arXiv subjects

Xuhai Xu

Publications and source records attributed to Xuhai Xu.

At least 37 records · Page 2Linked to original sources

arXiv2025

SensorLM: Learning the Language of Wearable Sensors

We present SensorLM, a family of sensor-language foundation models that enable wearable sensor data understanding with natural language. Despite its pervasive nature, aligning and interpreting sensor data with language remains challenging due to the lack of paired, richly annotated sensor-text descriptions in uncurated, real-world wearable data. We introduce a hierarchical caption generation pipeline designed to capture statistical, structural, and semantic information from sensor data. This approach enabled the curation of the largest sensor-language dataset to date, comprising over 59.7 million hours of data from more than 103,000 people. Furthermore, SensorLM extends prominent multimodal pretraining architectures (e.g., CLIP, CoCa) and recovers them as specific variants within a generic architecture. Extensive experiments on real-world tasks in human activity analysis and healthcare verify the superior performance of SensorLM over state-of-the-art in zero-shot recognition, few-shot learning, and cross-modal retrieval. SensorLM also demonstrates intriguing capabilities including scaling behaviors, label efficiency, sensor captioning, and zero-shot generalization to unseen tasks.

cs.LG↗

arXiv2025

LSM-2: Learning from Incomplete Wearable Sensor Data

Foundation models, a cornerstone of recent advancements in machine learning, have predominantly thrived on complete and well-structured data. Wearable sensor data frequently suffers from significant missingness, posing a substantial challenge for self-supervised learning (SSL) models that typically assume complete data inputs. This paper introduces the second generation of Large Sensor Model (LSM-2) with Adaptive and Inherited Masking (AIM), a novel SSL approach that learns robust representations directly from incomplete data without requiring explicit imputation. AIM's core novelty lies in its use of learnable mask tokens to model both existing ("inherited") and artificially introduced missingness, enabling it to robustly handle fragmented real-world data during inference. Pre-trained on an extensive dataset of 40M hours of day-long multimodal sensor data, our LSM-2 with AIM achieves the best performance across a diverse range of tasks, including classification, regression and generative modeling. Furthermore, LSM-2 with AIM exhibits superior scaling performance, and critically, maintains high performance even under targeted missingness scenarios, reflecting clinically coherent patterns, such as the diagnostic value of nighttime biosignals for hypertension prediction. This makes AIM a more reliable choice for real-world wearable data applications.

cs.LG↗

arXiv2025

A Survey of Passive Sensing for Workplace Wellbeing and Productivity

The modern workplace is undergoing a radical transformation, driven by technological advances that blur the boundaries between human capability and digital augmentation. At the forefront of this evolution is passive sensing technology - a suite of tools that quietly monitor and interpret human behavior without active user engagement. This paper examines how these technologies are reshaping our understanding of workplace dynamics, with a particular focus on employee wellbeing and productivity. Through a comprehensive review of recent research, we explore both the transformative potential and inherent challenges of passive sensing in professional environments. Our analysis reveals emerging patterns in how these technologies can support worker health and performance, while also highlighting critical gaps in current research and opportunities for future innovation. We conclude by outlining a roadmap for integrating passive sensing into future workplaces in ways that enhance human potential while preserving dignity and autonomy.

cs.HC↗

arXiv2025

More Modality, More AI: Exploring Design Opportunities of AI-Based Multi-modal Remote Monitoring Technologies for Early Detection of Mental Health Sequelae in Youth Concussion Patients

Anxiety, depression, and suicidality are common mental health sequelae following concussion in youth patients, often exacerbating concussion symptoms and prolonging recovery. Despite the critical need for early detection of these mental health symptoms, clinicians often face challenges in accurately collecting patients' mental health data and making clinical decision-making in a timely manner. Today's remote patient monitoring (RPM) technologies offer opportunities to objectively monitor patients' activities, but they were not specifically designed for youth concussion patients; moreover, the large amount of data collected by RPM technologies may also impose significant workloads on clinicians to keep up with and use the data. To address these gaps, we employed a three-stage study consisting of a formative study, interface design, and design evaluation. We first conducted a formative study through semi-structured interviews with six highly professional concussion clinicians and identified clinicians' key challenges in remotely collecting patient information and accessing patient treatment compliance. Subsequently, we proposed preliminary clinician-facing interface designs with the integration of AI-based RPM technologies (AI-RPM), followed by design evaluation sessions with highly professional concussion clinicians. Clinicians underscored the value of integrating multi-modal AI-RPM technologies to support their decision-making while emphasizing the importance of customizable interfaces through collaborative design and multiple responsible design considerations.

