arXiv subjects

Yiwen Dong

Publications and source records attributed to Yiwen Dong.

16 recordsLinked to original sources

arXiv2026

Designing Computerized Gait Analysis for Pediatric Care: Clinician Perspectives on Sensing, Workflow, and Care Environments

Computerized gait analysis (CGA) serves as an essential diagnostic tool for evaluating neuromuscular, musculoskeletal, and neurological disorders in children, from cerebral palsy to muscular dystrophy. By enabling objective and comprehensive gait analysis, CGA supports timely clinical interventions that can significantly improve pediatric mobility outcomes and quality of life. Yet pediatric gait analysis introduces unique design considerations often underexplored in existing CGA research, as children's ongoing development shapes assessment requirements. To understand how CGA technologies can be designed for pediatric care, we conducted a qualitative study with 12 pediatric clinicians and one system designer who routinely work with CGA. Participants identified child-specific challenges including managing heightened sensory sensitivities to wearable devices, accommodating body proportions in sensor placement and calibration, and maintaining patient engagement during data collection. Clinicians also articulated needs for workflow adaptations and expressed interest in extending gait analysis beyond controlled laboratory settings into naturalistic environments such as playgrounds and schools, where children's authentic movement patterns emerge. Drawing from these clinician perspectives, we present design recommendations for pediatric-centered CGA that address sensing modalities suitable for sensory-sensitive children and approaches for capturing gait data across diverse care environments. Our findings contribute to the broader challenge of adapting clinical technologies to meet the distinct needs of pediatric populations.

cs.HC↗

arXiv2026

Do VLMs See What Sensors Feel? A Scalable Expert-Guided Design for Wheelchair Accessibility Assessment from Street View

Assessing built-environment interaction, such as wheelchair accessibility, is difficult because real-world mobility is shaped by distributed, context-dependent, and temporary barriers that are hard to capture at scale. To support scalable assessment, this paper examines whether vision-language models (VLMs) can identify accessibility barriers from Google Street View (GSV) imagery. We propose an expert-guided retrieval-augmented framework that combines GSV images, ADA-informed guidance, and expert-derived rubrics to evaluate accessibility dimensions. We collect a campus-scale dataset at the University of Florida, linking 407 unique GSV locations with GPS-derived wheelchair dwell behavior as a mobility-friction signal. Results show that VLM ratings are both negatively correlated and distributionally similar with dwell time, indicating partial but consistent alignment with a behavioral proxy for mobility friction. Visual cue analysis shows that certain environmental objects, such as curb ramps and crosswalks, are associated with higher VLM accessibility scores, while alignment remains limited for subtle surface conditions, transient obstructions, and viewpoint-dependent barriers. Overall, our findings show the potential of expert-guided VLMs for scalable accessibility assessment aligning with sensor-derived indicators of real-world wheelchair navigation.

cs.CV↗

arXiv2026

Higher Satisfaction, Lower Cost: A Technical Report on How LLMs Revolutionize Meituan's Intelligent Interaction Systems

Enhancing customer experience is essential for business success, particularly as service demands grow in scale and complexity. Generative artificial intelligence and Large Language Models (LLMs) have empowered intelligent interaction systems to deliver efficient, personalized, and 24/7 support. In practice, intelligent interaction systems encounter several challenges: (1) Constructing high-quality data for cold-start training is difficult, hindering self-evolution and raising labor costs. (2) Multi-turn dialogue performance remains suboptimal due to inadequate intent understanding, rule compliance, and solution extraction. (3) Frequent evolution of business rules affects system operability and transferability, constraining low-cost expansion and adaptability. (4) Reliance on a single LLM is insufficient in complex scenarios, where the absence of multi-agent frameworks and effective collaboration undermines process completeness and service quality. (5) The open-domain nature of multi-turn dialogues, lacking unified golden answers, hampers quantitative evaluation and continuous optimization. To address these challenges, we introduce WOWService, an intelligent interaction system tailored for industrial applications. With the integration of LLMs and multi-agent architectures, WOWService enables autonomous task management and collaborative problem-solving. Specifically, WOWService focuses on core modules including data construction, general capability enhancement, business scenario adaptation, multi-agent coordination, and automated evaluation. Currently, WOWService is deployed on the Meituan App, achieving significant gains in key metrics, e.g., User Satisfaction Metric 1 (USM 1) -27.53% and User Satisfaction Metric 2 (USM 2) +25.51%, demonstrating its effectiveness in capturing user needs and advancing personalized service.

