Proceedings of the SIGCHI conference on human factors in computing systems. CHI Conference最新文献

Chronic liver disease can lead to neurological conditions that result in coma or death. Although early detection can allow for intervention, testing is infrequent and unstandardized. Beacon is a device for at-home patient self-measurement of cognitive function via critical flicker frequency, which is the frequency at which a flickering light appears steady to an observer. This paper presents our efforts in iterating on Beacon's hardware and software to enable at-home use, then reports on an at-home deployment with 21 patients taking measurements over 6 weeks. We found that measurements were stable despite being taken at different times and in different environments. Finally, through interviews with 15 patients and 5 hepatologists, we report on participant experiences with Beacon, preferences around how CFF data should be presented, and the role of caregivers in helping patients manage their condition. Informed by our experiences with Beacon, we further discuss design implications for home health devices.

μEMAs allow participants to answer a short survey quickly with a tap on a smartwatch screen or a brief speech input. The short interaction time and low cognitive burden enable researchers to collect self-reports at high frequency (once every 5-15 minutes) while maintaining participant engagement. Systems with single input modality, however, may carry different contextual biases that could affect compliance. We combined two input modalities to create a multimodal-μEMA system, allowing participants to choose between speech or touch input to self-report. To investigate system usability, we conducted a seven-day field study where we asked 20 participants to label their posture and/or physical activity once every five minutes throughout their waking day. Despite the intense prompting interval, participants responded to 72.4% of the prompts. We found participants gravitated towards different modalities based on personal preferences and contextual states, highlighting the need to consider these factors when designing context-aware multimodal μEMA systems.

Health information technologies are transforming how mental healthcare is paid for through value-based care programs, which tie payment to data quantifying care outcomes. But, it is unclear what outcomes data these technologies should store, how to engage users in data collection, and how outcomes data can improve care. Given these challenges, we conducted interviews with 30 U.S.-based mental health clinicians to explore the design space of health information technologies that support outcomes data specification, collection, and use in value-based mental healthcare. Our findings center clinicians' perspectives on aligning outcomes data for payment programs and care; opportunities for health technologies and personal devices to improve data collection; and considerations for using outcomes data to hold stakeholders including clinicians, health insurers, and social services financially accountable in value-based mental healthcare. We conclude with implications for future research designing and developing technologies supporting value-based care across stakeholders involved with mental health service delivery.

Large multimodal models (LMMs) have enabled new AI-powered applications that help people with visual impairments (PVI) receive natural language descriptions of their surroundings through audible text. We investigated how this emerging paradigm of visual assistance transforms how PVI perform and manage their daily tasks. Moving beyond basic usability assessments, we examined both the capabilities and limitations of LMM-based tools in personal and social contexts, while exploring design implications for their future development. Through interviews with 14 visually impaired users and analysis of image descriptions from both participants and social media using Be My AI (an LMM-based application), we identified two key limitations. First, these systems' context awareness suffers from hallucinations and misinterpretations of social contexts, styles, and human identities. Second, their intent-oriented capabilities often fail to grasp and act on users' intentions. Based on these findings, we propose design strategies for improving both human-AI and AI-AI interactions, contributing to the development of more effective, interactive, and personalized assistive technologies.

Video descriptions are crucial for blind and low vision (BLV) users to access visual content. However, current artificial intelligence models for generating descriptions often fall short due to limitations in the quality of human annotations within training datasets, resulting in descriptions that do not fully meet BLV users' needs. To address this gap, we introduce VideoA11y, an approach that leverages multimodal large language models (MLLMs) and video accessibility guidelines to generate descriptions tailored for BLV individuals. Using this method, we have curated VideoA11y-40K, the largest and most comprehensive dataset of 40,000 videos described for BLV users. Rigorous experiments across 15 video categories, involving 347 sighted participants, 40 BLV participants, and seven professional describers, showed that VideoA11y descriptions outperform novice human annotations and are comparable to trained human annotations in clarity, accuracy, objectivity, descriptiveness, and user satisfaction. We evaluated models on VideoA11y-40K using both standard and custom metrics, demonstrating that MLLMs fine-tuned on this dataset produce high-quality accessible descriptions. Code and dataset are available at https://people-robots.github.io/VideoA11y/.

