Review of deep learning: concepts, CNN architectures, challenges, applications, future directions

Journal Of Big Data, Journal Year: 2021, Volume and Issue: 8(1)

Published: March 31, 2021

In the last few years, deep learning (DL) computing paradigm has been deemed Gold Standard in machine (ML) community. Moreover, it gradually become most widely used computational approach field of ML, thus achieving outstanding results on several complex cognitive tasks, matching or even beating those provided by human performance. One benefits DL is ability to learn massive amounts data. The grown fast years and extensively successfully address a wide range traditional applications. More importantly, outperformed well-known ML techniques many domains, e.g., cybersecurity, natural language processing, bioinformatics, robotics control, medical information among others. Despite contributed works reviewing State-of-the-Art DL, all them only tackled one aspect which leads an overall lack knowledge about it. Therefore, this contribution, we propose using more holistic order provide suitable starting point from develop full understanding DL. Specifically, review attempts comprehensive survey important aspects including enhancements recently added field. particular, paper outlines importance presents types networks. It then convolutional neural networks (CNNs) utilized network type describes development CNNs architectures together with their main features, AlexNet closing High-Resolution (HR.Net). Finally, further present challenges suggested solutions help researchers understand existing research gaps. followed list major Computational tools FPGA, GPU, CPU are summarized along description influence ends evolution matrix, benchmark datasets, summary conclusion.

Language: Английский

---

Artificial intelligence in healthcare

Nature Biomedical Engineering, Journal Year: 2018, Volume and Issue: 2(10), P. 719 - 731

Published: Oct. 5, 2018

Language: Английский

---

CellProfiler 3.0: Next-generation image processing for biology

PLoS Biology, Journal Year: 2018, Volume and Issue: 16(7), P. e2005970 - e2005970

Published: July 3, 2018

CellProfiler has enabled the scientific research community to create flexible, modular image analysis pipelines since its release in 2005. Here, we describe 3.0, a new version of software supporting both whole-volume and plane-wise three-dimensional (3D) stacks, increasingly common biomedical research. CellProfiler's infrastructure is greatly improved, provide protocol for cloud-based, large-scale processing. New plugins enable running pretrained deep learning models on images. Designed by biologists, equips researchers with powerful computational tools via well-documented user interface, empowering biologists all fields quantitative, reproducible workflows.

Language: Английский

---

Albumentations: Fast and Flexible Image Augmentations

Information, Journal Year: 2020, Volume and Issue: 11(2), P. 125 - 125

Published: Feb. 24, 2020

Data augmentation is a commonly used technique for increasing both the size and diversity of labeled training sets by leveraging input transformations that preserve corresponding output labels. In computer vision, image augmentations have become common implicit regularization to combat overfitting in deep learning models are ubiquitously improve performance. While most frameworks implement basic transformations, list typically limited some variations flipping, rotating, scaling, cropping. Moreover, processing speed varies existing libraries. We present Albumentations, fast flexible open source library with many various transform operations available also an easy-to-use wrapper around other discuss design principles drove implementation Albumentations give overview key features distinct capabilities. Finally, we provide examples different vision tasks demonstrate faster than tools on operations.

Language: Английский

---

The rise of deep learning in drug discovery

Drug Discovery Today, Journal Year: 2018, Volume and Issue: 23(6), P. 1241 - 1250

Published: Jan. 31, 2018

Over the past decade, deep learning has achieved remarkable success in various artificial intelligence research areas. Evolved from previous on neural networks, this technology shown superior performance to other machine algorithms areas such as image and voice recognition, natural language processing, among others. The first wave of applications pharmaceutical emerged recent years, its utility gone beyond bioactivity predictions promise addressing diverse problems drug discovery. Examples will be discussed covering prediction, de novo molecular design, synthesis prediction biological analysis.

Language: Английский

---

Generating Focused Molecule Libraries for Drug Discovery with Recurrent Neural Networks

ACS Central Science, Journal Year: 2017, Volume and Issue: 4(1), P. 120 - 131

Published: Dec. 28, 2017

In

Language: Английский

---

A guide to machine learning for biologists

Nature Reviews Molecular Cell Biology, Journal Year: 2021, Volume and Issue: 23(1), P. 40 - 55

Published: Sept. 13, 2021

Language: Английский

---

Artificial intelligence in digital pathology — new tools for diagnosis and precision oncology

Nature Reviews Clinical Oncology, Journal Year: 2019, Volume and Issue: 16(11), P. 703 - 715

Published: Aug. 9, 2019

Language: Английский

---

A deep learning algorithm using CT images to screen for Corona virus disease (COVID-19)

European Radiology, Journal Year: 2021, Volume and Issue: 31(8), P. 6096 - 6104

Published: Feb. 24, 2021

Abstract Objective The outbreak of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-COV-2) has caused more than 26 million cases Corona virus disease (COVID-19) in the world so far. To control spread disease, screening large numbers suspected for appropriate quarantine and treatment are a priority. Pathogenic laboratory testing is typically gold standard, but it bears burden significant false negativity, adding to urgent need alternative diagnostic methods combat disease. Based on COVID-19 radiographic changes CT images, this study hypothesized that artificial intelligence might be able extract specific graphical features provide clinical diagnosis ahead pathogenic test, thus saving critical time control. Methods We collected 1065 images pathogen-confirmed along with those previously diagnosed typical viral pneumonia. modified inception transfer-learning model establish algorithm, followed by internal external validation. Results validation achieved total accuracy 89.5% specificity 0.88 sensitivity 0.87. dataset showed 79.3% 0.83 0.67. In addition, 54 first two nucleic acid test results were negative, 46 predicted as positive an 85.2%. Conclusion These demonstrate proof-of-principle using radiological timely accurate diagnosis. Key Points • evaluated performance deep learning algorithm screen during influenza season. As method, our relatively high image datasets. was used distinguish between other pneumonia, both which have quite similar radiologic characteristics.

Language: Английский

---

A deep learning algorithm using CT images to screen for Corona Virus Disease (COVID-19)

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2020, Volume and Issue: unknown

Published: Feb. 17, 2020

Abstract Background The outbreak of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-COV-2) has caused more than 2.5 million cases Corona Virus Disease (COVID-19) in the world so far, with that number continuing to grow. To control spread disease, screening large numbers suspected for appropriate quarantine and treatment is a priority. Pathogenic laboratory testing gold standard but time-consuming significant false negative results. Therefore, alternative diagnostic methods are urgently needed combat disease. Based on COVID-19 radiographical changes CT images, we hypothesized Artificial Intelligence’s deep learning might be able extract COVID-19’s specific graphical features provide clinical diagnosis ahead pathogenic test, thus saving critical time disease control. Methods Findings We collected 1,065 images pathogen-confirmed (325 images) along those previously diagnosed typical viral pneumonia (740 images). modified Inception transfer-learning model establish algorithm, followed by internal external validation. validation achieved total accuracy 89.5% specificity 0.88 sensitivity 0.87. dataset showed 79.3% 0.83 0.67. In addition, 54 first two nucleic acid test results were negative, 46 predicted as positive 85.2%. Conclusion These demonstrate proof-of-principle using artificial intelligence radiological timely accurate diagnosis. Author summary COVID-19, measures time. pneumonia. algorithm. Our study represents apply effectively COVID-19.

Language: Английский

---