cs.CL↗

arXiv2025

Unmasking the Genuine Type Inference Capabilities of LLMs for Java Code Snippets

Type inference is crucial for reusing online code snippets. Although snippets are prevalently shared on platforms like StackOverflow, they often lack essential type information, such as fully qualified names (FQNs). Recent studies have leveraged Large Language Models (LLMs) to perform type inference for such code snippets, showing promising results. However, these results may suffer from data leakage, as the benchmark, StatType-SO, used for evaluation has been publicly available on GitHub since 2017. Consequently, it remains uncertain whether the strong performance of LLMs reflects genuine semantic understanding of code or is due to the ground truth being included in the training set. This paper strives to comprehensively evaluate the genuine type inference capabilities of LLMs on Java code snippets and identify potential limitations of LLMs. First, we created ThaliaType, a new, previously unreleased benchmark suite designed for type inference evaluation. Second, using the StarCoder2 LLM as baseline, we uncovered data leakage from StatType-SO in StarCoder2's open-source training set and observed that other state-of-the-art LLMs exhibit similar performance drops when evaluated on ThaliaType, with precision decreasing by up to 59% and recall by up to 72%. Finally, we designed semantic-preserving code transformations to test the capabilities of LLMs in understanding the execution semantics of snippets. Results showed that LLMs' performance on StatType-SO is far less robust to these transformations than on ThaliaType, suggesting that the performance on StatType-SO may be biased by data leakage and have limited generalizability. These findings highlight the importance of carefully designed, leakage-free benchmarks for evaluating LLMs on type inference tasks. We recommend future studies adopt ThaliaType for rigorous and reliable assessments of LLMs' genuine type inference capabilities.

cs.SE↗

arXiv2025

Leveraging Audio Representations for Vibration-Based Crowd Monitoring in Stadiums

Crowd monitoring in sports stadiums is important to enhance public safety and improve the audience experience. Existing approaches mainly rely on cameras and microphones, which can cause significant disturbances and often raise privacy concerns. In this paper, we sense floor vibration, which provides a less disruptive and more non-intrusive way of crowd sensing, to predict crowd behavior. However, since the vibration-based crowd monitoring approach is newly developed, one main challenge is the lack of training data due to sports stadiums being large public spaces with complex physical activities. In this paper, we present ViLA (Vibration Leverage Audio), a vibration-based method that reduces the dependency on labeled data by pre-training with unlabeled cross-modality data. ViLA is first pre-trained on audio data in an unsupervised manner and then fine-tuned with a minimal amount of in-domain vibration data. By leveraging publicly available audio datasets, ViLA learns the wave behaviors from audio and then adapts the representation to vibration, reducing the reliance on domain-specific vibration data. Our real-world experiments demonstrate that pre-training the vibration model using publicly available audio data (YouTube8M) achieved up to a 5.8x error reduction compared to the model without audio pre-training.