While voice input offers a convenient alternative to traditional text editing on mobile devices, practical implementations face two key challenges: 1) reliably distinguishing between editing commands and content dictation, and 2) effortlessly pinpointing the intended edit location. We propose Tap&Say, a novel multimodal system that combines touch interactions with Large Language Models (LLMs) for accurate text correction. By tapping near an error, users signal their edit intent and location, addressing both challenges. Then, the user speaks the correction text. Tap&Say utilizes the touch location, voice input, and existing text to generate contextually relevant correction suggestions. We propose a novel touch location-informed attention layer that integrates the tap location into the LLM's attention mechanism, enabling it to utilize the tap location for text correction. We fine-tuned the touch location-informed LLM on synthetic touch locations and correction commands, achieving significantly higher correction accuracy than the state-of-the-art method VT [45]. A 16-person user study demonstrated that Tap&Say outperforms VT [45] with 16.4% shorter task completion time and 47.5% fewer keyboard clicks and is preferred by users.

This paper characterizes the mental health technology "kits" of individuals managing depression: the specific technologies on their digital devices and physical items in their environments that people turn to as part of their mental health management. We interviewed 28 individuals living across the United States who use bundles of connected tools for both individual and collaborative mental health activities. We contribute to the HCI community by conceptualizing these tool assemblages that people managing depression have constructed over time. We detail categories of tools, describe kit characteristics (intentional, adaptable, available), and present participant ideas for future mental health support technologies. We then discuss what a mental health technology kit perspective means for researchers and designers and describe design principles (building within current toolkits; creating new tools from current self-management strategies; and identifying gaps in people's current kits) to support depression self-management across an evolving set of tools.

Engaging with people's lived experiences is foundational for HCI research and design. This paper introduces a novel narrative elicitation method to empower people to easily articulate 'micro-narratives' emerging from their lived experiences, irrespective of their writing ability or background. Our approach aims to enable at-scale collection of rich, co-created datasets that highlight target populations' voices with minimal participant burden, while precisely addressing specific research questions. To pilot this idea, and test its feasibility, we: (i) developed an AI-powered prototype, which leverages LLM-chaining to scaffold the cognitive steps necessary for users' narrative articulation; (ii) deployed it in three mixed-methods studies involving over 380 users; and (iii) consulted with established academics as well as C-level staff at (inter)national non-profits to map out potential applications. Both qualitative and quantitative findings show the acceptability and promise of the micro-narrative method, while also identifying the ethical and safeguarding considerations necessary for any at-scale deployments.

Landmarks are critical in navigation, supporting self-orientation and mental model development. Similar to sighted people, people with low vision (PLV) frequently look for landmarks via visual cues but face difficulties identifying some important landmarks due to vision loss. We first conducted a formative study with six PLV to characterize their challenges and strategies in landmark selection, identifying their unique landmark categories (e.g., area silhouettes, accessibility-related objects) and preferred landmark augmentations. We then designed VisiMark, an AR interface that supports landmark perception for PLV by providing both overviews of space structures and in-situ landmark augmentations. We evaluated VisiMark with 16 PLV and found that VisiMark enabled PLV to perceive landmarks they preferred but could not easily perceive before, and changed PLV's landmark selection from only visually-salient objects to cognitive landmarks that are more important and meaningful. We further derive design considerations for AR-based landmark augmentation systems for PLV.

Managing complex chronic illness is challenging due to its unpredictability. This paper explores the potential of voice for automated flare-up forecasts. We conducted a six-week speculative design study with individuals with endometriosis, tasking participants to submit daily voice recordings and symptom logs. Through focus groups, we elicited their experiences with voice capture and perceptions of its usefulness in forecasting flare-ups. Participants were enthusiastic and intrigued at the potential of flare-up forecasts through the analysis of their voice. They highlighted imagined benefits from the experience of recording in supporting emotional aspects of illness and validating both day-to-day and overall illness experiences. Participants reported that their recordings revolved around their endometriosis, suggesting that the recordings' content could further inform forecasting. We discuss potential opportunities and challenges in leveraging the voice as a data modality in human-centered AI tools that support individuals with complex chronic conditions.