Publications
Machine Learning of Fracture Morphology and Growth in Geological Media: Preliminary Study
Abstract:
Rigorously predicting the evolving fractured subsurface is challenging because of the geometric representation of such dynamic systems, and because of challenges of accurate description of where fractures grow as a result of coupled processes and natural fractures. In this study, we explored applications of machine learning at two scales by using images and results from laboratory experiments and numerical modeling. First, we used U-Net based convolutional neural network (CNN) for studying the morphology of a rough fracture at the core scale where fracture asperities are explicitly represented. Then, the CNN algorithm was applied to study fracture growth at the discrete fracture scale. We show a preliminary result of CNN prediction of the rough fracture. A dice score of 72.54% suggests the accuracy of the CNN model is acceptable.
Prediction and Diagnosis of Cardiovascular Diseases using Machine Learning: A Review
Abstract:
Diagnosis and Prediction of cardiovascular diseases has often become a challenge faced by doctors and hospitals in India as well as abroad. Despite major transformations in lifestyles of people and advancements in medical domain; heart attacks still hold a major share in the global death rate. Various Machine Learning techniques can be used for classifying healthy people from the ones suffering from heart diseases. This work intends to present a comprehensive review of prediction of Cardiac diseases by using Machine Learning based approach.
Links: [paper]
Artificial Intelligence Making Driverless Cars Smarter
Abstract:
This paper explores the impact of driverless cars in the world of cars.Driverless cars are robotic vehicles designed to operate safely and autonomously without human intervention.This kind of technology has become a concrete reality and may pave the way for future systems where computers take over the art of driving.These cars won’t have a steering wheel,accelarator or a brake pedal because they don’t need them,artificial intelligence softwares and sensors will do all the work.
Links: [paper]
Mind Reading Computer Technology
Abstract:
Modern technology has led to many new inventions to cater the growing needs of people. One of the leading inventions is that of mind reading computer. Mind is an abstract entity, which consists of sensations emotions, feelings, desires and intentions. Mind reading is a way to detect or infer person's mental states. The technology is based on the ability to read human mind with use of computers. These computers analyse human brain to detect what it is trying to convey. The paper deals with brief study of major aspects involved in mind reading technology. It provides an effective way to blend human mind and computers.
Links: [paper]