cs.SD↗

arXiv2025

Personalized Emotion Detection from Floor Vibrations Induced by Footsteps

Emotion recognition is critical for various applications such as early detection of mental health disorders and emotion based smart home systems. Previous studies used various sensing methods for emotion recognition, such as wearable sensors, cameras, and microphones. However, these methods have limitations in long term domestic, including intrusiveness and privacy concerns. To overcome these limitations, this paper introduces a nonintrusive and privacy friendly personalized emotion recognition system, EmotionVibe, which leverages footstep induced floor vibrations for emotion recognition. The main idea of EmotionVibe is that individuals' emotional states influence their gait patterns, subsequently affecting the floor vibrations induced by their footsteps. However, there are two main research challenges: 1) the complex and indirect relationship between human emotions and footstep induced floor vibrations and 2) the large between person variations within the relationship between emotions and gait patterns. To address these challenges, we first empirically characterize this complex relationship and develop an emotion sensitive feature set including gait related and vibration related features from footstep induced floor vibrations. Furthermore, we personalize the emotion recognition system for each user by calculating gait similarities between the target person (i.e., the person whose emotions we aim to recognize) and those in the training dataset and assigning greater weights to training people with similar gait patterns in the loss function. We evaluated our system in a real-world walking experiment with 20 participants, summing up to 37,001 footstep samples. EmotionVibe achieved the mean absolute error (MAE) of 1.11 and 1.07 for valence and arousal score estimations, respectively, reflecting 19.0% and 25.7% error reduction compared to the baseline method.

eess.SY↗

arXiv2025

Bridging Structural Dynamics and Biomechanics: Human Motion Estimation through Footstep-Induced Floor Vibrations

Quantitative estimation of human joint motion in daily living spaces is essential for early detection and rehabilitation tracking of neuromusculoskeletal disorders (e.g., Parkinson's) and mitigating trip and fall risks for older adults. Existing approaches involve monitoring devices such as cameras, wearables, and pressure mats, but have operational constraints such as direct line-of-sight, carrying devices, and dense deployment. To overcome these limitations, we leverage gait-induced floor vibration to estimate lower-limb joint motion (e.g., ankle, knee, and hip flexion angles), allowing non-intrusive and contactless gait health monitoring in people's living spaces. To overcome the high uncertainty in lower-limb movement given the limited information provided by the gait-induced floor vibrations, we formulate a physics-informed graph to integrate domain knowledge of gait biomechanics and structural dynamics into the model. Specifically, different types of nodes represent heterogeneous information from joint motions and floor vibrations; Their connecting edges represent the physiological relationships between joints and forces governed by gait biomechanics, as well as the relationships between forces and floor responses governed by the structural dynamics. As a result, our model poses physical constraints to reduce uncertainty while allowing information sharing between the body and the floor to make more accurate predictions. We evaluate our approach with 20 participants through a real-world walking experiment. We achieved an average of 3.7 degrees of mean absolute error in estimating 12 joint flexion angles (38% error reduction from baseline), which is comparable to the performance of cameras and wearables in current medical practices.

cs.HC↗

arXiv2025

Continual Person Identification using Footstep-Induced Floor Vibrations on Heterogeneous Floor Structures

Person identification is important for smart buildings to provide personalized services such as health monitoring, activity tracking, and personnel management. However, previous person identification relies on pre-collected data from everyone, which is impractical in many buildings and public facilities in which visitors are typically expected. This calls for a continual person identification system that gradually learns people's identities on the fly. Existing studies use cameras to achieve this goal, but they require direct line-of-sight and also have raised privacy concerns in public. Other modalities such as wearables and pressure mats are limited by the requirement of device-carrying or dense deployment. Thus, prior studies introduced footstep-induced structural vibration sensing, which is non-intrusive and perceived as more privacy-friendly. However, this approach has a significant challenge: the high variability of vibration data due to structural heterogeneity and human gait variations, which makes online person identification algorithms perform poorly. In this paper, we characterize the variability in footstep-induced structural vibration data for accurate online person identification. To achieve this, we quantify and decompose different sources of variability and then design a feature transformation function to reduce the variability within each person's data to make different people's data more separable. We evaluate our approach through field experiments with 20 people. The results show a 70% variability reduction and a 90% accuracy for online person identification.

