when browsing the video and extracted slides separately, and using Slidecho they read fewer slides that were fully redundant with the speech.

Presenters commonly use slides as visual aids for informative talks. When presenters fail to verbally describe the content on their slides, blind and visually impaired audience members lose access to necessary content, making the presentation difficult to follow. Our analysis of 90 presentation videos revealed that 72% of 610 visual elements (e.g., images, text) were insufficiently described. To help presenters create accessible presentations, we introduce Presentation A11y, a system that provides real-time and post-presentation accessibility feedback. Our system analyzes visual elements on the slide and the transcript of the verbal presentation to provide element-level feedback on what visual content needs to be further described or even removed. Presenters using our system with their own slide-based presentations described more of the content on their slides, and identified 3.26 times more accessibility problems to fix after the talk than when using a traditional slide-based presentation interface. Integrating accessibility feedback into content creation tools will improve the accessibility of informational content for all.

Virtual Reality (VR) sickness is common with symptoms such as headaches, nausea, and disorientation, and is a major barrier to using VR. We propose WalkingVibe, which applies unobtrusive vibrotactile feedback for VR walking experiences, and also reduces VR sickness and discomfort while improving realism. Feedback is delivered through two small vibration motors behind the ears at a frequency that strikes a balance in inducing vestibular response while minimizing annoyance. We conducted a 240-person study to explore how visual, audio, and various tactile feedback designs affect the locomotion experience of users walking passively in VR while seated statically in reality. Results showed timing and location for tactile feedback have significant effects on VR sickness and realism. With WalkingVibe, 2-sided step-synchronized design significantly reduces VR sickness and discomfort while significantly improving realism. Furthermore, its unobtrusiveness and ease of integration make WalkingVibe a practical approach for improving VR experiences with new and existing VR headsets.

Strength training improves overall health, well-being, physical appearance, and sports performance, with training programs specifying variables such as sets, repetitions, rest time, weight, and tempo. The repetitive nature of strength training, typically performed at fixed tempo, has made it challenging to develop entertaining exergames for common strength training exercises. We present StrengthGaming, which uses scaling and shuffling technique to provide tempo variations to strength training while preserving the training volume and training goals. It affords game designers more flexibility in designing strength training-based exergames. We developed a prototype game, inspired by FlappyBird, that uses a wearable orientation sensor to track the repetition tempo to control a flying character, using both fixed tempo design and dynamic tempo design. Results from our 24-person user study showed that dynamic tempo was significantly more entertaining than fixed tempo (p<0.01), and was preferred by participants.

Modern touchscreen devices have recently introduced customizable touchscreen settings to improve accessibility for users with motor impairments. For example, iOS 10 introduced the following four Touch Accommodation settings: 1) Hold Duration, 2) Ignore Repeat, 3) Tap Assistance, and 4) Tap Assistance Gesture Delay. These four independent settings lead to a total of more than 1 million possible configurations, making it impractical to manually determine the optimal settings. We present PersonalTouch, which collects and analyzes touchscreen gestures performed by individual users, and recommends personalized, optimal touchscreen accessibility settings. Results from our user study show that PersonalTouch significantly improves touch input success rate for users with motor impairments (20.2%, N=12, p=.00054) and for users without motor impairments (1.28%, N=12, p=.032).

Augmented Reality (AR) provides real-time information by superimposing virtual information onto users' view of the real world. Our work is the first to explore how peripheral vision, instead of central vision, can be used to read text on AR and smart glasses. We present Peritext, a multiword reading interface using rapid serial visual presentation (RSVP). This enables users to observe the real world using central vision, while using peripheral vision to read virtual information. We first conducted a lab-based study to determine the effect of different text transformation by comparing reading efficiency among 3 capitalization schemes, 2 font faces, 2 text animation methods, and 3 different numbers of words for RSVP paradigm. We found that title case capitalization, sans-serif font and word-wise typewriter animation with multiword RSVP display resulted in better reading efficiency, which together formed our Peritext design. Another lab-based study followed, investigating the performance of the Peritext against control text, and the results showed significant better performance. Finally, we conducted a field study to collect user feedback while using Peritext in real-world walking scenarios, and all users reported a preference of 5° eccentricity over 8°.

Virtual Reality (VR) sickness occurs when exposure to a virtual environment causes symptoms that are similar to motion sickness, and has been one of the major user experience barriers of VR. To reduce VR sickness, prior work has explored dynamic field-of-view modification and galvanic vestibular stimulation (GVS) that recou-ples the visual and vestibular systems. We propose a new approach to reduce VR sickness, called PhantomLegs, that applies alternating haptic cues that are synchronized to users' footsteps in VR. Our prototype consists of two servos with padded swing arms, one set on each side of the head, that lightly taps the head as users walk in VR. We conducted a three-session, multi-day user study with 30 participants to evaluate its effects as users navigate through a VR environment while physically being seated. Results show that our approach significantly reduces VR sickness during the initial exposure while remaining comfortable to users.

Deaf and hard-of-hearing (DHH) individuals encounter difficulties when engaged in group conversations with hearing individuals, due to factors such as simultaneous utterances from multiple speakers and speakers whom may be potentially out of view. We interviewed and co-designed with eight DHH participants to address the following challenges: 1) associating utterances with speakers, 2) ordering utterances from different speakers, 3) displaying optimal content length, and 4) visualizing utterances from out-of-view speakers. We evaluated multiple designs for each of the four challenges through a user study with twelve DHH participants. Our study results showed that participants significantly preferred speechbubble visualizations over traditional captions. These design preferences guided our development of SpeechBubbles, a real-time speech recognition interface prototype on an augmented reality head-mounted display. From our evaluations, we further demonstrated that DHH participants preferred our prototype over traditional captions for group conversations.