Comparative analysis of deep learning methods of detection of diabetic retinopathy

The increasing amount of malware and cyberattacks on a host level increases the need for a reliable anomaly-based host IDS (HIDS) that would be able to deal with zero-day attacks and would ensure low false alarm rate (FAR), which is critical for the detection of such activity. Deep learning methods such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs) are considered to be highly suitable for solving data-driven security solutions. Therefore, it is necessary to perform the comparative analysis of such methods in order to evaluate their efficiency in attack classification as well as their ability to distinguish malicious and benign activity. In this article, we present the results achieved with the AWSCTD (attack-caused Windows OS system calls traces dataset), which can be considered as the most exhaustive set of host-level anomalies at the moment, including 112.56 million system calls from 12110 executable malware samples and 3145 benign software samples with 16.3 million system calls. The best results were obtained with CNNs with up to 90.0% accuracy for family classification and 95.0% accuracy for malicious/benign determination. RNNs demonstrated slightly inferior results. Furthermore, CNN tuning via an increase in the number of layers should make them practically applicable for host-level anomaly detection.

Download Full-text

Hybrid CNN-LSTM Model for Answer Identification

International Journal of Recent Technology and Engineering - 2 ◽

10.35940/ijrte.c4281.098319 ◽

2019 ◽

Vol 8 (3) ◽

pp. 1163-1166

Keyword(s):

Neural Network ◽

Feature Extraction ◽

Deep Learning ◽

Comparative Analysis ◽

Hybrid Model ◽

Deep Neural Network ◽

Extraction Methods ◽

Learning Methods ◽

Automatic Feature Extraction

User quest for information has led to development of Question Answer (QA) system to provide relevant answers to user questions. The QA task are different than normal NLP tasks as they heavily depend to semantics and context of given data. Retrieving and predicting answers to verity of questions require understanding of question, relevance with context and identifying and retrieving of suitable answers. Deep learning helps to produce impressive performance as it employs deep neural network with automatic feature extraction methods. The paper proposes a hybrid model to identify suitable answer for posed question. The proposes power exploits the power of CNN for extracting features and ability of LSTM for considering long term dependencies and semantic of context and question. Paper provides a comparative analysis on deep learning methods useful for predicting answer with the proposed method .The model is implemented on twenty tasks of babI dataset of Facebook .

Download Full-text

Application of Deep Learning Methods in Cybersecurity Tasks. Part 2

Voprosy kiberbezopasnosti ◽

10.21681/2311-3456-2020-04-11-21 ◽

2020 ◽

pp. 11-21

Author(s):

Diana Gaifulina ◽

◽

Igor Kotenko ◽

Keyword(s):

Deep Learning ◽

Comparative Analysis ◽

Subject Area ◽

Data Sets ◽

Learning Methods ◽

Systematic Analysis ◽

Network Traffic Analysis ◽

Current State ◽

The Subject

The purpose of the article: comparative analysis of methods for solving various cybersecurity problems based on the use of deep learning algorithms. Research method: Systematic analysis of modern methods of deep learning in various cybersecurity applications, including intrusion and malware detection, network traffic analysis, and some other tasks. The result obtained: classification scheme of the considered approaches to deep learning in cybersecurity, and their comparative characteristics by the used models, characteristics, and data sets. The analysis showed that many deeper architectures with a large number of neurons on each layer show better results. Recommendations are given for using deep learning methods in cybersecurity applications. The main contribution of the authors to the research of deep learning methods for cybersecurity tasks is the classification of the subject area; conducting a general and comparative analysis of existing approaches that reflect the current state of scientific research.

Download Full-text

Deep Learning Methods for the Assessment of Vascular Diameters for Diabetic Retinopathy Screening

10.1109/gcat52182.2021.9587870 ◽

2021 ◽

Author(s):

Yagna Sai Kalyan Rebba ◽

Shashank Rao Gujja ◽

Satyanarayana Teja Siripalli ◽

Mohammed Shoaib ◽

Lakshmi Kala Pampana ◽

...

Keyword(s):

Diabetic Retinopathy ◽

Deep Learning ◽

Learning Methods ◽

Diabetic Retinopathy Screening ◽

Retinopathy Screening

Download Full-text

Diabetic Retinopathy Detection Using Neural Network

International Journal of Innovative Technology and Exploring Engineering - Special Issue ◽

10.35940/ijitee.j1105.0881019 ◽

2019 ◽

Vol 8 (10) ◽

pp. 2936-2940

Keyword(s):

Neural Network ◽

Neural Networks ◽

Computer Vision ◽

Diabetic Retinopathy ◽

Artificial Neural Networks ◽

Deep Learning ◽

Blood Vessels ◽

Proliferative Diabetic Retinopathy ◽

Eye Disease ◽

Learning Methods

Now-a-days diabetics are affecting many people and it causes an eye disease called “diabetics retinopathy” but many are not aware of that, so it causes blindness. Diabetes aimed at protracted time harms the blood vessels of retina in addition to thereby affecting seeing ability of an individual in addition to leading to diabetic retinopathy. Diabetic retinopathy is classified hooked on twofold classes, non-proliferative diabetic retinopathy (NPDR) and proliferative diabetic retinopathy (PDR). Finding of diabetic retinopathy in fundus imaginary is done by computer vision and deep learning methods using artificial neural networks. The images of the diabetic retinopathy datasets are trained in neural networks. And based on the training datasets we can detect whether the person has (i)no diabetic retinopathy, (ii) mild non-proliferative diabetic retinopathy, (iii) severe non-proliferative diabetic retinopathy and (iv) proliferative diabetic retinopathy.

Download Full-text