cs.CV↗

arXiv2025

WeVibe: Weight Change Estimation Through Audio-Induced Shelf Vibrations In Autonomous Stores

Weight change estimation is crucial in various applications, particularly for detecting pick-up and put-back actions when people interact with the shelf while shopping in autonomous stores. Moreover, accurate weight change estimation allows autonomous stores to automatically identify items being picked up or put back, ensuring precise cost estimation. However, the conventional approach of estimating weight changes requires specialized weight-sensing shelves, which are densely deployed weight scales, incurring intensive sensor consumption and high costs. Prior works explored the vibration-based weight sensing method, but they failed when the location of weight change varies. In response to these limitations, we made the following contributions: (1) We propose WeVibe, a first item weight change estimation system through active shelf vibration sensing. The main intuition of the system is that the weight placed on the shelf influences the dynamic vibration response of the shelf, thus altering the shelf vibration patterns. (2) We model a physics-informed relationship between the shelf vibration response and item weight across multiple locations on the shelf based on structural dynamics theory. This relationship is linear and allows easy training of a weight estimation model at a new location without heavy data collection. (3) We evaluate our system on a gondola shelf organized as the real-store settings. WeVibe achieved a mean absolute error down to 38.07g and a standard deviation of 31.2g with one sensor and 10% samples from three weight classes on estimating weight change from 0g to 450g, which can be leveraged for differentiating items with more than 100g differences.

eess.SP↗

arXiv2024

Detecting Gait Abnormalities in Foot-Floor Contacts During Walking Through Footstep-Induced Structural Vibrations

Gait abnormality detection is critical for the early discovery and progressive tracking of musculoskeletal and neurological disorders, such as Parkinson's and Cerebral Palsy. Especially, analyzing the foot-floor contacts during walking provides important insights into gait patterns, such as contact area, contact force, and contact time, enabling gait abnormality detection through these measurements. Existing studies use various sensing devices to capture such information, including cameras, wearables, and force plates. However, the former two lack force-related information, making it difficult to identify the causes of gait health issues, while the latter has limited coverage of the walking path. In this study, we leverage footstep-induced structural vibrations to infer foot-floor contact profiles and detect gait abnormalities. The main challenge lies in modeling the complex force transfer mechanism between the foot and the floor surfaces, leading to difficulty in reconstructing the force and contact profile during foot-floor interaction using structural vibrations. To overcome the challenge, we first characterize the floor vibration for each contact type (e.g., heel, midfoot, and toe contact) to understand how contact forces and areas affect the induced floor vibration. Then, we leverage the time-frequency response spectrum resulting from those contacts to develop features that are representative of each contact type. Finally, gait abnormalities are detected by comparing the predicted foot-floor contact force and motion with the healthy gait. To evaluate our approach, we conducted a real-world walking experiment with 8 subjects. Our approach achieves 91.6% and 96.7% accuracy in predicting contact type and time, respectively, leading to 91.9% accuracy in detecting various types of gait abnormalities, including asymmetry, dragging, and midfoot/toe contacts.

eess.SP↗

arXiv2024

LPR: Large Language Models-Aided Program Reduction

Program reduction is a prevalent technique to facilitate compilers' debugging by automatically minimizing bug-triggering programs. Existing program reduction techniques are either generic across languages (e.g., Perses and Vulcan) or specifically customized for one certain language by employing language-specific features, like C-Reduce. However, striking the balance between generality across multiple programming languages and specificity to individual languages in program reduction is yet to be explored. This paper proposes LPR, the first technique utilizing LLMs to perform language-specific program reduction for multiple languages. The core insight is to utilize both the language-generic syntax level program reduction (e.g., Perses) and the language-specific semantic level program transformations learned by LLMs. Alternately, language-generic program reducers efficiently reduce programs into 1-tree-minimality, which is small enough to be manageable for LLMs; LLMs effectively transform programs via the learned semantics to expose new reduction opportunities for the language-generic program reducers to further reduce the programs. Our extensive evaluation on 50 benchmarks across three languages (C, Rust, and JavaScript) has highlighted LPR's practicality and superiority over Vulcan, the state-of-the-art language-generic program reducer. For effectiveness, LPR surpasses Vulcan by producing 24.93%, 4.47%, and 11.71% smaller programs on benchmarks in C, Rust and JavaScript. Moreover, LPR and Vulcan have demonstrated their potential to complement each other. By using Vulcan on LPR's output for C programs, we achieve program sizes comparable to those reduced by C-Reduce. For efficiency, LPR takes 10.77%, 34.88%, 36.96% less time than Vulcan to finish all benchmarks in C, Rust and JavaScript, separately.