cs.HC↗

arXiv2025

The Odyssey Journey: Top-Tier Medical Resource Seeking for Specialized Disorder in China

It is pivotal for patients to receive accurate health information, diagnoses, and timely treatments. However, in China, the significant imbalanced doctor-to-patient ratio intensifies the information and power asymmetries in doctor-patient relationships. Health information-seeking, which enables patients to collect information from sources beyond doctors, is a potential approach to mitigate these asymmetries. While HCI research predominantly focuses on common chronic conditions, our study focuses on specialized disorders, which are often familiar to specialists but not to general practitioners and the public. With Hemifacial Spasm (HFS) as an example, we aim to understand patients' health information and top-tier medical resource seeking journeys in China. Through interviews with three neurosurgeons and 12 HFS patients from rural and urban areas, and applying Actor-Network Theory, we provide empirical insights into the roles, interactions, and workflows of various actors in the health information-seeking network. We also identified five strategies patients adopted to mitigate asymmetries and access top-tier medical resources, illustrating these strategies as subnetworks within the broader health information-seeking network and outlining their advantages and challenges.

cs.HC↗

arXiv2025

WatchGuardian: Enabling User-Defined Personalized Just-in-Time Intervention on Smartwatch

While just-in-time interventions (JITIs) have effectively targeted common health behaviors, individuals often have unique needs to intervene in personal undesirable actions that can negatively affect physical, mental, and social well-being. We present WatchGuardian, a smartwatch-based JITI system that empowers users to define custom interventions for these personal actions with a small number of samples. For the model to detect new actions based on limited new data samples, we developed a few-shot learning pipeline that finetuned a pre-trained inertial measurement unit (IMU) model on public hand-gesture datasets. We then designed a data augmentation and synthesis process to train additional classification layers for customization. Our offline evaluation with 26 participants showed that with three, five, and ten examples, our approach achieved an average accuracy of 76.8%, 84.7%, and 87.7%, and an F1 score of 74.8%, 84.2%, and 87.2% We then conducted a four-hour intervention study to compare WatchGuardian against a rule-based intervention. Our results demonstrated that our system led to a significant reduction by 64.0 +- 22.6% in undesirable actions, substantially outperforming the baseline by 29.0%. Our findings underscore the effectiveness of a customizable, AI-driven JITI system for individuals in need of behavioral intervention in personal undesirable actions. We envision that our work can inspire broader applications of user-defined personalized intervention with advanced AI solutions.

cs.HC↗

arXiv2025

Human-inspired Perspectives: A Survey on AI Long-term Memory

With the rapid advancement of AI systems, their abilities to store, retrieve, and utilize information over the long term - referred to as long-term memory - have become increasingly significant. These capabilities are crucial for enhancing the performance of AI systems across a wide range of tasks. However, there is currently no comprehensive survey that systematically investigates AI's long-term memory capabilities, formulates a theoretical framework, and inspires the development of next-generation AI long-term memory systems. This paper begins by introducing the mechanisms of human long-term memory, then explores AI long-term memory mechanisms, establishing a mapping between the two. Based on the mapping relationships identified, we extend the current cognitive architectures and propose the Cognitive Architecture of Self-Adaptive Long-term Memory (SALM). SALM provides a theoretical framework for the practice of AI long-term memory and holds potential for guiding the creation of next-generation long-term memory driven AI systems. Finally, we delve into the future directions and application prospects of AI long-term memory.

cs.AI↗

arXiv2024

A Demonstration of Adaptive Collaboration of Large Language Models for Medical Decision-Making

