Edison Thomaz

Rebecca Adaimi, Edison Thomaz

Continual learning, also known as lifelong learning, is an emerging research topic that has been attracting increasing interest in the field of machine learning. With human activity recognition (HAR) playing a key role in enabling numerous real-world applications, an essential step towards the long-term deployment of such recognition systems is to extend the activity model to dynamically adapt to changes in people's everyday behavior. Current research in continual learning applied to HAR domain is still under-explored with researchers exploring existing methods developed for computer vision in HAR. Moreover, analysis has so far focused on task-incremental or class-incremental learning paradigms where task boundaries are known. This impedes the applicability of such methods for real-world systems since data is presented in a randomly streaming fashion. To push this field forward, we build on recent advances in the area of continual machine learning and design a lifelong adaptive learning framework using Prototypical Networks, LAPNet-HAR, that processes sensor-based data streams in a task-free data-incremental fashion and mitigates catastrophic forgetting using experience replay and continual prototype adaptation. Online learning is further facilitated using contrastive loss to enforce inter-class separation. LAPNet-HAR is evaluated on 5 publicly available activity datasets in terms of the framework's ability to acquire new information while preserving previous knowledge. Our extensive empirical results demonstrate the effectiveness of LAPNet-HAR in task-free continual learning and uncover useful insights for future challenges.

Xuewen Yao, Miriam Mikhelson, S. Craig Watkins et al.

In collaboration with Postpartum Support International (PSI), a non-profit organization dedicated to supporting caregivers with postpartum mood and anxiety disorders, we developed three chatbots to provide context-specific empathetic support to postpartum caregivers, leveraging both rule-based and generative models. We present and evaluate the performance of our chatbots using both machine-based metrics and human-based questionnaires. Overall, our rule-based model achieves the best performance, with outputs that are close to ground truth reference and contain the highest levels of empathy. Human users prefer the rule-based chatbot over the generative chatbot for its context-specific and human-like replies. Our generative chatbot also produced empathetic responses and was described by human users as engaging. However, limitations in the training dataset often result in confusing or nonsensical responses. We conclude by discussing practical benefits of rule-based vs. generative models for supporting individuals with mental health challenges. In light of the recent surge of ChatGPT and BARD, we also discuss the possibilities and pitfalls of large language models for digital mental healthcare.

Abduallah Mohamed, Fernando Lejarza, Stephanie Cahail et al.

The problem of human activity recognition from mobile sensor data applies to multiple domains, such as health monitoring, personal fitness, daily life logging, and senior care. A critical challenge for training human activity recognition models is data quality. Acquiring balanced datasets containing accurate activity labels requires humans to correctly annotate and potentially interfere with the subjects' normal activities in real-time. Despite the likelihood of incorrect annotation or lack thereof, there is often an inherent chronology to human behavior. For example, we take a shower after we exercise. This implicit chronology can be used to learn unknown labels and classify future activities. In this work, we propose HAR-GCCN, a deep graph CNN model that leverages the correlation between chronologically adjacent sensor measurements to predict the correct labels for unclassified activities that have at least one activity label. We propose a new training strategy enforcing that the model predicts the missing activity labels by leveraging the known ones. HAR-GCCN shows superior performance relative to previously used baseline methods, improving classification accuracy by about 25% and up to 68% on different datasets. Code is available at \url{https://github.com/abduallahmohamed/HAR-GCNN}.

Xuewen Yao, Miriam Mikhelson, Megan Micheletti et al.

Digital platforms, including online forums and helplines, have emerged as avenues of support for caregivers suffering from postpartum mental health distress. Understanding support seekers' experiences as shared on these platforms could provide crucial insight into caregivers' needs during this vulnerable time. In the current work, we provide a descriptive analysis of the concerns, psychological states, and motivations shared by healthy and distressed postpartum support seekers on two digital platforms, a one-on-one digital helpline and a publicly available online forum. Using a combination of human annotations, dictionary models and unsupervised techniques, we find stark differences between the experiences of distressed and healthy mothers. Distressed mothers described interpersonal problems and a lack of support, with 8.60% - 14.56% reporting severe symptoms including suicidal ideation. In contrast, the majority of healthy mothers described childcare issues, such as questions about breastfeeding or sleeping, and reported no severe mental health concerns. Across the two digital platforms, we found that distressed mothers shared similar content. However, the patterns of speech and affect shared by distressed mothers differed between the helpline vs. the online forum, suggesting the design of these platforms may shape meaningful measures of their support-seeking experiences. Our results provide new insight into the experiences of caregivers suffering from postpartum mental health distress. We conclude by discussing methodological considerations for understanding content shared by support seekers and design considerations for the next generation of support tools for postpartum parents.

