Global marine phytoplankton dynamics analysis with machine learning and reanalyzed remote sensing.

Phytoplankton are the world's largest oxygen producers found in oceans, seas and large water bodies, which play crucial roles in the marine food chain. Unbalanced biogeochemical features like salinity, pH, minerals, etc., can retard their growth. With advancements in better hardware, the usage of Artificial Intelligence techniques is rapidly increasing for creating an intelligent decision-making system. Therefore, we attempt to overcome this gap by using supervised regressions on reanalysis data targeting global phytoplankton levels in global waters. The presented experiment proposes the applications of different supervised machine learning regression techniques such as random forest, extra trees, bagging and histogram-based gradient boosting regressor on reanalysis data obtained from the Copernicus Global Ocean Biogeochemistry Hindcast dataset. Results obtained from the experiment have predicted the phytoplankton levels with a coefficient of determination score (R2) of up to 0.96. After further validation with larger datasets, the model can be deployed in a production environment in an attempt to complement in-situ measurement efforts.

Secret learning for lung cancer diagnosis-a study with homomorphic encryption, texture analysis and deep learning.

Advanced lung cancer diagnoses from radiographic images include automated detection of lung cancer from CT-Scan images of the lungs. Deep learning is a popular method for decision making which can be used to classify cancerous and non-cancerous lungs from CT-Scan images. There are many experiments which show the uses of deep learning for performing such classifications but very few of them have preserved the privacy of users. Among existing methods, federated learning limits data sharing to a central server and differential privacy although increases anonymity the original data is still shared. Homomorphic encryption can resolve the limitations of both of these. Homomorphic encryption is a cryptographic technique that allows computations to be performed on encrypted data. In our experiment, we have proposed a series of textural information extraction with the implementation of homomorphic encryption of the CT-Scan images of normal, adenocarcinoma, large cell carcinoma and squamous cell carcinoma. We have further processed the encrypted data to make it classifiable and later we have classified it with deep learning. The results from the experiments have obtained a classification accuracy of 0.9347.

Machine Learning-Based Regression Framework to Predict Health Insurance Premiums.

Artificial intelligence (AI) and machine learning (ML) in healthcare are approaches to make people's lives easier by anticipating and diagnosing diseases more swiftly than most medical experts. There is a direct link between the insurer and the policyholder when the distance between an insurance business and the consumer is reduced to zero with the use of technology, especially digital health insurance. In comparison with traditional insurance, AI and machine learning have altered the way insurers create health insurance policies and helped consumers receive services faster. Insurance businesses use ML to provide clients with accurate, quick, and efficient health insurance coverage. This research trained and evaluated an artificial intelligence network-based regression-based model to predict health insurance premiums. The authors predicted the health insurance cost incurred by individuals on the basis of their features. On the basis of various parameters, such as age, gender, body mass index, number of children, smoking habits, and geolocation, an artificial neural network model was trained and evaluated. The experimental results displayed an accuracy of 92.72%, and the authors analyzed the model's performance using key performance metrics.

n-Player Stochastic Duel Game Model with Applied Deep Learning and Its Modern Implications.

This paper provides a conceptual foundation for stochastic duels and contains a further study of the game models based on the theory of stochastic duels. Some other combat assessment techniques are looked upon briefly; a modern outlook on the applications of the theory through video games is provided; and the possibility of usage of data generated by popular shooter-type video games is discussed. Impactful works to date are carefully chosen; a timeline of the developments in the theory of stochastic duels is provided; and a brief literature review for the same is conducted, enabling readers to have a broad outlook at the theory of stochastic duels. A new evaluation model is introduced in order to match realistic scenarios. Improvements are suggested and, additionally, a trust mechanism is introduced to identify the intent of a player in order to make the model a better fit for realistic modern problems. The concept of teaming of players is also considered in the proposed mode. A deep-learning model is developed and trained on data generated by video games to support the results of the proposed model. The proposed model is compared to previously published models in a brief comparison study. Contrary to the conventional stochastic duel game combat model, this new proposed model deals with pair-wise duels throughout the game duration. This model is explained in detail, and practical applications of it in the context of the real world are also discussed. The approach toward solving modern-day problems through the use of game theory is presented in this paper, and hence, this paper acts as a foundation for researchers looking forward to an innovation with game theory.

ABCanDroid: A Cloud Integrated Android App for Noninvasive Early Breast Cancer Detection Using Transfer Learning.

Many patients affected by breast cancer die every year because of improper diagnosis and treatment. In recent years, applications of deep learning algorithms in the field of breast cancer detection have proved to be quite efficient. However, the application of such techniques has a lot of scope for improvement. Major works have been done in this field, however it can be made more efficient by the use of transfer learning to get impressive results. In the proposed approach, Convolutional Neural Network (CNN) is complemented with Transfer Learning for increasing the efficiency and accuracy of early detection of breast cancer for better diagnosis. The thought process involved using a pre-trained model, which already had some weights assigned rather than building the complete model from scratch. This paper mainly focuses on ResNet101 based Transfer Learning Model paired with the ImageNet dataset. The proposed framework provided us with an accuracy of 99.58%. Extensive experiments and tuning of hyperparameters have been performed to acquire the best possible results in terms of classification. The proposed frameworks aims to be an efficient tool for all doctors and society as a whole and help the user in early detection of breast cancer.

