Sokoban Research work project using DNN or Reinforcement Learning
Bütçe ₹12500-37500 INR
I am looking for a Freelancer to help me with a Sokoban research work project using either Deep Neural Network or Reinforcement Learning. No existing resources or datasets are needed, however the project will take 3 weeks. The research work involves experimenting and applying different deep learning algorithms or reinforcement learning algorithms to optimize strategies within the game of Sokoban. The goal of this project is to research the behavior of AI agents in handling tasks such as block relocation and identify the optimal solutions for each scenario. The desired outcome of the project is gaining insight into how to improve agent performance. This project will provide a great opportunity to gain experience in deep learning and/or reinforcement learning.
Objective: We need to create new heuristic function for the Sokoban game. Also we should vizualise the game for user interface and show how the model performs (moves by the algorithm).
Idea: I have refered to this paper ([login to view URL]). Also you refere ([login to view URL]). The main idea is to create own heuristic function and solve the game using a DNN model preferrably.
Code: So we need to come up with our own hueristic function and create a sokoban environment and show how the model works on the environment (vizualisation).
Evaluation and a graphical report similar to a paper.
Bu iş için 3 freelancer ortalamada ₹29067 teklif veriyor
Hello Abishek, I hope this message finds you well. I will be able to help you with your project. I read through the description and it sounds like I am just the right person for it. I have the right sk Daha Fazla
Hi there. I can help you with this. I have first authored research papers published in A* conferences like CVPR. I have used RL with Q learning and Deep RL based models. Let’s discuss about exact requirements. Thanks
As you are looking for someone to create a new heuristic function for Sokoban while simulating it, I would be the best fit. Let me tell you why… I have already developed a weather prediction model as part of Kaggle c Daha Fazla