Medical Decision-Making (MDM) is a multi-faceted process that requires clinicians to assess complex multi-modal patient data patient, often collaboratively. Large Language Models (LLMs) promise to streamline this process by synthesizing vast medical knowledge and multi-modal health data. However, single-agent are often ill-suited for nuanced medical contexts requiring adaptable, collaborative problem-solving. Our MDAgents addresses this need by dynamically assigning collaboration structures to LLMs based on task complexity, mimicking real-world clinical collaboration and decision-making. This framework improves diagnostic accuracy and supports adaptive responses in complex, real-world medical scenarios, making it a valuable tool for clinicians in various healthcare settings, and at the same time, being more efficient in terms of computing cost than static multi-agent decision making methods.

cs.CL↗

arXiv2024

MDAgents: An Adaptive Collaboration of LLMs for Medical Decision-Making

Foundation models are becoming valuable tools in medicine. Yet despite their promise, the best way to leverage Large Language Models (LLMs) in complex medical tasks remains an open question. We introduce a novel multi-agent framework, named Medical Decision-making Agents (MDAgents) that helps address this gap by automatically assigning a collaboration structure to a team of LLMs. The assigned solo or group collaboration structure is tailored to the medical task at hand, emulating real-world medical decision-making processes adapted to tasks of varying complexities. We evaluate our framework and baseline methods using state-of-the-art LLMs across a suite of real-world medical knowledge and medical diagnosis benchmarks, including a comparison of LLMs' medical complexity classification against human physicians. MDAgents achieved the best performance in seven out of ten benchmarks on tasks requiring an understanding of medical knowledge and multi-modal reasoning, showing a significant improvement of up to 4.2% (p < 0.05) compared to previous methods' best performances. Ablation studies reveal that MDAgents effectively determines medical complexity to optimize for efficiency and accuracy across diverse medical tasks. Notably, the combination of moderator review and external medical knowledge in group collaboration resulted in an average accuracy improvement of 11.8%. Our code can be found at https://github.com/mitmedialab/MDAgents.

cs.CL↗

arXiv2024

Scaling Wearable Foundation Models

Wearable sensors have become ubiquitous thanks to a variety of health tracking features. The resulting continuous and longitudinal measurements from everyday life generate large volumes of data; however, making sense of these observations for scientific and actionable insights is non-trivial. Inspired by the empirical success of generative modeling, where large neural networks learn powerful representations from vast amounts of text, image, video, or audio data, we investigate the scaling properties of sensor foundation models across compute, data, and model size. Using a dataset of up to 40 million hours of in-situ heart rate, heart rate variability, electrodermal activity, accelerometer, skin temperature, and altimeter per-minute data from over 165,000 people, we create LSM, a multimodal foundation model built on the largest wearable-signals dataset with the most extensive range of sensor modalities to date. Our results establish the scaling laws of LSM for tasks such as imputation, interpolation and extrapolation, both across time and sensor modalities. Moreover, we highlight how LSM enables sample-efficient downstream learning for tasks like exercise and activity recognition.

cs.LG↗

arXiv2024

Can AI Relate: Testing Large Language Model Response for Mental Health Support

Large language models (LLMs) are already being piloted for clinical use in hospital systems like NYU Langone, Dana-Farber and the NHS. A proposed deployment use case is psychotherapy, where a LLM-powered chatbot can treat a patient undergoing a mental health crisis. Deployment of LLMs for mental health response could hypothetically broaden access to psychotherapy and provide new possibilities for personalizing care. However, recent high-profile failures, like damaging dieting advice offered by the Tessa chatbot to patients with eating disorders, have led to doubt about their reliability in high-stakes and safety-critical settings. In this work, we develop an evaluation framework for determining whether LLM response is a viable and ethical path forward for the automation of mental health treatment. Our framework measures equity in empathy and adherence of LLM responses to motivational interviewing theory. Using human evaluation with trained clinicians and automatic quality-of-care metrics grounded in psychology research, we compare the responses provided by peer-to-peer responders to those provided by a state-of-the-art LLM. We show that LLMs like GPT-4 use implicit and explicit cues to infer patient demographics like race. We then show that there are statistically significant discrepancies between patient subgroups: Responses to Black posters consistently have lower empathy than for any other demographic group (2%-13% lower than the control group). Promisingly, we do find that the manner in which responses are generated significantly impacts the quality of the response. We conclude by proposing safety guidelines for the potential deployment of LLMs for mental health response.