Ebrahim Farahmand, Reza Rahimi Azghan, Nooshin Taheri Chatrudi et al.

Diabetes is a chronic metabolic disorder characterized by persistently high blood glucose levels (BGLs), leading to severe complications such as cardiovascular disease, neuropathy, and retinopathy. Predicting BGLs enables patients to maintain glucose levels within a safe range and allows caregivers to take proactive measures through lifestyle modifications. Continuous Glucose Monitoring (CGM) systems provide real-time tracking, offering a valuable tool for monitoring BGLs. However, accurately forecasting BGLs remains challenging due to fluctuations due to physical activity, diet, and other factors. Recent deep learning models show promise in improving BGL prediction. Nonetheless, forecasting BGLs accurately from multimodal, irregularly sampled data over long prediction horizons remains a challenging research problem. In this paper, we propose AttenGluco, a multimodal Transformer-based framework for long-term blood glucose prediction. AttenGluco employs cross-attention to effectively integrate CGM and activity data, addressing challenges in fusing data with different sampling rates. Moreover, it employs multi-scale attention to capture long-term dependencies in temporal data, enhancing forecasting accuracy. To evaluate the performance of AttenGluco, we conduct forecasting experiments on the recently released AIREADI dataset, analyzing its predictive accuracy across different subject cohorts including healthy individuals, people with prediabetes, and those with type 2 diabetes. Furthermore, we investigate its performance improvements and forgetting behavior as new cohorts are introduced. Our evaluations show that AttenGluco improves all error metrics, such as root mean square error (RMSE), mean absolute error (MAE), and correlation, compared to the multimodal LSTM model. AttenGluco outperforms this baseline model by about 10% and 15% in terms of RMSE and MAE, respectively.

Yufei Shen, Ji Hwan Park, Minchao Huang et al.

Early detection of cognitive impairment is critical for timely diagnosis and intervention, yet infrequent clinical assessments often lack the sensitivity and temporal resolution to capture subtle cognitive declines in older adults. Passive smartphone sensing has emerged as a promising approach for naturalistic and continuous cognitive monitoring. Building on this potential, we implemented a Long Short-Term Memory (LSTM) model to detect cognitive impairment from sequences of daily behavioral features, derived from multimodal sensing data collected in an ongoing one-year study of older adults. Our key contributions are two techniques to enhance model generalizability across participants: (1) routine-aware augmentation, which generates synthetic sequences by replacing each day with behaviorally similar alternatives, and (2) demographic personalization, which reweights training samples to emphasize those from individuals demographically similar to the test participant. Evaluated on 6-month data from 36 older adults, these techniques jointly improved the Area Under the Precision-Recall Curve (AUPRC) of the model trained on sensing and demographic features from 0.637 to 0.766, highlighting the potential of scalable monitoring of cognitive impairment in aging populations with passive sensing.

Reza Rahimi Azghan, Gautham Krishna Gudur, Mohit Malu et al.

The rise of deep learning has greatly advanced human behavior monitoring using wearable sensors, particularly human activity recognition (HAR). While deep models have been widely studied, most assume stationary data distributions - an assumption often violated in real-world scenarios. For example, sensor data from one subject may differ significantly from another, leading to distribution shifts. In continual learning, this shift is framed as a sequence of tasks, each corresponding to a new subject. Such settings suffer from catastrophic forgetting, where prior knowledge deteriorates as new tasks are learned. This challenge is compounded by the scarcity and inconsistency of labeled data in human studies. To address these issues, we propose CLAD-Net (Continual Learning with Attention and Distillation), a framework enabling wearable-sensor models to be updated continuously without sacrificing performance on past tasks. CLAD-Net integrates a self-supervised transformer, acting as long-term memory, with a supervised Convolutional Neural Network (CNN) trained via knowledge distillation for activity classification. The transformer captures global activity patterns through cross-attention across body-mounted sensors, learning generalizable representations without labels. Meanwhile, the CNN leverages knowledge distillation to retain prior knowledge during subject-wise fine-tuning. On PAMAP2, CLAD-Net achieves 91.36 percent final accuracy with only 8.78 percent forgetting, surpassing memory-based and regularization-based baselines such as Experience Replay and Elastic Weight Consolidation. In semi-supervised settings with only 10-20 percent labeled data, CLAD-Net still delivers strong performance, demonstrating robustness to label scarcity. Ablation studies further validate each module's contribution.

