What are the types of Agents in AI | What do they do
The term “Artificial Intelligence” is getting quite a buzz for the last two decades but most people still don’t have the basic knowledge about the different things of AI. For example, even after knowing about AI and using so many AI-powered devices you may be still wondering what an Agent in AI does or how many types of agents in Ai are there, you may not even know what exactly an agent is in Ai.
If you want to everything about agents in artificial intelligence and what are the different kinds of agents are there you came to the right place.
An Artificial Intelligence system, as its name suggests works by mimicking human intelligence and thinking, an Ai system works depending on two parts, its environment, and the rational agents. Agents basically sense the environment through sensors and act on their environments through actuators. Based on their role differences different agents have different properties. Let’s learn about this in more depth and learn what are the types of agents in Ai.
What AI agent actually is?
As long as something is using sensors and actuators to sense and act upon its environments it can be called an agent. Not only in AI systems, but Agents can also be seen in our everyday lives. Something that sees then thinks and then perceives can be called an agent. So an agent can be,
Types of agents around us
Human-Agent– As humans can see and sense and humans can also think and act onto it using arms, legs, or voice humans are a type of agent.
Robotic-Agent– As robots mimic human acts through sensing via camera and integrating with various motors, robots are a type of agent
Software-Agent– As software acts upon similar example, for example seeing, thinking, and acting upon using various technologies software are the third kind of agent.
How does an Agent work?
An Agent has three steps working process,
- Seeing
- Thinking
- Doing
Seeing- Using its sensors agents perceives the world. Here it only sees the world but does not make any changes to it.
Thinking- Once it has taken all the information needed it starts calculating and analyzing them according to how it is programmed.
Doing- Here the agent does something that makes physical changes or modifies something in the environment.
The Types of Agents in Ai
Ai agent is a part of a software agent which is further divided into categories. There are 5 types of agents in the AI system.
- Reflex Agent
- Goal Based Agent
- Utility Agent
- Learning Agent
- Monitoring Agent
Reflex Agent
As its name suggests Reflex Agents work just like the human body. Just like a quick response of the body from heat or fear of environmental elements. Reflex Agent in An Artificial Intelligence system also does a prompt action based on the current situation. It cant act based on previous environmental scenarios. Just like different types of agents in Ai, they are further divided into categories.
Simple Reflex Agent
Simple reflex agents act based upon only the current scenarios and will ignore all the previous histories. The actions of this agent are based on some simple This-That conditions. For example in a car body parts painting tunnel, the AI will keep spraying that particular part unless it’s completely colored. They can only work in observable environments. As for the previous example, if they can’t see something properly they will not be able to paint it.
Model Based Reflex Agent
It’s the more advanced version of the simple reflex agent it works by finding a rule which is suitable for the current situation. It is given a model of the environment, as it can work depending on the previous situation too, it can use the model to work on a partially observable environment. The other types of agents in Ai are,
Goal-Based Agents
These kinds of agents work on certain goals in order to minimize the time reaching it. It helps the model to accelerate things as here the model can simulate and choose what situation will be best to decrease the end time for reaching the goal. The knowledge that supports its decisions is represented explicitly and can be modified, which makes these agents more flexible. They usually require search and planning. The goal-based agent’s behavior can easily be changed. For the next types of agents in Ai we have,
Utility-Based Agents
For these types of agents in Ai, the end goal for the Ai works as a foundation. For example, in a smart temperature control system here if the goal is keeping the temperature at 24 degrees centigrade, once it is reached it will stop cooling any further and when the temperature rises it will get started automatically. For the utility-based AI agent, there are multiple possible alternatives like for example,
Another place where a utility-based AI agent is used is a GPS system were reaching the destination is not enough it needs to be quick and convenient. Here the Utility-Based Ai agent will take these decisions keeping convenience in mind.
Learning Agent
it’s the fourth kind of Ai agent and by far the smartest one of types of agents in Ai. It learns from past experiences. At first steps during the Ai development give it some basic knowledge and commands, from here they act and adapt and keep learning and keep getting smart automatically. It has four components.
Components Of Learning Based AI Agent
Just as there are different types of agents in the Ai system this one also has 4 different types.
Learning Element: As its name suggests learning agents’ basic task is to learn and make improvements through the environment.
Critic Element: It gives feedback to the learning element on how good it is doing or how good it is learning.
performance Element: It chooses which external factor should it be subject to work for.
Problem Generator: This component suggests actions that will lead to new and informative experiences.
How a Learning Based AI Agent is different from Machine Learning?
Machine learning is a subset of Ai. Most of the Ai application today requires some kind of ML algorithms as intelligent software always requires some kind of knowledge which ml gives to the AI software. As there are differences between different types of agents in Ai there is a difference here too.
Though the difference here is a pure AI application depends completely on the algorithms that engineers develop simulating human behavior and Machine Learning on the other hand ML is fueled by data and learns from there.
How does a Perfect Ai Agent work?
- It should be able to analyze a correct right action among many other options
- An Agent should understand the action that will accrue, even its exact the same it may not result the same in reality
- It should be able to measure its performance
- Doing something until the problem is solved
- It needs to have knowledge of the environment
Conclusion
There are mainly three types of agents in the environment, Human, Robotic and Software agents. Software agents are divided further into Ai agents and there are also 5 types of agents in Ai that you just saw above. Even those have further divided categories and components. All of them have different roles in an Ai application, all of them work together to get desired results. As we learned just now. Hope we could solve your every question regarding the types of agents in Ai. Hope you like it. Have a great day.