The problem of finding adequate population models in ecology is important for understanding essential aspects of their dynamic nature. Since analyzing and accurately predicting the intelligent adaptation of multiple species is difficult due to their complex interactions, the study of population dynamics still remains a challenging task in computational biology. In this paper, we use a modern deep reinforcement learning (RL) approach to explore a new avenue for understanding predator-prey ecosystems. Recently, reinforcement learning methods have achieved impressive results in areas, such as games and robotics. RL agents generally focus on building strategies for taking actions in an environment in order to maximize their expected returns. Here we frame the co-evolution of predators and preys in an ecosystem as allowing agents to learn and evolve toward better ones in a manner appropriate for multi-agent reinforcement learning. Recent significant advancements in reinforcement learning allow for new perspectives on these types of ecological issues. Our simulation results show that throughout the scenarios with RL agents, predators can achieve a reasonable level of sustainability, along with their preys.
By incorporating a growing number of sensors and adopting machine learning technologies, wearable devices have recently become a prominent health care application domain. Among the related research topics in this field, one of the most important issues is detecting falls while walking. Since such falls may lead to serious injuries, automatically and promptly detecting them during daily use of smartphones and/or smart watches is a particular need. In this paper, we investigate the use of Gaussian process (GP) methods for characterizing dynamic walking patterns and detecting falls while walking with built-in wearable sensors in smartphones and/or smartwatches. For the task of characterizing dynamic walking patterns in a low-dimensional latent feature space, we propose a novel approach called auto-encoded Gaussian process dynamical model, in which we combine a GP-based state space modeling method with a nonlinear dimensionality reduction method in a unique manner. The Gaussian process methods are fit for this task because one of the most import strengths of the Gaussian process methods is its capability of handling uncertainty in the model parameters. Also for detecting falls while walking, we propose to recycle the latent samples generated in training the auto-encoded Gaussian process dynamical model for GP-based novelty detection, which can lead to an efficient and seamless solution to the detection task. Experimental results show that the combined use of these GP-based methods can yield promising results for characterizing dynamic walking patterns and detecting falls while walking with the wearable sensors.
31 March 2016 in International Journal of Fuzzy Logic and Intelligent Systems
Jooyoung Park, Seongman Heo, Taehwan Kim, Jeongho Park, Jaein Kim, and Kyungwook Park. IJFIS 2016;16:44-51. http://dx.doi.org/10.5391/IJFIS.2016.16.1.44
Background: The two most commonly used methods to identify frailty are the frailty phenotype and the frailty index. However, both methods have limitations in clinical application. In addition, methods for measuring frailty have not yet been standardized. Objective: We aimed to develop and validate a classification model for predicting frailty status using vocal biomarkers in community-dwelling older adults, based on voice recordings obtained from the picture description task (PDT). Methods: We recruited 127 participants aged 50 years and older and collected clinical information through a short form of the Comprehensive Geriatric Assessment scale. Voice recordings were collected with a tablet device during the Korean version of the PDT, and we preprocessed audio data to remove background noise before feature extraction. Three artificial intelligence (AI) models were developed for identifying frailty status: SpeechAI (using speech data only), DemoAI (using demographic data only), and DemoSpeechAI (combining both data types). Results: Our models were trained and evaluated on the basis of 5-fold cross-validation for 127 participants and compared. The SpeechAI model, using deep learning–based acoustic features, outperformed in terms of accuracy and area under the receiver operating characteristic curve (AUC), 80.4% (95% CI 76.89%‐83.91%) and 0.89 (95% CI 0.86‐0.92), respectively, while the model using only demographics showed an accuracy of 67.96% (95% CI 67.63%‐68.29%) and an AUC of 0.74 (95% CI 0.73‐0.75). The SpeechAI model outperformed the model using only demographics significantly in AUC (t4=8.705 [2-sided]; P<.001). The DemoSpeechAI model, which combined demographics with deep learning–based acoustic features, showed superior performance (accuracy 85.6%, 95% CI 80.03%‐91.17% and AUC 0.93, 95% CI 0.89‐0.97), but there was no significant difference in AUC between the SpeechAI and DemoSpeechAI models (t4=1.057 [2-sided]; P=.35). Compared with models using traditional acoustic features from the openSMILE toolkit, the SpeechAI model demonstrated superior performance (AUC 0.89) over traditional methods (logistic regression: AUC 0.62; decision tree: AUC 0.57; random forest: AUC 0.66). Conclusions: Our findings demonstrate that vocal biomarkers derived from deep learning–based acoustic features can be effectively used to predict frailty status in community-dwelling older adults. The SpeechAI model showed promising accuracy and AUC, outperforming models based solely on demographic data or traditional acoustic features. Furthermore, while the combined DemoSpeechAI model showed slightly improved performance over the SpeechAI model, the difference was not statistically significant. These results suggest that speech-based AI models offer a noninvasive, scalable method for frailty detection, potentially streamlining assessments in clinical and community settings.
