What you will learn

• AI search
• Knowledge representation
• Neural networks
• NLP

Course content (49 lessons)

1. Unit 1: Definition of Artificial Intelligence — Artificial Intelligence (AI) is a branch of Computer Science aimed at creating machines capable of performing tasks that typically require human intelligence. The 4 Definitions…
2. Definition and Scope — Definitions of AI There is no single agreed-upon definition, but they generally fall into four categories: 1. Thinking Humanly: "The exciting new effort to make computers think...…
3. History and Evolution — The evolution of AI has been a roller coaster of "Hypes" and "Winters": Era Milestone Description :--- :--- :--- 1943-1955 Gestation McCulloch & Pitts propose the first artificial…
4. Goals and Objectives — The ultimate goal of AI is to create intelligent agents. Specific objectives include: 1. Reasoning & Problem Solving: Developing algorithms to solve puzzles, play games, and…
5. Types of AI — AI is classified based on capability and functionality. Based on Capability Type Name Capability Examples :--- :--- :--- :--- ANI Artificial Narrow Intelligence Good at one…
6. AI vs Human Intelligence — A comparative analysis of Artificial vs. Biological Intelligence: Parameter Artificial Intelligence (AI) Human Intelligence (HI) :--- :--- :--- Origin Created by humans…
7. Applications of AI — AI has moved from research labs to the real world: Healthcare: IBM Watson for diagnosis, Da Vinci systems for robotic surgery, Drug discovery. Finance: Fraud detection systems…
8. Challenges and Limitations — Despite the hype, AI faces critical hurdles: Technical Challenges 1. Data Quality: Garbage In, Garbage Out. AI needs massive, clean datasets. 2. Explainability: The "Black Box"…
9. Unit 2: Intelligent Agents — An Intelligent Agent is anything that can be viewed as perceiving its environment through sensors and acting upon that environment through actuators . The Standard Model Agent:…
10. Characteristics and PEAS — To design a rational agent, we must first specify the task environment. We use the PEAS framework: PEAS Description Agent Type P erformance Measure E nvironment A ctuators S…
11. Environment Types — The environment determines the complexity of the agent. We classify environments based on these properties: Property Definition Example (Easy) Example (Hard) :--- :--- :--- :---…
12. Rationality — Rationality does not mean "omniscience" (knowing everything). It means doing the "right thing" given what the agent knows. Definition of a Rational Agent For each possible percept…
13. Problem Solving in AI — When an agent needs to achieve a goal but doesn't know how, it uses Problem Solving (Search). The Problem Solving Loop 1. Goal Formulation: Agent adopts a goal (e.g., "Reach…
14. State Space Representation — To solve a problem computationally, we must model the real world as a State Space . Elements of State Space 1. State: A snapshot of the world. (e.g., In Chess, the position of all…
15. Problem Characteristics — Before choosing an algorithm, analyze the problem type: Decomposable vs. Non-decomposable: Can it be split? (Integration is decomposable; 8-puzzle is not). Recoverable vs.…
16. Unit 3: Introduction to Search Strategies — Search is the universal problem-solving mechanism in AI when the sequence of steps is not known in advance. Problem Formulation To search effectively, we define: 1. Initial State:…
17. Uninformed Search Techniques — Uninformed Search (Blind Search) has no clue how close a state is to the goal. It just traverses systematically. Comparison of Algorithms Strategy Description Data Structure…
18. Informed Search Techniques — Informed Search (Heuristic Search) uses a "rule of thumb" or Heuristic Function $h(n)$ to estimate the cost to the goal. The Heuristic Function $h(n)$ $h(n)$ = estimated cost from…
19. Comparison of Search Algorithms — Choosing the right algorithm depends on the constraints (Time vs. Space vs. Optimality). Property BFS DFS UCS A :--- :--- :--- :--- :--- Time Complexity $O(b^d)$ (Exp) $O(b^m)$…
20. Applications of Search — Search is the foundation of planning and navigation in AI. Real-World Examples 1. GPS Navigation: (Google Maps) uses A (or contractions hierarchies) to find the fastest route…
21. Unit 4: Why Knowledge Representation? — Search is useful, but it doesn't "know" anything about the world. A doctor AI needs medical knowledge, not just a search tree. Knowledge Representation (KR) is about encoding…
22. Types of Knowledge — Knowledge isn't just one thing. It comes in various flavors: Type Definition Example :--- :--- :--- Declarative Knowing what . Facts, concepts, and relations. "Canberra is the…
23. KR Techniques — How do we actually encode knowledge? Here are the 4 main approaches: 1. Logical Representation Using formal logic (Propositional or First-Order Logic). Unambiguous but complex.…
24. Reasoning in AI — Reasoning is the process of drawing inferences (new facts) from existing knowledge. Types of Reasoning Method Logic Direction Certainty Example :--- :--- :--- :--- Deductive…
