Revision notes

AI Basics: Foundation

The key points, rules to remember, tables and diagrams from all 5 lessons, on one page for quick revision. To keep a copy, use your browser's Print and choose "Save as PDF".

Lesson 1

What is Artificial Intelligence? Meaning, narrow AI and everyday examples

How a learning AI system works

  1. Examples (data) The system is shown many examples instead of being told every rule.
  2. Learn patterns It finds the patterns in those examples.
  3. Apply to new cases It uses the patterns on new cases it has never seen.
Rule-following program vs learning AI
PointOrdinary programAI system (learning type)
Who decides the rulesThe programmer writes every ruleThe system finds patterns in example data
New, unplanned situationsFails or gives an errorMakes a best guess from learned patterns
Gets better with more dataNoUsually yes, after retraining
Everyday exampleCalculator, ATM PIN checkSpam filter, face unlock, song suggestions
TypeAlso calledWhat it can doDoes it exist?
Narrow AIWeak AIOne task or a small set of related tasks, often very wellYes - the AI systems in use are classed as narrow AI
General AIStrong AI, AGI (Artificial General Intelligence)Any intellectual task a human can do, and switch between themNo - it remains a research goal

Narrow AI vs general AI

  • Narrow AI (weak AI)
    • Does one task or a small set of related tasks, often very well
    • Excellent inside its own job, helpless outside it
    • Exists: the AI systems in use are narrow AI
  • General AI (strong AI, AGI)
    • Would do any intellectual task a human can
    • Could switch between tasks
    • Does not exist: it remains a research goal
SystemAI or not?Reason
A calculator working out 245 + 380 = 625Not AIFixed arithmetic rule, no judgement or learning
A traffic signal that changes every 60 seconds by timerNot AIFixed timer, ignores real traffic
A maps app predicting 6 pm travel time from Delhi to GurugramAIPredicts from patterns in past and live traffic data
A music app suggesting songs you have never playedAILearns your taste from listening patterns

AI = machines doing tasks that normally need human intelligence. Narrow (weak) AI does one job well; general (strong) AI would do any intellectual job a human can - and has not been achieved.

Key takeaways: AI is about performing tasks that need intelligence, such as recognising, predicting and deciding. Ordinary programs follow fixed rules; learning AI finds patterns in data. Every AI in use is narrow AI, while general AI (AGI) is still a goal. Maps, spam filters, voice assistants and recommendations are everyday examples.

Read the full lesson

Lesson 2

A short history of AI: Turing test, Dartmouth 1956, expert systems and deep learning

WhenMilestoneWhy it matters
1950Alan Turing's paper 'Computing Machinery and Intelligence' proposes the imitation game, now called the Turing testFirst famous way to ask 'can machines think?' in a testable form
1956Dartmouth workshop, USA; the term 'artificial intelligence' was coined by John McCarthy for itAI is named and becomes a field of research
1970s-1980sExpert systems built from if-then rules given by human expertsFirst AI used widely in business and medicine
Periods after big promisesAI winters - funding and interest fall when results disappointShows that hype and progress go in cycles
1997IBM's Deep Blue defeats world chess champion Garry Kasparov in a matchA computer beats the best human at a game long seen as a test of intellect
2010s onwardsDeep-learning era: neural networks with many layersBig gains in speech, images, translation and chatbots

How an expert system works

  1. Human experts Experts give their knowledge.
  2. Knowledge base The knowledge is stored as if-then rules.
  3. Inference engine It applies the rules to a new case.
  4. Advice For example, MYCIN advised doctors on bacterial infections.

Hype and progress go in cycles

  1. Big promises
  2. Results fall short
  3. AI winter Funding and interest drop sharply.
  4. Renewed progress

Then back to step 1: Big promises

Remember the chain: 1950 Turing (the test) → 1956 McCarthy at Dartmouth (the name) → expert systems (rules) → AI winters (setbacks) → 1997 Deep Blue (chess) → 2010s deep learning (data + GPUs).

Key takeaways: Turing proposed the imitation game in 1950; McCarthy coined 'artificial intelligence' for the 1956 Dartmouth workshop; expert systems used hand-written rules; AI winters were periods of lost funding; Deep Blue beat Kasparov in 1997 by fast search; deep learning took off when data, GPUs and better methods came together.

Read the full lesson

Lesson 3

How AI works at a glance: data, algorithms, models, training and inference

TermPlain meaningExam-preparation comparison
Data / datasetThe examples the system learns from - numbers, text, images, soundsPrevious papers and practice questions
AlgorithmThe step-by-step method used to learn from the dataYour study method
ModelWhat the algorithm produces after learning: the stored patterns used to make predictionsThe knowledge in your head after studying
TrainingThe stage in which the model learns patterns from dataThe months of preparation
InferenceUsing the trained model on new input to produce an answerWriting the exam on new questions

From data to a model

  1. Data The examples the system learns from.
  2. Algorithm The step-by-step method used to learn from the data.
  3. Training The model learns patterns from the data.
  4. Model The stored patterns used to make predictions.