Ebrahim Farahmand, Reza Rahimi Azghan, Nooshin Taheri Chatrudi et al.

This paper proposes GluMind, a transformer-based multimodal framework designed for continual and long-term blood glucose forecasting. GluMind devises two attention mechanisms, including cross-attention and multi-scale attention, which operate in parallel and deliver accurate predictive performance. Cross-attention effectively integrates blood glucose data with other physiological and behavioral signals such as activity, stress, and heart rate, addressing challenges associated with varying sampling rates and their adverse impacts on robust prediction. Moreover, the multi-scale attention mechanism captures long-range temporal dependencies. To mitigate catastrophic forgetting, GluMind incorporates a knowledge retention technique into the transformer-based forecasting model. The knowledge retention module not only enhances the model's ability to retain prior knowledge but also boosts its overall forecasting performance. We evaluate GluMind on the recently released AIREADI dataset, which contains behavioral and physiological data collected from healthy people, individuals with prediabetes, and those with type 2 diabetes. We examine the performance stability and adaptability of GluMind in learning continuously as new patient cohorts are introduced. Experimental results show that GluMind consistently outperforms other state-of-the-art forecasting models, achieving approximately 15% and 9% improvements in root mean squared error (RMSE) and mean absolute error (MAE), respectively.

Alice Zhang, Callihan Bertley, Dawei Liang et al.

Social interactions play a crucial role in shaping human behavior, relationships, and societies. It encompasses various forms of communication, such as verbal conversation, non-verbal gestures, facial expressions, and body language. In this work, we develop a novel computational approach to detect a foundational aspect of human social interactions, in-person verbal conversations, by leveraging audio and inertial data captured with a commodity smartwatch in acoustically-challenging scenarios. To evaluate our approach, we conducted a lab study with 11 participants and a semi-naturalistic study with 24 participants. We analyzed machine learning and deep learning models with 3 different fusion methods, showing the advantages of fusing audio and inertial data to consider not only verbal cues but also non-verbal gestures in conversations. Furthermore, we perform a comprehensive set of evaluations across activities and sampling rates to demonstrate the benefits of multimodal sensing in specific contexts. Overall, our framework achieved 82.0$\pm$3.0% macro F1-score when detecting conversations in the lab and 77.2$\pm$1.8% in the semi-naturalistic setting.

Alice Zhang, Edison Thomaz, Lie Lu

Advancements in audio neural networks have established state-of-the-art results on downstream audio tasks. However, the black-box structure of these models makes it difficult to interpret the information encoded in their internal audio representations. In this work, we explore the semantic interpretability of audio embeddings extracted from these neural networks by leveraging CLAP, a contrastive learning model that brings audio and text into a shared embedding space. We implement a post-hoc method to transform CLAP embeddings into concept-based, sparse representations with semantic interpretability. Qualitative and quantitative evaluations show that the concept-based representations outperform or match the performance of original audio embeddings on downstream tasks while providing interpretability. Additionally, we demonstrate that fine-tuning the concept-based representations can further improve their performance on downstream tasks. Lastly, we publish three audio-specific vocabularies for concept-based interpretability of audio embeddings.

Haoxiang Yu, Jingyi An, Evan King et al.

Understanding the complexity of human activities solely through an individual's data can be challenging. However, in many situations, surrounding individuals are likely performing similar activities, while existing human activity recognition approaches focus almost exclusively on individual measurements and largely ignore the context of the activity. Consider two activities: attending a small group meeting and working at an office desk. From solely an individual's perspective, it can be difficult to differentiate between these activities as they may appear very similar, even though they are markedly different. Yet, by observing others nearby, it can be possible to distinguish between these activities. In this paper, we propose an approach to enhance the prediction accuracy of an individual's activities by incorporating insights from surrounding individuals. We have collected a real-world dataset from 20 participants with over 58 hours of data including activities such as attending lectures, having meetings, working in the office, and eating together. Compared to observing a single person in isolation, our proposed approach significantly improves accuracy. We regard this work as a first step in collaborative activity recognition, opening new possibilities for understanding human activity in group settings.

