(searched for: doi:10.34104/ajeit.021.0970118)
Published: 2 February 2022
Australian Journal of Engineering and Innovative Technology pp 13-26; https://doi.org/10.34104/ajeit.022.013026
The rapid growth of Deep Learning (DL) based applications is taking place in this modern world. Deep Learning is used to solve so many critical problems such as big data analysis, computer vision, and human brain interfacing. The advancement of deep learning can also causes some national and some international threats to privacy, democracy, and national security. Deepfake videos are growing so fast having an impact on political, social, and personal life. Deepfake videos use artificial intelligence and can appear very convincing, even to a trained eye. Often obscene videos are made using deepfakes which tarnishes people's reputation. Deepfakes are a general public concern, thus it's important to develop methods to detect them. This survey paper includes a survey of deepfake creation algorithms and, more crucially we added some approaches of deepfake detection that proposed by researchers to date. Here we go over the problems, trends in the field, and future directions for deepfake technology in detail. This paper gives a complete overview of deepfake approaches and supports the implementation of novel and more reliable methods to cope with the highly complicated deepfakes by studying the background of deepfakes and state-of-the-art deepfake detection methods.
Published: 30 October 2021
International Journal of Management and Accounting pp 114-121; https://doi.org/10.34104/ijma.021.01140121
Machine Learning Applications have been well accepted for various financial processes throughout the world. Supervised Learning processes for objective classification by Naïve Bayes classifiers have been supporting many definitive segregation processes. Various banks in Bangladesh have found challenging moments to identify financially and ethically qualified loan applicants. In this research process, we have confirmed the safe applicant’s list using definitive variable measures through identifiable questions. Our research process has successfully segregated the given applicants using Naïve Bayes classifier with the proof of lowering loan default rate from an average of 23.26%% to 11.76% and development of financial ratios as performance indicators of these banks through various financial ratios as indicators of these banks.