Goal: Use the agent to explore the Android UI to reach as many new UI states as possible. By exploring the states, agent hopes to find possible combinations of events/inputs that lead to unintended behavior like app crashes.
I have integration Droidbot ([login to view URL]) with two agents: Actor Critic and DQN. Since I made it a gym environment, you can use any type of algorithm from any RL package that works with gym.
I modified Droidbot to work as a gym environment. This allows you to integrate Droidbot with any major Reinforcement Learning library that works with gym environments (which is all of the major ones). I then intergrated the Droidbot gym environment with the Stable Baselines Reinforcement Learning library. I trained multiple agents and provided a script for running the Actor-Critic agent in the code's README file. This will generate an agent that learns to explore the android app UI and the agent will train a policy for the interaction. I also included a Deep Q Network agent and instructions for how to run that agent.
I added an option to add unexplored events to the action space in the env so the agent can prioritize them.
Going forward here are the issues and improvements that can be done:
Modify the agent to prioritize taking unexplored events. Those unexplored events would then be added to the DQN agent's replay buffer. Only when the agent is choosing between explored events does the agent's selection process kick in.
Explore parellelization for greater speed. Android emulators run a bit slow. Droidbot can be parallelized and RL libraries usually can though it's unclear the best way to do so.
Train agents with various hyperparameters to find the ideal agent parameter settings.
Decide on a better observation/state space representation. I created an observation space for the past four frames using image data. The papers you sent me all have different and interesting ways of doing the observation space representation before inputting it into the model. These are all quite involved and many possible choices can be made. I like the way Humanoid describes it and some code may be available from the github repo to make the transition easier.
Decide on a better action space. This env is complicated because the number of actions at each step is different. I made the env so that the number can be regenerated at each step or fixed at the beginning. I defaulted it to a fixed number since RL libraries typically prefer it this way. You can do a variable number of actions but that typically involves inputting a state representation and an action representation into the model for each action, scoring them, and then taking the best one. However since there is no action representation, I couldn't model it that way. Humanoid has a good option for modelling action representations in either a fixed action space or a variable action space and some of the code may be available from that github repo. The other research papers you had have some other alternatives.
13 freelance font une offre moyenne de $531 pour ce travail
Hello, I have 6 years of experienced on mobile application ( Android +iphone/ ipad ) and App admimn with development web services where admin can manage the app from backend ,. I have developed the Transport App , R Plus
Hello Sir, How are you? you are doing well!!! I am Pratik, a Business Development Manager at Shreeyait Solutions ([login to view URL]) and we are giving best Mobile app services since 2013 and have ver Plus
Dear!!! Hiring Manager, As a web/mobile app programming expert with 10+ years of experience and powerful mathematics basics I always have designed and built awesome software products based on the strong logic and abs Plus
Hello, I have gone throw details of your project , i am highly experience developer and working with a small team as i have designer and developer for android ,ios and backed... as i have understood you need android Plus
Hello Greetings...!! I have read your project requirements. I am familiar with all required technologies and I have experienced resource to start with your project. Either we can go with fix cost or you can hire us o Plus
Hi, Dear! I have read through your job details very carefully! Practical and Possible! I have 5+ years of experiences in many AI projects includig Finacial anova, ancova regression, Stock Management with R program and Plus
Hello, Thanks for posting this project. Here we read your requirement and understand it. We have some queries . So we want to discuss with you. Please message us then we can discuss in detail and provide you exact e Plus
First:- Chatbot --> This can be integrated by any Facebook page, Slack etc. Its main aim is to reduce the human effort. For example, if this Chatbot is integrated with company page so it can easily answer you some bas Plus
Dear Client. First of all, it's my pleasure to bid on your project. I am very interested in your proposal. Because I have rich experience in this field. If you choose me, you'll never be disappointed in me, and I'll do Plus
Dear Sir. Glad to meet you. I'm very interested in your job post involving these skills. I'll bring your idea to the next level or extend your live project with new awesome features. I'm web Developer specializing in Plus
Highly interested with your project and I'm ready to start right now. My completion rate is always 100% that I STRICTLY BITE ONLY WHICH I CAN CHEW. Please message me to discuss more!
Hi, I AM THE BEST CHOICE HERE. I am writing my thesis for my phd which is about RL and IRL, and all about the algorithms related. I am using gym and creating gym environments as well. I am a MuJoCo expert as well. I am Plus