ALBERT

Beschreibung: Sensor-based human activity recognition has various applications in the arena of healthcare, elderly smart-home, sports, etc. There are numerous works in this field—to recognize various human activities from sensor data. However, those works are based on data patterns that are clean data and have almost no missing data, which is a genuine concern for real-life healthcare centers. Therefore, to address this problem, we explored the sensor-based activity recognition when some partial data were lost in a random pattern. In this paper, we propose a novel method to improve activity recognition while having missing data without any data recovery. For the missing data pattern, we considered data to be missing in a random pattern, which is a realistic missing pattern for sensor data collection. Initially, we created different percentages of random missing data only in the test data, while the training was performed on good quality data. In our proposed approach, we explicitly induce different percentages of missing data randomly in the raw sensor data to train the model with missing data. Learning with missing data reinforces the model to regulate missing data during the classification of various activities that have missing data in the test module. This approach demonstrates the plausibility of the machine learning model, as it can learn and predict from an identical domain. We exploited several time-series statistical features to extricate better features in order to comprehend various human activities. We explored both support vector machine and random forest as machine learning models for activity classification. We developed a synthetic dataset to empirically evaluate the performance and show that the method can effectively improve the recognition accuracy from 80.8% to 97.5%. Afterward, we tested our approach with activities from two challenging benchmark datasets: the human activity sensing consortium (HASC) dataset and single chest-mounted accelerometer dataset. We examined the method for different missing percentages, varied window sizes, and diverse window sliding widths. Our explorations demonstrated improved recognition performances even in the presence of missing data. The achieved results provide persuasive findings on sensor-based activity recognition in the presence of missing data.

Beschreibung: Wearable sensor-based systems and devices have been expanded in different application domains, especially in the healthcare arena. Automatic age and gender estimation has several important applications. Gait has been demonstrated as a profound motion cue for various applications. A gait-based age and gender estimation challenge was launched in the 12th IAPR International Conference on Biometrics (ICB), 2019. In this competition, 18 teams initially registered from 14 countries. The goal of this challenge was to find some smart approaches to deal with age and gender estimation from sensor-based gait data. For this purpose, we employed a large wearable sensor-based gait dataset, which has 745 subjects (357 females and 388 males), from 2 to 78 years old in the training dataset; and 58 subjects (19 females and 39 males) in the test dataset. It has several walking patterns. The gait data sequences were collected from three IMUZ sensors, which were placed on waist-belt or at the top of a backpack. There were 67 solutions from ten teams—for age and gender estimation. This paper extensively analyzes the methods and achieved-results from various approaches. Based on analysis, we found that deep learning-based solutions lead the competitions compared with conventional handcrafted methods. We found that the best result achieved 24.23% prediction error for gender estimation, and 5.39 mean absolute error for age estimation by employing angle embedded gait dynamic image and temporal convolution network.

Beschreibung: In this paper, we address Zero-shot learning for sensor activity recognition using word embeddings. The goal of Zero-shot learning is to estimate an unknown activity class (i.e., an activity that does not exist in a given training dataset) by learning to recognize components of activities expressed in semantic vectors. The existing zero-shot methods use mainly 2 kinds of representation as semantic vectors, attribute vector and embedding word vector. However, few zero-shot activity recognition methods based on embedding vector have been studied; especially for sensor-based activity recognition, no such studies exist, to the best of our knowledge. In this paper, we compare and thoroughly evaluate the Zero-shot method with different semantic vectors: (1) attribute vector, (2) embedding vector, and (3) expanded embedding vector and analyze their correlation to performance. Our results indicate that the performance of the three spaces is similar but the use of word embedding leads to a more efficient method, since this type of semantic vector can be generated automatically. Moreover, our suggested method achieved higher accuracy than attribute-vector methods, in cases when there exist similar information in both the given sensor data and in the semantic vector; the results of this study help select suitable classes and sensor data to build a training dataset.

Beschreibung: Integrating speech recondition technology into an electronic health record (EHR) has been studied in recent years. However, the full adoption of the system still faces challenges such as handling speech errors, transforming raw data into an understandable format and controlling the transition from one field to the next field with speech commands. To reduce errors, cost, and documentation time, we propose a dialogue system care record (DSCR) based on a smartphone for nursing documentation. We describe the effects of DSCR on (1) documentation speed, (2) document accuracy and (3) user satisfaction. We tested the application with 12 participants to examine the usability and feasibility of DSCR. The evaluation shows that DSCR can collect data efficiently by achieving 96% of documentation accuracy. Average documentation speed was increased by 15% (P = 0.012) compared to traditional electronic forms (e-forms). The participants’ average satisfaction rating was 4.8 using DSCR compared to 3.6 using e-forms on a scale of 1–5 (P = 0.032).