cs.CL↗

arXiv2024

Boosting Gesture Recognition with an Automatic Gesture Annotation Framework

Training a real-time gesture recognition model heavily relies on annotated data. However, manual data annotation is costly and demands substantial human effort. In order to address this challenge, we propose a framework that can automatically annotate gesture classes and identify their temporal ranges. Our framework consists of two key components: (1) a novel annotation model that leverages the Connectionist Temporal Classification (CTC) loss, and (2) a semi-supervised learning pipeline that enables the model to improve its performance by training on its own predictions, known as pseudo labels. These high-quality pseudo labels can also be used to enhance the accuracy of other downstream gesture recognition models. To evaluate our framework, we conducted experiments using two publicly available gesture datasets. Our ablation study demonstrates that our annotation model design surpasses the baseline in terms of both gesture classification accuracy (3-4% improvement) and localization accuracy (71-75% improvement). Additionally, we illustrate that the pseudo-labeled dataset produced from the proposed framework significantly boosts the accuracy of a pre-trained downstream gesture recognition model by 11-18%. We believe that this annotation framework has immense potential to improve the training of downstream gesture recognition models using unlabeled datasets.

cs.CV↗

arXiv2024

G-VOILA: Gaze-Facilitated Information Querying in Daily Scenarios

Modern information querying systems are progressively incorporating multimodal inputs like vision and audio. However, the integration of gaze -- a modality deeply linked to user intent and increasingly accessible via gaze-tracking wearables -- remains underexplored. This paper introduces a novel gaze-facilitated information querying paradigm, named G-VOILA, which synergizes users' gaze, visual field, and voice-based natural language queries to facilitate a more intuitive querying process. In a user-enactment study involving 21 participants in 3 daily scenarios (p = 21, scene = 3), we revealed the ambiguity in users' query language and a gaze-voice coordination pattern in users' natural query behaviors with G-VOILA. Based on the quantitative and qualitative findings, we developed a design framework for the G-VOILA paradigm, which effectively integrates the gaze data with the in-situ querying context. Then we implemented a G-VOILA proof-of-concept using cutting-edge deep learning techniques. A follow-up user study (p = 16, scene = 2) demonstrates its effectiveness by achieving both higher objective score and subjective score, compared to a baseline without gaze data. We further conducted interviews and provided insights for future gaze-facilitated information querying systems.

cs.HC↗

arXiv2024

Illuminating the Unseen: Investigating the Context-induced Harms in Behavioral Sensing

Behavioral sensing technologies are rapidly evolving across a range of well-being applications. Despite its potential, concerns about the responsible use of such technology are escalating. In response, recent research within the sensing technology has started to address these issues. While promising, they primarily focus on broad demographic categories and overlook more nuanced, context-specific identities. These approaches lack grounding within domain-specific harms that arise from deploying sensing technology in diverse social, environmental, and technological settings. Additionally, existing frameworks for evaluating harms are designed for a generic ML life cycle, and fail to adapt to the dynamic and longitudinal considerations for behavioral sensing technology. To address these gaps, we introduce a framework specifically designed for evaluating behavioral sensing technologies. This framework emphasizes a comprehensive understanding of context, particularly the situated identities of users and the deployment settings of the sensing technology. It also highlights the necessity for iterative harm mitigation and continuous maintenance to adapt to the evolving nature of technology and its use. We demonstrate the feasibility and generalizability of our framework through post-hoc evaluations on two real-world behavioral sensing studies conducted in different international contexts, involving varied population demographics and machine learning tasks. Our evaluations provide empirical evidence of both situated identity-based harm and more domain-specific harms, and discuss the trade-offs introduced by implementing bias mitigation techniques.