Federated learning based Covid-19 detection.

The world is affected by COVID-19, an infectious disease caused by the SARS-CoV-2 virus. Tests are necessary for everyone as the number of COVID-19 affected individual's increases. So, the authors developed a basic sequential CNN model based on deep and federated learning that focuses on user data security while simultaneously enhancing test accuracy. The proposed model helps users detect COVID-19 in a few seconds by uploading a single chest X-ray image. A deep learning-aided architecture that can handle client and server sides efficiently has been proposed in this work. The front-end part has been developed using StreamLit, and the back-end uses a Flower framework. The proposed model has achieved a global accuracy of 99.59% after being trained for three federated communication rounds. The detailed analysis of this paper provides the robustness of this work. In addition, the Internet of Medical Things (IoMT) will improve the ease of access to the aforementioned health services. IoMT tools and services are rapidly changing healthcare operations for the better. Hopefully, it will continue to do so in this difficult time of the COVID-19 pandemic and will help to push the envelope of this work to a different extent.

BRISK: Dynamic Encryption Based Cipher for Long Term Security.

Several emerging areas like the Internet of Things, sensor networks, healthcare and distributed networks feature resource-constrained devices that share secure and privacy-preserving data to accomplish some goal. The majority of standard cryptographic algorithms do not fit with these constrained devices due to heavy cryptographic components. In this paper, a new block cipher, BRISK, is proposed with a block size of 32-bit. The cipher design is straightforward due to simple round operations, and these operations can be efficiently run in hardware and suitable for software. Another major concept used with this cipher is dynamism during encryption for each session; that is, instead of using the same encryption algorithm, participants use different ciphers for each session. Professor Lars R. Knudsen initially proposed dynamic encryption in 2015, where the sender picks a cipher from a large pool of ciphers to encrypt the data and send it along with the encrypted message. The receiver does not know about the encryption technique used before receiving the cipher along with the message. However, in the proposed algorithm, instead of choosing a new cipher, the process uses the same cipher for each session, but varies the cipher specifications from a given small pool, e.g., the number of rounds, cipher components, etc. Therefore, the dynamism concept is used here in a different way.

Internet of Things Based Blockchain for Temperature Monitoring and Counterfeit Pharmaceutical Prevention.

The top priority of today's healthcare system is delivering medicine directly from the manufacturer to end-user. The pharmaceutical supply chain involves some level of commingling of a collection of stakeholders such as distributors, manufacturers, wholesalers, and customers. The biggest challenge associated with this supply chain is temperature monitoring as well as counterfeit drug prevention. Many drugs and vaccines remain viable within a specific range of temperatures. If exposed beyond this temperature range, the medicine no longer works as intended. In this paper, an Internet of Things (IoT) sensor-based blockchain framework is proposed that tracks and traces drugs as they pass slowly through the entire supply chain. On the one hand, these new technologies of blockchain and IoT sensors play an essential role in supply chain management. On the other hand, they also pose new challenges of security for resource-constrained IoT devices and blockchain scalability issues to handle this IoT sensor-based information. In this paper, our primary focus is on improving classic blockchain systems to make it suitable for IoT based supply chain management, and as a secondary focus, applying these new promising technologies to enable a viable smart healthcare ecosystem through a drug supply chain.

A Decentralized Privacy-Preserving Healthcare Blockchain for IoT.

Medical care has become one of the most indispensable parts of human lives, leading to a dramatic increase in medical big data. To streamline the diagnosis and treatment process, healthcare professionals are now adopting Internet of Things (IoT)-based wearable technology. Recent years have witnessed billions of sensors, devices, and vehicles being connected through the Internet. One such technology-remote patient monitoring-is common nowadays for the treatment and care of patients. However, these technologies also pose grave privacy risks and security concerns about the data transfer and the logging of data transactions. These security and privacy problems of medical data could result from a delay in treatment progress, even endangering the patient's life. We propose the use of a blockchain to provide secure management and analysis of healthcare big data. However, blockchains are computationally expensive, demand high bandwidth and extra computational power, and are therefore not completely suitable for most resource-constrained IoT devices meant for smart cities. In this work, we try to resolve the above-mentioned issues of using blockchain with IoT devices. We propose a novel framework of modified blockchain models suitable for IoT devices that rely on their distributed nature and other additional privacy and security properties of the network. These additional privacy and security properties in our model are based on advanced cryptographic primitives. The solutions given here make IoT application data and transactions more secure and anonymous over a blockchain-based network.