The global adoption of smartphone technology affords many conveniences, and not surprisingly, healthcare applications using wearable sensors like smartphones have received much attention. Among the various potential applications and research related to healthcare, recent studies have been conducted on recognizing human activities and characterizing human motions, often with wearable sensors, and with sensor signals that generally operate in the form of time series. In most studies, these sensor signals are used after pre-processing, e.g., by converting them into an image format rather than directly using the sensor signals themselves. Several methods have been used for converting time series data to image formats, such as spectrograms, raw plots, and recurrence plots. In this paper, we deal with the health care task of predicting human motion signals obtained from sensors attached to persons. We convert the motion signals into image formats with the recurrence plot method, and use it as an input into a deep learning model. For predicting subsequent motion signals, we utilize a recently introduced deep learning model combining neural networks and the Fourier transform, the Fourier neural operator. The model can be viewed as a Fourier-transform-based extension of a convolution neural network, and in these experiments, we compare the results of the model to the convolution neural network (CNN) model. The results of the proposed method in this paper show better performance than the results of the CNN model and, furthermore, we confirm that it can be utilized for detecting potential accidental falls more quickly via predicted motion signals.
With many conveniences afforded by advances in smartphone technology, developing advanced data analysis methods for health-related information from smartphone users has become a fast-growing research topic in the healthcare field. Along these lines, this paper addresses smartphone sensor-based characterization of human motions with neural stochastic differential equations (NSDEs) and a Transformer model. NSDEs and modeling via Transformer networks are two of the most prominent deep learning-based modeling approaches, with significant performance yields in many applications. For the problem of modeling dynamical features, stochastic differential equations and deep neural networks are frequently used paradigms in science and engineering, respectively. Combining these two paradigms in one unified framework has drawn significant interest in the deep learning community, and NSDEs are among the leading technologies for combining these efforts. The use of attention has also become a widely adopted strategy in many deep learning applications, and a Transformer is a deep learning model that uses the mechanism of self-attention. This concept of a self-attention based Transformer was originally introduced for tasks of natural language processing (NLP), and due to its excellent performance and versatility, the scope of its applications is rapidly expanding. By utilizing the techniques of neural stochastic differential equations and a Transformer model along with data obtained from smartphone sensors, we present a deep learning method capable of efficiently characterizing human motions. For characterizing human motions, we encode the high-dimensional sequential data from smartphone sensors into latent variables in a low-dimensional latent space. The concept of the latent variable is particularly useful because it can not only carry condensed information concerning motion data, but also learn their low-dimensional representations. More precisely, we use neural stochastic differential equations for modeling transitions of human motion in a latent space, and rely on a Generative Pre-trained Transformer 2 (GPT2)-based Transformer model for approximating the intractable posterior of conditional latent variables. Our experiments show that the proposed method can yield promising results for the problem of characterizing human motion patterns and some related tasks including user identification.