25. Forward vs Backward Chaining — The way an inference engine applies rules matters: Forward Chaining (Data-Driven) Starts with: Known facts (Data). Method: Applies rules to infer new facts until the goal is…
26. Basics of Expert Systems — An Expert System is a computer program that simulates the judgment and behavior of a human or an organization that has expert knowledge and experience in a particular field.…
27. Unit 5: Introduction to Machine Learning — Machine Learning (ML) is a subset of AI that focuses on building systems that learn from data, identifying patterns, and making decisions with minimal human intervention.…
28. Types of Machine Learning — ML algorithms are classified by how they learn (learning style). Type Description Key Variable Example :--- :--- :--- :--- Supervised Learning with labeled data (Teacher). Input…
29. Supervised Learning Algorithms — In Supervised Learning, the algorithm learns a mapping function from input to output. 1. Regression (Predicting Numbers) Used when the output variable is continuous. Linear…
30. Unsupervised Learning Algorithms — In Unsupervised Learning, there are no correct answers. The goal is to discover hidden structures. 1. Clustering Grouping similar data points together. K-Means: Partitions data…
31. Key Terminology in ML — Understanding these concepts is crucial for building good models: Training Set: Data used to teach the model (80%). Test Set: Data used to evaluate the model (20%). Overfitting:…
32. The ML Usage Process — The lifecycle of an ML project typically involves 7 steps: 1. Data Gathering: Collecting raw data from sources. 2. Data Preparation: Cleaning, formatting, and normalizing data. 3.…
33. Unit 6: Intro to NLP — Natural Language Processing (NLP) is the branch of AI that gives computers the ability to understand, interpret, and generate human language in a valuable way. The Challenge of…
34. Representing Text (Vectorization) — Computers can't understand strings; they only understand numbers. We must convert text to numbers ( Vectorization ). Techniques 1. Bag of Words (BoW): Counts word frequency. Loses…
35. Common NLP Tasks — NLP powers many tools we use daily: Sentiment Analysis: Detecting mood (Positive/Negative). Used in brand monitoring. Machine Translation: (Google Translate). Sequence-to-Sequence…
36. Introduction to Computer Vision — Computer Vision (CV) enables computers to "see" and interpret the visual world. It aims to reconstruct 3D scene properties from 2D images. How Computers See Images To a computer,…
37. Core CV Tasks — What can we actually do with these images? Task Question Answered Output :--- :--- :--- Image Classification "What is in this picture?" Label (e.g., "Cat") Object Detection "Where…
38. Applications of CV — CV is revolutionizing the physical world: 1. Healthcare: Detecting tumors in Radiology scans (often better than human doctors). 2. Autonomous Driving: Tesla Autopilot uses CV to…
39. Unit 7: Ethical Issues in AI — As AI systems pervade society, they raise profound ethical questions that go beyond code. Key Ethical Dilemmas 1. Job Displacement: Automation replacing human labor (e.g., Truck…
40. Bias and Fairness — AI models learn from human data, and human data contains human prejudice. This leads to Algorithmic Bias . Types of Bias Bias Type Definition Example :--- :--- :--- Historical…
41. AI Safety and Security — Refers to two different but related fields: Safety vs. Security Feature AI Safety AI Security :--- :--- :--- Focus Preventing accidents/unintended consequences. Preventing…
42. Explainable AI (XAI) — Modern Deep Learning models are often Black Boxes —accurate but inscrutable. The Need for XAI In high-stakes fields (Law, Medicine, Finance), "The computer said so" is not enough.…
43. The Future: AGI and Singularitry — Where is this all going? Definitions of future AI levels: Narrow vs General vs Super 1. ANI (Narrow AI): Super-human at one task (Chess, Protein folding). (We are here) . 2. AGI…
44. Human-AI Collaboration — The near-term future is not Human vs AI, but Human + AI ( Augmented Intelligence ). Synergy Examples Medicine: AI scans 1000 X-rays, Doctor reviews the flagged 10. Coding:…
45. Unit 8: Development Environments & Platforms — Building AI doesn't require a supercomputer anymore. Cloud platforms make it accessible. Popular Platforms Platform Best For Key Features :--- :--- :--- Google Colab Students &…
46. Key AI Libraries (Python) — Python is the lingua franca of AI. Here is the standard stack: The AI Stack Library Purpose Code Example :--- :--- :--- NumPy Math & Matrices np.array([1, 2]).sum() Pandas Data…
47. Case Study: AlphaGo (Reinforcement Learning) — In 2016, DeepMind's AlphaGo defeated Lee Sedol (18-time world champion) at Go, a game previously thought impossible for AI due to its high branching factor. How it worked 1.…
48. Case Study: ChatGPT (Generative LLM) — Released by OpenAI in 2022, ChatGPT became the fastest-growing app in history. Technology: The Transformer It uses the Transformer architecture (specifically GPT - Generative…
49. Mini Project: Sentiment Classifier — Let's build a simple AI that detects if a movie review is Positive or Negative. Steps 1. Data: Get 100 movie reviews labelled (Position/Neg). 2. Vectorization: Convert words to…

