Unit 1: Fundamentals of Artificial intelligence and Intelligent Agent
What is AI?
Acting Humanly: The Turing Test Approach
This approach to AI aims to build systems that can pass the Turing Test, a benchmark for
machine intelligence. Proposed by Alan Turing in 1950, the test involves a human interrogator
communicating with a machine and another human via text. If the interrogator can't reliably tell
which is the machine and which is the human, the machine has passed the test. To succeed, an
AI needs capabilities like:
● Natural Language Processing (NLP) to communicate.
● Knowledge Representation to store what it knows.
● Automated Reasoning to answer questions and draw new conclusions.
● Machine Learning to adapt to new situations.
Thinking Humanly: The Cognitive Modeling Approach
This approach focuses on creating AI that thinks and reasons like a human. It's not just about
getting the right answer but about simulating the cognitive processes and inner workings of the
human mind. This field, often called cognitive science, combines AI with psychology and
neuroscience to build computational models of human cognition. The goal is to understand how
humans think and then replicate that process in a machine.
Thinking Rationally: The “Laws of Thought” Approach
This approach uses logic to create AI systems that can reason and make sound deductions.
The idea is to formalize human reasoning into a set of logical rules and then build systems that
can apply these rules to solve problems. This approach is powerful for tasks that can be clearly
defined by rules, but it can struggle with uncertainty and situations that don't fit a neat logical
framework.
Acting Rationally: The Rational Agent Approach
This is the most widely accepted and practiced approach to AI today. A rational agent is an
entity that acts to achieve the best possible outcome. Instead of trying to mimic human thought
or behavior, this approach focuses on building agents that are rational, meaning they make
optimal decisions based on the information they have. This is a pragmatic approach that
prioritizes performance and effectiveness over mimicking human processes.
State of the Art (Applications of AI)
AI is everywhere today, driving significant advancements across various fields. Some key
applications include:
● Natural Language Processing (NLP): Virtual assistants (Siri, Alexa), machine
translation (Google Translate), and sentiment analysis.
● Computer Vision: Facial recognition, medical image analysis, and self-driving cars.
● Robotics: Manufacturing, surgical robots, and autonomous drones.
● Expert Systems: Medical diagnosis, financial fraud detection, and geological survey
analysis.
● Machine Learning: Personalized recommendations (Netflix), spam filters, and stock
market prediction.
Agents and Environments
An agent is anything that can be viewed as perceiving its environment through sensors and
acting upon that environment through actuators.
● Sensors: These are how an agent perceives its environment. Examples include cameras
for a self-driving car or a microphone for a voice assistant.
● Actuators: These are how an agent acts on its environment. Examples include the
steering wheel for a self-driving car or a speaker for a voice assistant.
● Environment: This is the world that the agent exists in and interacts with. It can be
anything from a physical room to a digital chessboard.
The Concept of Rationality
Rationality in AI means making the right decision to achieve the best possible outcome, given
the available information. A rational agent maximizes its performance measure, which is a set
of criteria used to evaluate its success. It's about doing the right thing, which means the action
that causes the agent to be most successful.
The Nature of the Environment
Environments can be classified based on several properties that affect the design of an agent.
● Fully vs. Partially Observable: Can the agent see the entire state of the environment? A
chess game is fully observable; a poker game is partially observable.
● Single-agent vs. Multi-agent: Is the agent acting alone or with other agents?
● Deterministic vs. Stochastic: Is the next state of the environment completely
determined by the current state and the agent's action, or is there an element of
randomness?
● Episodic vs. Sequential: Is the agent's experience divided into a series of independent
episodes, or do current actions affect future states?
● Static vs. Dynamic: Does the environment change while the agent is deliberating?
● Discrete vs. Continuous: Are the states and actions a finite set or a continuous range?
The Structure of Agents
The agent program is the function that implements the agent mapping from percepts to actions.
There are four basic types of agent programs:
1. Simple reflex agents: Act based on the current percept, ignoring the history.
2. Model-based reflex agents: Maintain an internal model of the world to handle partially
observable environments.
3. Goal-based agents: Use a goal to decide which actions to take to achieve a desired
state.
4. Utility-based agents: Use a utility function to choose the best action, especially when
there are multiple goals or when the best action is uncertain.
Case Study: Creating a New Healthcare Market with AI
AI can revolutionize healthcare by creating new markets focused on personalized, efficient, and
accessible care.
1. AI-Powered Diagnostic Services: AI can analyze medical images (X-rays, MRIs, CT scans)
and patient data to assist doctors in diagnosing diseases with greater accuracy and speed. This
creates a market for AI diagnostic tools that can be used in clinics worldwide.
2. Personalized Medicine and Drug Discovery: By analyzing a patient's genetic data, lifestyle,
and medical history, AI can predict disease risks and recommend personalized treatment plans.
AI can also accelerate drug discovery by simulating molecular interactions, reducing the time
and cost of bringing new drugs to market.
3. Virtual Health Assistants: AI-powered chatbots and virtual assistants can provide 24/7
support, answering patient questions, scheduling appointments, and monitoring chronic
conditions. This creates a market for virtual care platforms that improve patient engagement and
reduce the burden on healthcare professionals.
4. Predictive Analytics for Public Health: AI can analyze population data to predict disease
outbreaks, track the spread of epidemics, and optimize resource allocation in hospitals. This
creates a market for public health intelligence platforms that can help governments and
organizations prepare for and respond to health crises.
