Search for: All records

The recent success of deep neural networks in prediction tasks on wearable sensor data is evident. However, in more practical online learning scenarios, where new data arrive sequentially, neural networks suffer severely from the ``catastrophic forgetting`` problem. In real-world settings, given a pre-trained model on the old data, when we collect new data, it is practically infeasible to re-train the model on both old and new data because the computational costs will increase dramatically as more and more data arrive in time. However, if we fine-tune the model only with the new data because the new data might be different from the old data, the neural network parameters will change to fit the new data. As a result, the new parameters are no longer suitable for the old data. This phenomenon is known as catastrophic forgetting, and continual learning research aims to overcome this problem with minimal computational costs. While most of the continual learning research focuses on computer vision tasks, implications of catastrophic forgetting in wearable computing research and potential avenues to address this problem have remained unexplored. To address this knowledge gap, we study continual learning for activity recognition using wearable sensor data. We show that the catastrophic forgetting problem is a critical challenge for real-world deployment of machine learning models for wearables. Moreover, we show that the catastrophic forgetting problem can be alleviated by employing various training techniques. 

Stress detection and monitoring is an active area of research with important implications for an individual's personal, professional, and social health. Current approaches for stress classification use traditional machine learning algorithms trained on features computed from multiple sensor modalities. These methods are data and computation-intensive, rely on hand-crafted features, and lack reproducibility. These limitations impede the practical use of stress detection and classification systems in the real world. To overcome these shortcomings, we propose Stressalyzer, a novel stress classification and personalization framework from single-modality sensor data without feature computation and selection. Stressalyzer uses only Electrodermal activity (EDA) sensor data while providing competitive results compared to the state-of-the-art techniques that use multiple sensor modalities and are computationally expensive due to the calculation of large number of features. Using the dataset collected in a laboratory setting from $15$ subjects, our single-channel neural network-based model achieves a classification accuracy of 92.9% and an f1 score of 0.89 for binary stress classification. Our leave-one-subject-out analysis establishes the subjective nature of stress and shows that personalizing stress models using Stressalyzer significantly improves the model performance. Without model personalization, we found a performance decline in 40% of the subjects, suggesting the need for model personalization. 

Detecting stress from wearable sensor data enables those struggling with unhealthy stress coping mechanisms to better manage their stress. Previous studies have investigated how mechanisms for detecting stress from sensor data can be optimized, comparing alternative algorithms and approaches to find the best possible outcome. One strategy to make these mechanisms more accessible is to reduce the number of sensors that wearable devices must support. Reducing the number of sensors will enable wearable devices to be a smaller size, require less battery, and last longer, making use of these wearable devices more accessible. To progress towards this more convenient stress detection mechanism, we investigate how learning algorithms perform on singular modalities and compare the outcome with results from multiple modalities. We found that singular modalities performed comparably or better than combined modalities on two stress-detection datasets, suggesting that there is promise for detecting stress with fewer sensor requirements. From the four modalities we tested, acceleration, blood volume pulse, and electrodermal activity, we saw acceleration and electrodermal activity to stand out in a few cases, but all modalities showed potential. Our results are acquired from testing with random holdout and leave-one-subject-out validation, using several machine learning techniques. Our results can inspire work on optimizing stress detection with singular modalities to make the benefits of these detection mechanisms more convenient.