Dawei Liang, Yangyang Shi, Yun Wang et al.

Detection of common events and scenes from audio is useful for extracting and understanding human contexts in daily life. Prior studies have shown that leveraging knowledge from a relevant domain is beneficial for a target acoustic event detection (AED) process. Inspired by the observation that many human-centered acoustic events in daily life involve voice elements, this paper investigates the potential of transferring high-level voice representations extracted from a public speaker dataset to enrich an AED pipeline. Towards this end, we develop a dual-branch neural network architecture for the joint learning of voice and acoustic features during an AED process and conduct thorough empirical studies to examine the performance on the public AudioSet [1] with different types of inputs. Our main observations are that: 1) Joint learning of audio and voice inputs improves the AED performance (mean average precision) for both a CNN baseline (0.292 vs 0.134 mAP) and a TALNet [2] baseline (0.361 vs 0.351 mAP); 2) Augmenting the extra voice features is critical to maximize the model performance with dual inputs.

Sarnab Bhattacharya, Keum San Chun, Edison Thomaz

With the ubiquity of touchscreens, touch input modality has become a popular way of interaction. However, current touchscreen technology is limiting in its design as it restricts touch interactions to specially instrumented touch surfaces. Surface contaminants like water can also hinder proper interactions. In this paper, we propose the use of magnetic field sensing to enable finger tracking on a surface with minimal instrumentation. Our system, MagSurface, turns everyday surfaces into a touch medium, thus allowing more flexibility in the types of touch surfaces. The evaluation of our system consists of quantifying the accuracy of the system in locating an object on 2D flat surfaces. We test our system on three different surface materials to validate its usage scenarios. A qualitative user experience study was also conducted to get feedback on the ease of use and comfort of the system. Localization error as low as a few millimeters was achieved

Congyu Wu, Amanda N. Barczyk, R. Cameron Craddock et al.

Loneliness is a widely affecting mental health symptom and can be mediated by and co-vary with patterns of social exposure. Using momentary survey and smartphone sensing data collected from 129 Android-using college student participants over three weeks, we (1) investigate and uncover the relations between momentary loneliness experience and companionship type and (2) propose and validate novel geosocial features of smartphone-based Bluetooth and GPS data for predicting loneliness and companionship type in real time. We base our features on intuitions characterizing the quantity and spatiotemporal predictability of an individual's Bluetooth encounters and GPS location clusters to capture personal significance of social exposure scenarios conditional on their temporal distribution and geographic patterns. We examine our features' statistical correlation with momentary loneliness through regression analyses and evaluate their predictive power using a sliding window prediction procedure. Our features achieved significant performance improvement compared to baseline for predicting both momentary loneliness and companionship type, with the effect stronger for the loneliness prediction task. As such we recommend incorporation and further evaluation of our geosocial features proposed in this study in future mental health sensing and context-aware computing applications.

Congyu Wu, Hagen Fritz, Zoltan Nagy et al.

As mobile technologies become ever more sensor-rich, portable, and ubiquitous, data captured by smart devices are lending rich insights into users' daily lives with unprecedented comprehensiveness, unobtrusiveness, and ecological validity. A number of human-subject studies have been conducted in the past decade to examine the use of mobile sensing to uncover individual behavioral patterns and health outcomes. While understanding health and behavior is the focus for most of these studies, we find that minimal attention has been placed on measuring personal environments, especially together with other human-centric data modalities. Moreover, the participant cohort size in most existing studies falls well below a few hundred, leaving questions open about the reliability of findings on the relations between mobile sensing signals and human outcomes. To address these limitations, we developed a home environment sensor kit for continuous indoor air quality tracking and deployed it in conjunction with established mobile sensing and experience sampling techniques in a cohort study of up to 1584 student participants per data type for 3 weeks at a major research university in the United States. In this paper, we begin by proposing a conceptual framework that systematically organizes human-centric data modalities by their temporal coverage and spatial freedom. Then we report our study design and procedure, technologies and methods deployed, descriptive statistics of the collected data, and results from our extensive exploratory analyses. Our novel data, conceptual development, and analytical findings provide important guidance for data collection and hypothesis generation in future human-centric sensing studies.

