Improving application launch performance in smartphones using recurrent neural network

Mobile phones became indispensable tools in our lives, withAndroid being the most used mobile OS. These devices depend onmanaging application lifecycles to improve launch performance,but the management of application processes is not done in anefficient way. The standard low memory killer, that is responsiblefor freeing memory, does not consider any user information and itfrequently kills applications that are going to be launched.Context-awareness presents several possibilities to make mobilesystems more efficient and user-driven in terms of user experience.In this paper, we introduce a context-based launcher usingrecurrent neural network (RNN), a special branch of neuralnetworks capable of remembering dependencies, taking inconsideration not just previous inputs, but also previous outputs,providing high accuracy without any extra sensor context. Oursystem guarantees that most of the applications are ready to use inthe background, substantially improving the launch time enablinga better user experience. Experimental results demonstrate that thenovel scheme can reduce application launch latency in asignificant manner.