Training vs inference

  • Training
    • Happens first, usually once or now and then
    • Needs a lot of data and computing power
    • Like years of medical college
  • Inference
    • Happens every time someone uses the system
    • Example: each time you unlock your phone with your face
    • Like the doctor seeing one patient
FeatureRules-based systemLearning system
Source of knowledgeRules written by peoplePatterns learned from data
Needs large data?NoYes
Easy to explain its decision?Yes - you can read the ruleOften hard - patterns are spread across the model
Handles messy, varied inputPoorlyWell, if trained on similar examples
ExampleExpert system, fixed fraud rule 'block payments above a set limit'Face recognition, speech recognition, recommendations

Data + algorithm → training → model. Model + new input → inference → output. Garbage in, garbage out: a model is only as good as its data.

Key takeaways: data are the examples, the algorithm is the learning method, and the model is the learned result. Training builds the model; inference uses it. Test on unseen data. Rules-based systems follow human-written rules and are easy to explain; learning systems find patterns in data and handle messy input.

Read the full lesson

Lesson 4

Branches of AI: machine learning, deep learning, NLP, computer vision, speech and robotics

Layers of a neural network

  1. Input layer Receives the data, such as the pixels of a photo.
  2. Hidden layers Process it; many hidden layers make a network deep.
  3. Output layer Gives the answer, such as 'cat'.

Three ways machines learn

  • Supervised learning From examples labelled with the right answer.
    • Emails marked spam or not spam
  • Unsupervised learning From unlabelled data, finding groups or patterns on its own.
    • Grouping shoppers with similar buying habits
  • Reinforcement learning By trial and error, with rewards and penalties.
    • A program learning to play a game by playing it many times
Type of machine learningHow it learnsExample
Supervised learningFrom examples labelled with the right answerEmails marked spam or not spam
Unsupervised learningFrom unlabelled data, finding groups or patterns on its ownGrouping shoppers with similar buying habits
Reinforcement learningBy trial and error, earning rewards for good actions and penalties for bad onesA program learning to play a game by playing it many times
Branches of AI
BranchWhat it doesEveryday example
Machine learningLearns patterns from data to predict or decideFraud alerts on unusual payments, product recommendations
Deep learning and neural networksMachine learning with many-layered networks for complex dataFace recognition, modern speech recognition
Natural language processing (NLP)Understands and produces human languageHindi-English translation, chatbots, grammar suggestions
Computer visionUnderstands images and videoFace unlock, number-plate reading, spotting crop disease in leaf photos
Speech recognitionTurns spoken words into textVoice typing, voice assistants
RoboticsMachines that sense and act in the physical worldWarehouse robots, robot vacuum cleaners, factory robot arms
Expert systemsApply if-then rules taken from human expertsEarly medical-advice systems
SituationBranchClue
A toll plaza camera reads vehicle number platesComputer visionUnderstanding an image
An app translates a Tamil sign into EnglishNLP (with computer vision to read the sign)Language
A bank groups customers by spending habits without labelsMachine learning (unsupervised)Finding groups in unlabelled data
A farmer speaks a query and it appears as textSpeech recognitionVoice to text

AI ⊃ machine learning ⊃ deep learning. Language → NLP; images → computer vision; voice to text → speech recognition; physical action → robotics; if-then expert rules → expert systems.

Key takeaways: deep learning sits inside machine learning, which sits inside AI. Neural networks have input, hidden and output layers; 'deep' means many hidden layers. ML learns in supervised, unsupervised or reinforcement ways. Identify a branch by its input: language, images, speech, physical action or expert rules.

Read the full lesson

Lesson 5

Strengths and limits of AI: errors, hallucination, human checking and key terms

ErrorWhat happensExample
BiasUnfair results because the training data were unbalanced or reflected past unfairnessA face system trained mostly on one group of faces works worse for others
False positiveSays 'yes' when the true answer is 'no'A genuine payment blocked as fraud
False negativeSays 'no' when the true answer is 'yes'A spam email lets a scam through to the inbox
HallucinationA chatbot gives fluent, confident, but false or made-up informationInvents a book title, a quotation, a date or a law section

Why a chatbot hallucinates

  1. Learns patterns from text A language model is built from patterns in its training text.
  2. Produces likely words It writes the words most likely to follow.
  3. Sounds right It is designed to sound right, not to look facts up in a checked record.
  4. Plausible but may be wrong When the pattern is weak, it still produces something plausible.

A person in the loop

  1. AI gives an output
  2. Person checks it Facts such as dates, fees and figures are checked against a reliable source.
  3. Person adds context Context the AI lacks.
  4. Person decides And takes responsibility for the final decision.
TermMeaning
AlgorithmA step-by-step method for solving a problem or for learning from data
ModelThe learned patterns produced by training, used to make predictions
DatasetThe collection of examples used to train or test a model
TrainingThe stage in which a model learns patterns from data
InferenceUsing a trained model on new input to get an output
ChatbotA program that holds a conversation with people in text or voice
AGIArtificial General Intelligence: hypothetical AI able to do any intellectual task a human can
HallucinationConfident but false or made-up output from an AI system
BiasSystematic unfairness in AI output, often caused by unbalanced data

Use AI as an assistant, not an authority. Fluent is not the same as true: check every fact that matters against a reliable source.

Key takeaways: AI is fast, consistent and good at patterns, but it lacks understanding, common sense and accountability. It makes errors - bias, false positives and negatives, and hallucination. A person checks the output and takes the decision. Learn the glossary terms exactly; they appear in exam questions.

Read the full lesson

Privacy · Terms · Refunds · Delivery