Unit 1: Definition of Artificial Intelligence

Artificial Intelligence (AI) is a branch of Computer Science aimed at creating machines capable of performing tasks that typically require human intelligence.

The 4 Definitions (The Quad)

AI is defined in four categories based on Thinking vs. Acting and Humanly vs. Rationally:

Definition and Scope

Definitions of AI

There is no single agreed-upon definition, but they generally fall into four categories:

1. Thinking Humanly: "The exciting new effort to make computers think... machines with minds." (Haugeland, 1985)
2. Thinking Rationally: "The study of mental faculties through the use of computational models." (Charniak, 1985)
3. Acting Humanly: "The art of creating machines that perform functions that require intelligence when performed by people." (Kurzweil, 1990). This is linked to the Turing Test.
4. Acting Rationally: "Computational Intelligence is the study of the design of intelligent agents." (Poole et al., 1998). This is the modern standard.

Scope of AI

AI is a multidisciplinary field:

• Machine Learning (ML): Teaching computers to learn from data.
• Natural Language Processing (NLP): Interaction between computers and human language.
• Computer Vision: Deriving meaningful information from images/videos.
• Robotics: Designing physical robots to perform tasks.
• Expert Systems: Simulating the decision-making ability of a human expert.
• Speech Recognition: Converting spoken language into text.

Frequently asked questions

Is the AI Fundamentals course really free?

Yes. The entire AI Fundamentals course on Siksha Sarovar is free to read with no account required. You can optionally sign in with Google to save your progress.

Do I get a certificate for AI Fundamentals?

Yes — finish the lessons and pass the quiz to earn a free, verifiable certificate you can share on LinkedIn or with recruiters.

Can I run code while learning?

Yes. The built-in online compiler runs C, C++, Python, Java, PHP, JavaScript, C# and SQL directly in your browser — no installation needed.