Deep Transfer Learning Approach for Identification of Daily Living Human Activities

Abstract

Deep learning is one of the of AI methods where there are number of layers of information which are addressed as neurons and assists with understanding the information proficiently. AI helps the machines and frameworks to comprehend the human activities themselves and afterward answer in a manner that is controlled effectively toward the end client of that specific application, framework and so forth. Various deep learning calculations are utilized to execute the idea where the profound gaining begins the interaction by taking information from one layer and give it to the following layer of information. A great deal of data and information is put away as layers and order where they are associated with one another by organization of neurons which go about as data of interest for each layer. Transfer learning is the new concept in deep learning where the data and information is transferred from one model to other model and thus it saves time and resources and cost of utilising it. Deep transfer learning contributes to get results in different model by using already existing model or by utilising few components of already existing model. There are many profound learning systems which are utilized across different spaces to simple and work on the errand of the business. The paper will explain the types of deep transfer learning methods, its benefits as well as challenges to implement the concept. It is widely used to identify daily human activities.