Beschreibung: Labeling activity data is a central part of the design and evaluation of human activity recognition systems. The performance of the systems greatly depends on the quantity and “quality” of annotations; therefore, it is inevitable to rely on users and to keep them motivated to provide activity labels. While mobile and embedded devices are increasingly using deep learning models to infer user context, we propose to exploit on-device deep learning inference using a long short-term memory (LSTM)-based method to alleviate the labeling effort and ground truth data collection in activity recognition systems using smartphone sensors. The novel idea behind this is that estimated activities are used as feedback for motivating users to collect accurate activity labels. To enable us to perform evaluations, we conduct the experiments with two conditional methods. We compare the proposed method showing estimated activities using on-device deep learning inference with the traditional method showing sentences without estimated activities through smartphone notifications. By evaluating with the dataset gathered, the results show our proposed method has improvements in both data quality (i.e., the performance of a classification model) and data quantity (i.e., the number of data collected) that reflect our method could improve activity data collection, which can enhance human activity recognition systems. We discuss the results, limitations, challenges, and implications for on-device deep learning inference that support activity data collection. Also, we publish the preliminary dataset collected to the research community for activity recognition.

Beschreibung: We detail the solution to the UCAmI Cup Challenge to recognizing on going activities at home from sensor measurements. We use binary sensors and proximity sensor measurements for the recognition. We use an hybrid strategy, combining a probabilistic model and a definition-based model. The former consists of a Hidden Markov Model using the result of a neural network as emission probabilities. It is trained with the labelled data provided by the Cup. The latter approach takes advantage of the descriptions provided for each of the activities which are expressed in logical statements based on the sensors states. We then combine the results with a weighted average. We compare the performance of each individual strategy and of the combined strategy.

Beschreibung: One of the biggest challenges of activity data collection is the need to rely on users and keep them engaged to continually provide labels. Recent breakthroughs in mobile platforms have proven effective in bringing deep neural networks powered intelligence into mobile devices. This study proposes a novel on-device personalization for data labeling for an activity recognition system using mobile sensing. The key idea behind this system is that estimated activities personalized for a specific individual user can be used as feedback to motivate user contribution and improve data labeling quality. First, we exploited fine-tuning using a Deep Recurrent Neural Network to address the lack of sufficient training data and minimize the need for training deep learning on mobile devices from scratch. Second, we utilized a model pruning technique to reduce the computation cost of on-device personalization without affecting the accuracy. Finally, we built a robust activity data labeling system by integrating the two techniques outlined above, allowing the mobile application to create a personalized experience for the user. To demonstrate the proposed model’s capability and feasibility, we developed and deployed the proposed system to realistic settings. For our experimental setup, we gathered more than 16,800 activity windows from 12 activity classes using smartphone sensors. We empirically evaluated the proposed quality by comparing it with a baseline using machine learning. Our results indicate that the proposed system effectively improved activity accuracy recognition for individual users and reduced cost and latency for inference for mobile devices. Based on our findings, we highlight critical and promising future research directions regarding the design of efficient activity data collection with on-device personalization.

Beschreibung: Parkinson’s disease (PD) patients experience varying symptoms related to their illness. Therefore, each patient needs a tailored treatment program from their doctors. One approach is the use of anti-PD medicines. However, a “wearing-off” phenomenon occurs when these medicines lose their effect. As a result, patients start to experience the symptoms again until their next medicine intake. In the long term, the duration of “wearing-off” begins to shorten. Thus, patients and doctors have to work together to manage PD symptoms effectively. This study aims to develop a prediction model that can determine the “wearing-off” of anti-PD medicine. We used fitness tracker data and self-reported symptoms from a smartphone application in a real-world environment. Two participants wore the fitness tracker for a month while reporting any symptoms using the Wearing-Off Questionnaire (WoQ-9) on a smartphone application. Then, we processed and combined the datasets for each participant’s models. Our analysis produced prediction models for each participant. The average balanced accuracy with the best hyperparameters was at 70.0–71.7% for participant 1 and 76.1–76.9% for participant 2, suggesting that our approach would be helpful to manage the “wearing-off” of anti-PD medicine, motor fluctuations of PD patients, and customized treatment for PD patients.