cs.PL↗

arXiv2023

FreeAL: Towards Human-Free Active Learning in the Era of Large Language Models

Collecting high-quality labeled data for model training is notoriously time-consuming and labor-intensive for various NLP tasks. While copious solutions, such as active learning for small language models (SLMs) and prevalent in-context learning in the era of large language models (LLMs), have been proposed and alleviate the labeling burden to some extent, their performances are still subject to human intervention. It is still underexplored how to reduce the annotation cost in the LLMs era. To bridge this, we revolutionize traditional active learning and propose an innovative collaborative learning framework FreeAL to interactively distill and filter the task-specific knowledge from LLMs. During collaborative training, an LLM serves as an active annotator inculcating its coarse-grained knowledge, while a downstream SLM is incurred as a student to filter out high-quality in-context samples to feedback LLM for the subsequent label refinery. Extensive experiments on eight benchmark datasets demonstrate that FreeAL largely enhances the zero-shot performances for both SLM and LLM without any human supervision. The code is available at https://github.com/Justherozen/FreeAL .

cs.CL↗

arXiv2023

ProMix: Combating Label Noise via Maximizing Clean Sample Utility

Learning with Noisy Labels (LNL) has become an appealing topic, as imperfectly annotated data are relatively cheaper to obtain. Recent state-of-the-art approaches employ specific selection mechanisms to separate clean and noisy samples and then apply Semi-Supervised Learning (SSL) techniques for improved performance. However, the selection step mostly provides a medium-sized and decent-enough clean subset, which overlooks a rich set of clean samples. To fulfill this, we propose a novel LNL framework ProMix that attempts to maximize the utility of clean samples for boosted performance. Key to our method, we propose a matched high confidence selection technique that selects those examples with high confidence scores and matched predictions with given labels to dynamically expand a base clean sample set. To overcome the potential side effect of excessive clean set selection procedure, we further devise a novel SSL framework that is able to train balanced and unbiased classifiers on the separated clean and noisy samples. Extensive experiments demonstrate that ProMix significantly advances the current state-of-the-art results on multiple benchmarks with different types and levels of noise. It achieves an average improvement of 2.48\% on the CIFAR-N dataset. The code is available at https://github.com/Justherozen/ProMix

cs.LG↗

arXiv2023

TelecomTM: A Fine-Grained and Ubiquitous Traffic Monitoring System Using Pre-Existing Telecommunication Fiber-Optic Cables as Sensors

