WhatsApp or SMS: 07087083227
Click here to Get The Complete Research Project Chapter 1-5

Full Project – Artificial intelligence system for chrome offline game

Click here to Get this Complete Project Chapter 1-5

CHAPTER ONE

INTRODUCTION

1.0 BAKGROUND OF STUDY

Learning human-level control policies directly from high dimensional sensory data such as vision is a long-standing challenge for controlling system design. Many classical control algorithms are based on accurate modelling or domain knowledge of the underlying dynamics in the system. However, for systems with unknown dynamics and high dimensional input, these methods are unrealistic and hard to generalize.

In this project, Deep Q-learning, online learning Multi Layered Perceptron (MLP) and rule-based decision-making (Human Optimization) for learning to control the game agent in T-Rex, the classic game embedded in chrome offline mode, directly from high-dimensional image input. According to Muller (2015) all methods are based on two ways of state space simplification. The first way is to extract pixel based features and detect objects using computer vision approaches for imitating human player, such as MLP. The other is to automatically learn patters from resized raw game image without manual feature extraction, such as Deep Q-learning. To be specific, the input to Deep Q-learning to update the training samples in neural network and leverage a Convolutional Neural Network (CNN) to predict which action to take under given circumstances. Similarly, I take the extracted pixel features as input and use MLP or rule-based decision making (Human Optimization) for prediction.

The result reveal that the approaches significantly out-perform even the experienced human player in the game. I implemented the extraction of pixel-based features from the game screenshot and the MLP algorithm. I also focus on the implementation of Q-learning framework, and the comparison between hands coded online learning methods and deep reinforcement learning. Also focus efforts on the deep Q-learning network model, the choices of hyper-parameter of training process. Analyzing the effect of acceleration in T-Rex and compare the game with other game

1.1 PROBLEM STATEMENT

AI is developing with such an incredible speed, sometimes it seems magical. There is an
opinion among researchers and developers that AI could grow so immensely strong that it would be difficult for humans to control. Humans developed AI systems by introducing into them every possible intelligence they could, for which the humans themselves now seem threatened.
An AI program that recognizes speech and understands natural language is theoretically
capable of understanding each conversation on e-mails and telephones.

Threat to Human Dignity
AI systems have already started replacing the human beings in few industries. It should not replace people in the sectors where they are holding dignified positions which are pertaining to ethics such as nursing, surgeon, judge, police officer, etc.
Threat to Safety
the self-improving AI systems can become so mighty than humans that could be very difficult to stop from achieving their goals, which may lead to unintended consequences.

1.2 AIMS AND OBJECTIVES

The well-known story of TD-gammon is one of the milestones in reinforcement learning. It used a model-free TD-learning algorithm similar to Q-learning and achieved human-expert level performance. Since then, the use of reinforcement learning has popularized and various attempts have been made to apply reinforcement learning on games. In 2013, Google Deep mind proposed the use of deep reinforcement learning on training agents to play the 2600 Atari games. Taking just the pixels and reward received from the game as inputs, they were able to reach human-expect performance in multiple Atari games. The main advantage of Deep-Q learning is that no specification of the game dynamics is needed in spite of the high dimensional image input. John (2012), argued that the agent is able to learn to play the game without knowing the underlying game logic. To process the image data, they use a deep Q-network (DQN) to directly evaluate the Q function for Q-learning. An experience replay is also applied to de-correlate experiences. This framework is model-free and can generalize to a lot of similar problems. After their research, many papers tried to make improvements. Further improvements involve prioritizing experience replay, more efficient training, and better stability when training.

SIGNIFICANCE OF STUDY

Artificial Intelligence is the machines which are designed and programmed in such a manner

It thinks and act like a human.

Artificial Intelligence becomes the important part of our daily life. Our life is changed by AI because this technology is used in a wide area of day to day services.

These technologies reduce human effort. Now in many industries, people are using this technology to develop machine slaves to perform the different activity (John, 2012).

Using the machine for the work speed up your process of doing work and give you an accurate result. The introduction of AI brings the idea of error free world. This technology will slowly introduce in all the sector to reduce human effort and give accurate and faster result

AI role in Gaming Zone

Computer and TV games got more development and updates in its fields. There was a time when “Super Mario” was considered as the best g.ame. But nowadays there are different gaming bots are introduced and you don’t have to wait for other to play with yours. Both are developed who will play with you.

1.4 SCOPE OF THE STUDY

This topic “AI chrome offline game” has a very wide scope, which include the following areas:

Self-learn with Deep Q-learning.
Keeping the user busy when internet connection is down.
Games tend to relax the mind of users and increase eye sight and vision.
Increase in key proficiency and speed.
- LIMITATIONS OF STUDY

The assessment of relevant information and materials for this study was limited by the following factors;

Inadequate finance to exhaustively access the needed materials and internet.
Inadequate internet facility available for the research.
Inadequate power supply.
Time available for the researcher.

AI TERMINOLOGY

Here is the list of frequently used terms in the domain of AI:

Term	Meaning
Agent	Agents are systems or software programs capable of autonomous, purposeful and reasoning directed towards one or more goals. They are also called assistants, brokers, bots, droids, intelligent agents, and software agents.
Autonomous Robot	Robot free from external control or influence and able to control itself independently.
Backward Chaining	Strategy of working backward for Reason/Cause of a problem.
Blackboard	It is the memory inside computer, which is used for communication between the cooperating expert systems.
Environment	It is the part of real or computational world inhabited by the agent.
Forward Chaining	Strategy of working forward for conclusion/solution of a problem.
Heuristics	It is the knowledge based on Trial-and-error, evaluations, and experimentation.
Knowledge Engineering	Acquiring knowledge from human experts and other resources.
Percepts	It is the format in which the agent obtains information about the environment.
Pruning	Overriding unnecessary and irrelevant considerations in AI systems.
Rule	It is a format of representing knowledge base in Expert System. It is in the form of IF-THEN-ELSE.
Shell	A shell is a software that helps in designing inference engine, knowledge base, and user interface of an expert system.
Task	It is the goal the agent is tries to accomplish.
Turing Test	A test developed by Allan Turing to test the intelligence of a machine as compared to human intelligence.

Get the Complete Project

This is a premium project material and the complete research project plus questionnaires and references can be gotten at an affordable rate of N3,000 for Nigerian clients and $8 for international clients.

Click here to Get this Complete Project Chapter 1-5

You can also check other Research Project here: