Deep Learning using Convolutional LSTM estimates Biological Age from Physical Activity

Rahman, Syed Ashiqur; Adjeroh, Donald A.

doi:10.1038/s41598-019-46850-0

Cited by 82 publications

(69 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Deep neural networks are natural tools for learning non-trivial and highly non-linear representations of the input data. Convolutional and recurrent networks have been used for the analysis of intraday physical activity data streams from wearable devices and predictive modeling of health outcomes [ 23 ] including biological age [ 17 , 24 ]. Often such models demonstrate a moderate improvement in accuracy at a price of a decreased transferability across datasets with different baseline feature levels.…”

Section: Discussionmentioning

confidence: 99%

Deep longitudinal phenotyping of wearable sensor data reveals independent markers of longevity, stress, and resilience

Pyrkov

Sokolov

Федичев

2021

Aging

View full text Add to dashboard Cite

Biological age acceleration (BAA) models based on blood tests or DNA methylation emerge as a de facto standard for quantitative characterizations of the aging process. We demonstrate that deep neural networks trained to predict morbidity risk from wearable sensor data can provide a high-quality and cheap alternative for BAA determination. The GeroSense BAA model was trained and validated using steps per minute recordings from 103,830 one-week long and 2,599 of up to 2 years-long longitudinal samples and exhibited a superior association with life-expectancy over the average number of steps per day in, e.g., groups stratified by professional occupations. The association between the BAA and effects of lifestyles, the prevalence of future incidence of diseases was comparable to that of BAA from models based on blood test results. Wearable sensors let sampling of BAA fluctuations at time scales corresponding to days and weeks and revealed the divergence of organism state recovery time (resilience) as a function of chronological age. The number of individuals suffering from the lack of resilience increased exponentially with age at a rate compatible with Gompertz mortality law. We speculate that due to the stochastic character of BAA fluctuations, its mean and auto-correlation properties together comprise the minimum set of biomarkers of aging in humans.

show abstract

Section: Discussionmentioning

confidence: 99%

Deep longitudinal phenotyping of wearable sensor data reveals independent markers of longevity, stress, and resilience

Pyrkov

Sokolov

Федичев

2021

Aging

View full text Add to dashboard Cite

show abstract

“…Simultaneously, it can be used to screen out hub genes related to cancer tissue proteins (Liu et al, 2009). The expression level of hub genes can be analyzed by deep learning (Khan et al, 2001;Rahman and Adjeroh, 2019;Zeng et al, 2019;Mallik et al, 2020). This analysis can achieve good results at the genetic level.…”

Section: Introductionmentioning

confidence: 99%

Breast Cancer Case Identification Based on Deep Learning and Bioinformatics Analysis

Jia

Chen

et al. 2021

Front. Genet.

View full text Add to dashboard Cite

Mastering the molecular mechanism of breast cancer (BC) can provide an in-depth understanding of BC pathology. This study explored existing technologies for diagnosing BC, such as mammography, ultrasound, magnetic resonance imaging (MRI), computed tomography (CT), and positron emission tomography (PET) and summarized the disadvantages of the existing cancer diagnosis. The purpose of this article is to use gene expression profiles of The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) to classify BC samples and normal samples. The method proposed in this article triumphs over some of the shortcomings of traditional diagnostic methods and can conduct BC diagnosis more rapidly with high sensitivity and have no radiation. This study first selected the genes most relevant to cancer through weighted gene co-expression network analysis (WGCNA) and differential expression analysis (DEA). Then it used the protein–protein interaction (PPI) network to screen 23 hub genes. Finally, it used the support vector machine (SVM), decision tree (DT), Bayesian network (BN), artificial neural network (ANN), convolutional neural network CNN-LeNet and CNN-AlexNet to process the expression levels of 23 hub genes. For gene expression profiles, the ANN model has the best performance in the classification of cancer samples. The ten-time average accuracy is 97.36% (±0.34%), the F1 value is 0.8535 (±0.0260), the sensitivity is 98.32% (±0.32%), the specificity is 89.59% (±3.53%) and the AUC is 0.99. In summary, this method effectively classifies cancer samples and normal samples and provides reasonable new ideas for the early diagnosis of cancer in the future.

show abstract

“…2, and the difference between LSTM models and CLSTM models is precisely the presence of a convolutional block of layers before the LSTM block. This approach follows previous works [27], [28] that demonstrated good classification results; again, we remark that the goal of our article is to assess the scaling generalization aspect, which was not addressed at these works, instead of an architecture evaluation. Our training set is composed of different channel temporal sequences for each observed pixel, and our assumption is that convolutional layers help to identify curve patterns in the input temporal sequences, and LSTM layers help to identify the temporal dependence of these patterns for each crop type.…”

Section: B Crop Identification Using Clstm Networkmentioning

confidence: 94%

Generalization of Convolutional LSTM Models for Crop Area Estimation

Macedo

Mattos

Oliveira

2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

The population growth and consequent global rise in food demand require increasingly efficient agricultural solutions, in what is commonly called digital agriculture. Among promising initiatives, the use of remotely sensed data combined with machine learning algorithms enables handling faster agricultural operations with lower associated cost. One of the most important activities in digital agriculture is crop identification, which is fundamental for managing the inventory of a farm by producers and governmental authorities, and has been addressed by several prior works. In this article, we explore crop identification from a scalability perspective using the premise that data trained at a set of labeled geo-referenced regions in the agricultural pole at central western Brazil may be used for identifying crops at the entire municipality area. We propose to use convolutional long-short term memory networks for identifying crop types using a public labeled training set, and then apply the trained model for estimating crop area in a larger area involving the entire municipality unlabeled data. Our results were evaluated against governmental census data and report evidences that the tested crop identification network is able to successfully estimate crop area from much larger unlabeled data for different crop types.

show abstract

Deep Learning using Convolutional LSTM estimates Biological Age from Physical Activity

Cited by 82 publications

References 37 publications

Deep longitudinal phenotyping of wearable sensor data reveals independent markers of longevity, stress, and resilience

Deep longitudinal phenotyping of wearable sensor data reveals independent markers of longevity, stress, and resilience

Breast Cancer Case Identification Based on Deep Learning and Bioinformatics Analysis

Generalization of Convolutional LSTM Models for Crop Area Estimation

Contact Info

Product

Resources

About