Learning Agent

In the field of Artificial Intelligence (AI), learning agents refer to intelligent systems or algorithms that can improve
their performance over time by acquiring knowledge and experience from their environment. These agents use
various learning techniques to adapt and make better decisions based on the information they gather.

A learning agent in AI is the type of agent that can learn from its past experiences or it has learning capabilities. It
starts to act with basic knowledge and then is able to act and adapt automatically through learning.

Learning Agents is an Unreal Engine (UE) plugin that allows you to train AI characters using machine learning
(ML). This enables you to augment or replace traditional game AI, such as those written with behavior trees or
state machines. In particular, the plugin allows you to use reinforcement (RL) and imitation learning (IL)
approaches. In the long term, Learning Agents aims to be useful in a range of applications, including physics-based
animations, game-playing NPCs, and automated QA testing, etc.
Learning Agents is not a general purpose ML framework. Each aspect of the plugin has been created with
character decision-making in mind, so you wouldn't use.

Learning Agents for the following:

As the name itself suggests, in this type of learning, the agent is supervised in every means in prior itself. What it
simply means is that the correct answer for almost each example problem is fed in the Knowledge Base of the
system initially in its development phase. SO, whenever the agent confronts ay problem, it tries to find the same
problem or a similar problem in its knowledge base whose solution it already has embedded in its system. If the
problem is not there or is a lot different from those already residing in its system, then in those cases, the agent
fails to function or perform any necessary action.

2. Unsupervised Learning:-

In the unsupervised learning agents, the answers to the problems are not available with the agent in advance. In
this type of learning, the agent has to itself find the solution to the problem by learning from its past actions and
experiences. However, the required information which forms the foundation of the Knowledge Base is provided
to the agent in its development phase, but it has to find the solutions by itself. This type of agent is smarter than
the Supervised Learning agent as it has the ability to find a relevant solution to those problems also which the
agent have faced for the first time and has no prior knowledge or experience regarding it.
3. Reinforcement Learning:-

In the Reinforcement Learning method, the learning process is almost the same as in Unsupervised learning. But
the difference is that, in Reinforcement Learning, the agent is given some reward occasionally for completing any
task. Here, the goal of the agent is to get the maximum of such rewards. So, whenever any agent tries to find the
solution to any problem, it searches for an alternative which would give him the maximum reward points. This
type of learning not only makes the agent smart but also helps it to take the best possible decision according to
the utility of the developer or the user. The utility based agents use this type of learning in their systems.
Components of Learning Agents :-
1.Learning element:
It is responsible for making improvements by learning from environment. The learning element is responsible for
improvements this can make a change to any of the knowledge components in the agents. One way of learning is
to observe pairs of successive states in the percept sequence; from this the agent can learn how the world evolves.
For utility based agents an external performance standard is needed to tell the critic if the agent’s action has a
good or a bad effect on the world.

2.Critic:
Learning element takes feedback from critic which describes that how well the agent is doing with respect to a
fixed performance standard.The learning agent gains feedback from the critic on how well the agent is doing and
determines how the performance element should be modified if at all to improve the agent.
For example when you were in school you would do a test and it would be marked the test is the critic. The
teacher would mark the test and see what could be improved and instructs you how to do better next time, the
teacher is the learning element and you are the performance element.
3.Performance element:
It is responsible for selecting external action. the performance generator only suggests actions that it can already
do so we need a way of getting the agent to experience new situations, and this is what the performance generator
is for.This way the agent keeps on learning.

4.Problem generator:
This component is responsible for suggesting actions that will lead to new and informative experiences. The
performance generator only suggests actions that it can already do so we need a way of getting the agent to
experience new situations, and this is what the performance generator is for. This way the agent keeps on
learning.