Xuewen Yao, Megan Micheletti, Mckensey Johnson et al.

Most existing cry detection models have been tested with data collected in controlled settings. Thus, the extent to which they generalize to noisy and lived environments is unclear. In this paper, we evaluate several established machine learning approaches including a model leveraging both deep spectrum and acoustic features. This model was able to recognize crying events with F1 score 0.613 (Precision: 0.672, Recall: 0.552), showing improved external validity over existing methods at cry detection in everyday real-world settings. As part of our evaluation, we collect and annotate a novel dataset of infant crying compiled from over 780 hours of labeled real-world audio data, captured via recorders worn by infants in their homes, which we make publicly available. Our findings confirm that a cry detection model trained on in-lab data underperforms when presented with real-world data (in-lab test F1: 0.656, real-world test F1: 0.236), highlighting the value of our new dataset and model.

Priyanka Khante, Mai Lee Chang, Domingo Martinez et al.

Acoustic environments vary dramatically within the home setting. They can be a source of comfort and tranquility or chaos that can lead to less optimal cognitive development in children. Research to date has only subjectively measured household chaos. In this work, we use three unsupervised machine learning techniques to quantify household chaos in infants' homes. These unsupervised techniques include hierarchical clustering using K-Means, clustering using self-organizing map (SOM) and deep learning. We evaluated these techniques using data from 9 participants which is a total of 197 hours. Results show that these techniques are promising to quantify household chaos.

Dawei Liang, Edison Thomaz

Over the years, activity sensing and recognition has been shown to play a key enabling role in a wide range of applications, from sustainability and human-computer interaction to health care. While many recognition tasks have traditionally employed inertial sensors, acoustic-based methods offer the benefit of capturing rich contextual information, which can be useful when discriminating complex activities. Given the emergence of deep learning techniques and leveraging new, large-scaled multi-media datasets, this paper revisits the opportunity of training audio-based classifiers without the onerous and time-consuming task of annotating audio data. We propose a framework for audio-based activity recognition that makes use of millions of embedding features from public online video sound clips. Based on the combination of oversampling and deep learning approaches, our framework does not require further feature processing or outliers filtering as in prior work. We evaluated our approach in the context of Activities of Daily Living (ADL) by recognizing 15 everyday activities with 14 participants in their own homes, achieving 64.2% and 83.6% averaged within-subject accuracy in terms of top-1 and top-3 classification respectively. Individual class performance was also examined in the paper to further study the co-occurrence characteristics of the activities and the robustness of the framework.

Vinay Bettadapura, Edison Thomaz, Aman Parnami et al.

The pervasiveness of mobile cameras has resulted in a dramatic increase in food photos, which are pictures reflecting what people eat. In this paper, we study how taking pictures of what we eat in restaurants can be used for the purpose of automating food journaling. We propose to leverage the context of where the picture was taken, with additional information about the restaurant, available online, coupled with state-of-the-art computer vision techniques to recognize the food being consumed. To this end, we demonstrate image-based recognition of foods eaten in restaurants by training a classifier with images from restaurant's online menu databases. We evaluate the performance of our system in unconstrained, real-world settings with food images taken in 10 restaurants across 5 different types of food (American, Indian, Italian, Mexican and Thai).

Daniel Castro, Steven Hickson, Vinay Bettadapura et al.

We present a method to analyze images taken from a passive egocentric wearable camera along with the contextual information, such as time and day of week, to learn and predict everyday activities of an individual. We collected a dataset of 40,103 egocentric images over a 6 month period with 19 activity classes and demonstrate the benefit of state-of-the-art deep learning techniques for learning and predicting daily activities. Classification is conducted using a Convolutional Neural Network (CNN) with a classification method we introduce called a late fusion ensemble. This late fusion ensemble incorporates relevant contextual information and increases our classification accuracy. Our technique achieves an overall accuracy of 83.07% in predicting a person's activity across the 19 activity classes. We also demonstrate some promising results from two additional users by fine-tuning the classifier with one day of training data.