cs.HC↗

arXiv2024

Health-LLM: Large Language Models for Health Prediction via Wearable Sensor Data

Large language models (LLMs) are capable of many natural language tasks, yet they are far from perfect. In health applications, grounding and interpreting domain-specific and non-linguistic data is crucial. This paper investigates the capacity of LLMs to make inferences about health based on contextual information (e.g. user demographics, health knowledge) and physiological data (e.g. resting heart rate, sleep minutes). We present a comprehensive evaluation of 12 state-of-the-art LLMs with prompting and fine-tuning techniques on four public health datasets (PMData, LifeSnaps, GLOBEM and AW_FB). Our experiments cover 10 consumer health prediction tasks in mental health, activity, metabolic, and sleep assessment. Our fine-tuned model, HealthAlpaca exhibits comparable performance to much larger models (GPT-3.5, GPT-4 and Gemini-Pro), achieving the best performance in 8 out of 10 tasks. Ablation studies highlight the effectiveness of context enhancement strategies. Notably, we observe that our context enhancement can yield up to 23.8% improvement in performance. While constructing contextually rich prompts (combining user context, health knowledge and temporal information) exhibits synergistic improvement, the inclusion of health knowledge context in prompts significantly enhances overall performance.

cs.CL↗

arXiv2024

"I Wish There Were an AI": Challenges and AI Potential in Cancer Patient-Provider Communication

Patient-provider communication has been crucial to cancer patients' survival after their cancer treatments. However, the research community and patients themselves often overlook the communication challenges after cancer treatments as they are overshadowed by the severity of the patient's illness and the variety and rarity of the cancer disease itself. Meanwhile, the recent technical advances in AI, especially in Large Language Models (LLMs) with versatile natural language interpretation and generation ability, demonstrate great potential to support communication in complex real-world medical situations. By interviewing six healthcare providers and eight cancer patients, our goal is to explore the providers' and patients' communication barriers in the post-cancer treatment recovery period, their expectations for future communication technologies, and the potential of AI technologies in this context. Our findings reveal several challenges in current patient-provider communication, including the knowledge and timing gaps between cancer patients and providers, their collaboration obstacles, and resource limitations. Moreover, based on providers' and patients' needs and expectations, we summarize a set of design implications for intelligent communication systems, especially with the power of LLMs. Our work sheds light on the design of future AI-powered systems for patient-provider communication under high-stake and high-uncertainty situations.

cs.HC↗

arXiv2024

Fast-Forward Reality: Authoring Error-Free Context-Aware Policies with Real-Time Unit Tests in Extended Reality

Advances in ubiquitous computing have enabled end-user authoring of context-aware policies (CAPs) that control smart devices based on specific contexts of the user and environment. However, authoring CAPs accurately and avoiding run-time errors is challenging for end-users as it is difficult to foresee CAP behaviors under complex real-world conditions. We propose Fast-Forward Reality, an Extended Reality (XR) based authoring workflow that enables end-users to iteratively author and refine CAPs by validating their behaviors via simulated unit test cases. We develop a computational approach to automatically generate test cases based on the authored CAP and the user's context history. Our system delivers each test case with immersive visualizations in XR, facilitating users to verify the CAP behavior and identify necessary refinements. We evaluated Fast-Forward Reality in a user study (N=12). Our authoring and validation process improved the accuracy of CAPs and the users provided positive feedback on the system usability.

cs.HC↗

arXiv2024

From Text to Self: Users' Perceptions of Potential of AI on Interpersonal Communication and Self

In the rapidly evolving landscape of AI-mediated communication (AIMC), tools powered by Large Language Models (LLMs) are becoming integral to interpersonal communication. Employing a mixed-methods approach, we conducted a one-week diary and interview study to explore users' perceptions of these tools' ability to: 1) support interpersonal communication in the short-term, and 2) lead to potential long-term effects. Our findings indicate that participants view AIMC support favorably, citing benefits such as increased communication confidence, and finding precise language to express their thoughts, navigating linguistic and cultural barriers. However, the study also uncovers current limitations of AIMC tools, including verbosity, unnatural responses, and excessive emotional intensity. These shortcomings are further exacerbated by user concerns about inauthenticity and potential overreliance on the technology. Furthermore, we identified four key communication spaces delineated by communication stakes (high or low) and relationship dynamics (formal or informal) that differentially predict users' attitudes toward AIMC tools. Specifically, participants found the tool is more suitable for communicating in formal relationships than informal ones and more beneficial in high-stakes than low-stakes communication.

cs.HC↗