Controlling 3D gaming agent in an adversarial setting with Deep Reinforcement Learning
In order to perform a large variety of tasks and to achieve human-level performance in complex real-world environments, Artificial Intelligence (AI) Agents must be able to learn from their past experiences and gain both knowledge and an accurate representation of their environment. Traditionally
2025-06-28 16:30:56 - Adil Khan
Controlling 3D gaming agent in an adversarial setting with Deep Reinforcement Learning
Project Area of Specialization Artificial IntelligenceProject SummaryIn order to perform a large variety of tasks and to achieve human-level performance in complex real-world environments, Artificial Intelligence (AI)
Agents must be able to learn from their past experiences and gain both knowledge and an accurate representation of their environment. Traditionally, AI
agents have suffered from difficulties in obtaining a good representation of their environment and then mapping this representation to an efficient control
policy. Deep reinforcement learning algorithms have provided a solution to this issue. This project aims to train 3D gaming agents using different deep
reinforcement learning models.
To train a 3D adversarial gaming Agents to take actions on its own. Making a restricted environment for different games to train different agents.
Agents can challenge the professional human game Tekken players and even the world's top Tekken player champion human-level intelligence in beating all 3 games of Tekken.
To use different reinforcement learning techniques. e.g. Q-learning, Deep Q-learning, DDQN, etc.
We are using the Agile Method of software implementation. We are more focused on learning and implementation and result oriented approach. Try to get the result as soon as possible. We want to do it professionally by starting from designing the problem and meanwhile working on the literature review and also gathering knowledge from the different books about our project. Then start working on the prototype and produce a prototype and then produce a different version of the product as go further.
Benefits of the ProjectWe will build Gaming training agents. Those agents will help the newbies to play and also to the players to improve their game.
(Already built agents that are in-game are not so active and intelligent, but our agent will be same act like a human player so it will be more interactive with
player according to its level.
We will involve with the gaming industry and for the new coming games when they look into our agent model playing on previous games they will try to push our
model in their game for the training agent in their game.
We are also involving with Robotics Industry to bring automation due to robotics technology in the industry.
Automated agents trained on different games and try to beat the human by using their experience which agents gathered using different Q learning and DQN learning Model. This will also help us to train a robot to work as a human because games are a constraint environment if robots learn the game and beat the human then the robot will also learn the things in the real-world in a constrained environment and do the task as a human do.
Final Deliverable of the Project Software SystemCore Industry SecurityOther Industries Manufacturing , Others Core Technology RoboticsOther Technologies Artificial Intelligence(AI), Big DataSustainable Development Goals Decent Work and Economic Growth, Industry, Innovation and InfrastructureRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 70000 | |||
| Graphic Card | Equipment | 1 | 30000 | 30000 |
| Azure Credits | Equipment | 1 | 40000 | 40000 |