We introduce the TelecomTM system that uses pre-existing telecommunication fiber-optic cables as virtual strain sensors to sense vehicle-induced ground vibrations for fine-grained and ubiquitous traffic monitoring and characterization. Here we call it a virtual sensor because it is a software-based representation of a physical sensor. Due to the extensively installed telecommunication fiber-optic cables at the roadside, our system using redundant dark fibers enables to monitor traffic at low cost with low maintenance. Many existing traffic monitoring approaches use cameras, piezoelectric sensors, and smartphones, but they are limited due to privacy concerns and/or deployment requirements. Previous studies attempted to use telecommunication cables for traffic monitoring, but they were only exploratory and limited to simple tasks at a coarse granularity, e.g., vehicle detection, due to their hardware constraints and real-world challenges. In particular, those challenges are 1) unknown and heterogeneous properties of virtual sensors and 2) large and complex noise conditions. To this end, our TelecomTM system first characterizes the geographic location and analyzes the signal pattern of each virtual sensor through driving tests. We then develop a spatial-domain Bayesian filtering and smoothing algorithm to detect, track, and characterize each vehicle. Our approach uses the spatial dependency of multiple virtual sensors and Newton's laws of motion to combine the distributed sensor data to reduce uncertainties in vehicle detection and tracking. In our real-world evaluation on a two-way traffic road with 1120 virtual sensors, TelecomTM achieved 90.18% vehicle detection accuracy, 27$\times$ and 5$\times$ error reduction for vehicle position and speed tracking compared to a baseline method, and $\pm$3.92% and $\pm$11.98% percent error for vehicle wheelbase and weight estimation, respectively.

eess.SY↗

arXiv2022

PigV$^2$: Monitoring Pig Vital Signs through Ground Vibrations Induced by Heartbeat and Respiration

Pig vital sign monitoring (e.g., estimating the heart rate (HR) and respiratory rate (RR)) is essential to understand the stress level of the sow and detect the onset of parturition. It helps to maximize peri-natal survival and improve animal well-being in swine production. The existing approach mainly relies on manual measurement, which is labor-intensive and only provides a few points of information. Other sensing modalities such as wearables and cameras are developed to enable more continuous measurement, but are still limited due to animal discomfort, data transfer, and storage challenges. In this paper, we introduce PigV$^2$, the first system to monitor pig heart rate and respiratory rate through ground vibrations. Our approach leverages the insight that both heartbeat and respiration generate ground vibrations when the sow is lying on the floor. We infer vital information by sensing and analyzing these vibrations. The main challenge in developing PigV$^2$ is the overlap of vital- and non-vital-related information in the vibration signals, including pig movements, pig postures, pig-to-sensor distances, and so on. To address this issue, we first characterize their effects, extract their current status, and then reduce their impact by adaptively interpolating vital rates over multiple sensors. PigV$^2$ is evaluated through a real-world deployment with 30 pigs. It has 3.4% and 8.3% average errors in monitoring the HR and RR of the sows, respectively.

eess.SP↗

arXiv2022

GaitVibe+: Enhancing Structural Vibration-based Footstep Localization Using Temporary Cameras for In-home Gait Analysis

In-home gait analysis is important for providing early diagnosis and adaptive treatments for individuals with gait disorders. Existing systems include wearables and pressure mats, but they have limited scalability. Recent studies have developed vision-based systems to enable scalable, accurate in-home gait analysis, but it faces privacy concerns due to the exposure of people's appearances. Our prior work developed footstep-induced structural vibration sensing for gait monitoring, which is device-free, wide-ranged, and perceived as more privacy-friendly. Although it has succeeded in temporal gait event extraction, it shows limited performance for spatial gait parameter estimation due to imprecise footstep localization. In particular, the localization error mainly comes from the estimation error of the wave arrival time at the vibration sensors and its error propagation to wave velocity estimations. Therefore, we present GaitVibe+, a vibration-based footstep localization method fused with temporarily installed cameras for in-home gait analysis. Our method has two stages: fusion and operating. In the fusion stage, both cameras and vibration sensors are installed to record only a few trials of the subject's footstep data, through which we characterize the uncertainty in wave arrival time and model the wave velocity profiles for the given structure. In the operating stage, we remove the camera to preserve privacy at home. The footstep localization is conducted by estimating the time difference of arrival (TDoA) over multiple vibration sensors, whose accuracy is improved through the reduced uncertainty and velocity modeling during the fusion stage. We evaluate GaitVibe+ through a real-world experiment with 50 walking trials. With only 3 trials of multi-modal fusion, our approach has an average localization error of 0.22 meters, which reduces the spatial gait parameter error from 111% to 27%.

eess.SP↗