Before going into details about AI Intelligent Agents, let us first study about what an agent actually is in Artificial Intelligence.
What is an Agent?
Anything that perceives things from the environment with the help of sensors and acts upon it through effectors (robotic arm) or actuators are known as agents. The cycle in which an agent works is: Perceiving, Thinking and then Acting accordingly.
A Sensor is a device of the agent which senses the changes made in the environment and it then sends that data to other electronic devices for processing. Hence, an agent sees its environment with the help of these sensors.
An Actuator is a component which converts the energy in the machine into motion. The machine is moved and controlled by these actuators. Electric motors, gears are some examples of actuators.
Effectors are the components which finally act on the environment after all the processing. Some examples of effectors are the Robotic arms, legs, display screen, wheels or even wings attached to the machine.
An agent can be any of these:
- Human Agent
- Robotic Agent
- Software Agent
For a Human Agent then the sensors for the human agent are its eyes, ears and the effectors would be legs, hands, or mouth. Similarly, for a Robotic Agent its sensors would be its cameras, range finders and its effectors would be the motors which are attached to it. And if the agent is a Software Agent, then various keystrokes and file content act as sensors and the screen on which the output is displayed after processing acts as actuator.
Intelligent Agent –Definition
An Intelligent Agent is a goal-oriented agent. It can be referred to as an entity which tries to achieve its goals by acting on the environment with the help of sensors and actuators.
These agents are called intelligent because they have a certain level of autonomy with the help of which they learn continuously and perform specific, predictable, and repetitive tasks for the users or different applications.
Intelligent Agent – Characteristics
AI Intelligent agents must show the following characteristics:
- The AI Intelligent Agents must have a certain level of autonomy with the help of which they can perform the tasks on their own.
- They should be able to interact with other entities like humans, agents and systems.
- AI Intelligent Agent must be goal oriented.
- They must have a learning ability and they must be able to learn even when they are carrying out various tasks.
- An intelligent agent must be able to accommodate new rules incrementally.
Intelligent Agent – Structure
An AI intelligent agent is made up of three main parts. These are as follows:
- Architecture– Intelligent systems consist of actuators and sensors as their main architectural component. Hence, they would include camera, sensors, computer etc.
- Agent Function–This is a function where the mapping between the action and the percept sequence occurs. Percept sequence refers to the history of what the agent had already perceived.
- Agent Program–The agent programme finally executes the agent function.
Hence, we can say that the basic task of AI is to help design such an agent program that is able to implement the agent function. So, an intelligent agent is a concatenation of a good architecture and an agent program.
That is, AI Intelligent Agent = Architecture + Agent Program.
AI Intelligent Agents – Working Model
Intelligent agents work on the PEAS model. In this model, performance measure defines the success or failure of an agent.
- Performance Measures– This represents the sequence of steps taken by an agent in order to achieve its goal.
- Environment–It shows the interaction of agent with different types of environments.
- Actuators– It specifies the ways by which the agent would affect the environment.
- Sensors–This is the way through which the agent collects information from the environment.
Let us suppose an example to understand this model. Let us suppose an example of a self-driven car. Here,
Performance Measures are: Comfort, Safety, Time Taken, Correct Navigation.
Environment represents: Roads, Signals, other Vehicles, and Pedestrians
Actuators are: Steering Wheel, Brake, Horn, Accelerator, Indicators etc
And Sensors are: The Cameras attached, Speedometer of the car, GPS, Odometer etc.
Now, the success or failure of this self-driven car depends on its type of performance in relation to the points under the performance measures.
AI Intelligent Agents – Applications
In artificial intelligence, intelligent agents have been applied in different real life situations. Some of these applications are as follows:
- Medical Diagnostics– Intelligent agents are also being used in healthcare services. The vital characteristics of the patient are entered via keyboard into the intelligent agent then it uses this information and decides the best course of medical procedures the patient must follow. Intelligent agents also administer certain tests and treatments via the actuators to the patients.
- Repetitive Office Work – In order to lower the administrative costs now companies make use of intelligent agents in the fields of customer support and sales. With the help of intelligent agents, the office productivity is also being enhanced.
- Automatic Driving – Self driving cars make use of intelligent agents for their navigation. These cars make use of GPS, cameras and radars to collect information from the environment. This information includes pedestrians, other vehicles, roads, signals etc. Based on the information recorded the actuators (like break or accelerator here) then take appropriate actions.
- Information Searching, Retrieval and Navigation– The search engines we used to search for information on Internet use intelligent agents. Intelligent agents perform the task of searching the information on behalf of the users in a very short time. Hence, they help the user to get accurate and precise information within a short span of time.
- Vacuum Cleaners – Now even AI enabled vacuum cleaners are available which complete the task of cleaning with the help of data collected by various cameras, dirt indicator sensors, bump detectors, wall sensors attached to it. It completes the tasks via wheels, brushes and vacuum extractor attached to it.
With more and more advancements in technology AI Intelligent Agents are helping in automation of various time consuming and difficult tasks. With further more efforts intelligent agents can be applied to more complex AI driven devices which would help solve many challenges in the world ahead.