Revision notes

Information Technology and AI: Practice

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

Network generations, network types and IoT: revision table and common traps

Generations: what to remember
GenerationSignalMain additionStandards and names to recognise
1GAnalogMobile voice onlyAMPS (an early analog system)
2GDigitalDigital voice, SMS, slow packet dataGSM, CDMA; GPRS and EDGE are data upgrades often called 2.5G and 2.75G
3GDigitalPractical mobile internet, video callsUMTS/WCDMA, CDMA2000, HSPA; ITU family IMT-2000
4GDigital, all-IPMobile broadband; voice carried as data (VoLTE)LTE and LTE-Advanced; ITU family IMT-Advanced
5GDigital, all-IPHigh speed, very low latency, massive device numbers5G NR (New Radio); ITU family IMT-2020
Network types
TypeCoversTypical example
PAN - personal area networkA few metres around one personBluetooth earphones linked to a phone
LAN - local area networkOne room, building or campusOffice computers and Wi-Fi in a school
MAN - metropolitan area networkA cityA city-wide network linking government offices
WAN - wide area networkCountries and continentsThe internet itself

1G analog voice; 2G digital + SMS; 3G mobile internet (IMT-2000); 4G all-IP broadband, VoLTE (IMT-Advanced); 5G eMBB + URLLC + mMTC (IMT-2020). Higher frequency = more capacity, less range.

Key takeaways: learn each generation by its one main addition and its ITU family name. Know the three 5G use families. Range decides PAN, LAN, MAN and WAN. In numerical questions, convert bytes to bits before dividing.

Read the full lesson

Lesson 2

Cloud service models, deployment models and big data: revision and exam traps

Responsibility by service model
LayerOn-premises (own servers)IaaSPaaSSaaS
ApplicationCustomerCustomerCustomerProvider
DataCustomerCustomerCustomerCustomer (the provider stores it)
Runtime and middlewareCustomerCustomerProviderProvider
Operating systemCustomerCustomerProviderProvider
Virtualisation, servers, storage, networkCustomerProviderProviderProvider
Deployment models
Deployment modelMeaning
Public cloudOwned by a provider and shared by many customers over the internet
Private cloudUsed by one organisation only, run by itself or a provider
Community cloudShared by several organisations with common needs
Hybrid cloudA combination of public and private clouds working together
The five Vs
VQuestion it answers
VolumeHow much data is there?
VelocityHow fast does it arrive and need handling?
VarietyHow many forms does it take (structured tables, text, images, video)?
VeracityHow accurate and trustworthy is it?
ValueCan useful insight be drawn from it?

IaaS: you manage OS upwards. PaaS: you manage only the app and data. SaaS: you only use it. Hybrid = public + private. Edge = process near the source. Veracity = trustworthiness.

Key takeaways: remember the layer table, the five essential features, the four deployment models and the five Vs. Watch for the hybrid/multi-cloud trap and for questions that turn on who manages the operating system.

Read the full lesson

Lesson 3

Artificial intelligence and machine learning: key terms, learning types and exam traps

Three learning types
TypeDataTypical tasksMethods to recognise
SupervisedLabelledClassification (category), regression (number)Linear regression, decision trees, support vector machines
UnsupervisedUnlabelledClustering, finding unusual cases, reducing many features to a fewk-means clustering, principal component analysis
ReinforcementRewards and penalties from an environmentGames, robot control, sequences of decisionsAn agent learns a policy that earns the most reward
Vocabulary to know
TermMeaning
Training dataThe examples a model learns from
ModelWhat learning produces: the learnt pattern used to make predictions
Parameters (weights)The numbers inside a model adjusted during training
InferenceUsing a trained model to produce an output for new input
OverfittingThe model memorises the training data, including its noise, and does poorly on new data
UnderfittingThe model is too simple to capture the pattern even in the training data
PromptThe instruction or question given to a generative model
HallucinationA confident but false or invented output from a generative model

AI ⊃ machine learning ⊃ deep learning. Category = classification; number = regression. Great on training, poor on new data = overfitting. Inference = using the trained model.

Key takeaways: place each term in the nesting of AI, ML and deep learning; match tasks to learning types by asking 'are there labels?' and 'is the output a category or a number?'. Judge models on unseen data.

Read the full lesson

Lesson 4

Blockchain, quantum computing and semiconductors: revision tables and traps

Blockchain terms
TermMeaning
Distributed ledgerA record copied and kept in step on many computers (nodes) rather than one central database
BlockA batch of records, plus the hash of the previous block
HashA fixed-length digital fingerprint of data; any change to the data changes it completely
Consensus mechanismThe rule by which nodes agree which new block to add
Proof of workNodes compete to solve a hard computing puzzle; uses much computing power and energy
Proof of stakeValidators are chosen according to the amount of value they lock up (their stake) and lose it if they cheat
Smart contractA program stored on a blockchain that runs automatically when set conditions are met
Public versus permissionedAnyone may join a public blockchain; only approved members may join a permissioned one
Quantum terms
TermMeaning
QubitThe basic unit of quantum information
SuperpositionA qubit can be in a combination of 0 and 1 until measured
EntanglementQubits whose states are linked, so measuring one tells you about the other
DecoherenceLoss of the fragile quantum state through disturbance from the surroundings
Quantum key distributionSharing encryption keys using quantum states, so that eavesdropping disturbs them and is detected

Blockchain: ledger + hash links + consensus; PoW spends computing work, PoS puts value at stake. Qubit: superposition, entanglement, decoherence. Dopant with 5 valence electrons → n-type; with 3 → p-type.

Key takeaways: know each term in the three tables, the four properties of a good hash, and the n-type/p-type rule. Avoid the hype traps: blockchain is not bitcoin, quantum is not a faster laptop, Moore's law is not physics.

Read the full lesson

Lesson 5

India's digital public infrastructure and cyber laws: Aadhaar, UPI, DigiLocker, IT Act and DPDP Act

Three building blocks
SystemLayerRun byPoints to remember
AadhaarIdentityUnique Identification Authority of India (UIDAI)12-digit number for residents; based on biometric (fingerprints, iris, photograph) and demographic data; given statutory backing by the Aadhaar Act, 2016; used for authentication and e-KYC; not a proof of citizenship
UPIPaymentsNational Payments Corporation of India (NPCI)Launched in 2016; instant bank-to-bank transfers from a mobile phone using a UPI ID (virtual payment address), round the clock; works across banks and apps
DigiLockerDocumentsMinistry of Electronics and Information Technology, under Digital IndiaDocuments are issued into it digitally by the issuing authority, so they can be verified as genuine and shared without paper copies
IT Act provisions often asked about
ProvisionSubject
Section 43Penalty for unauthorised access to, or damage of, a computer system
Section 66Computer-related offences such as hacking done dishonestly or fraudulently
Section 66CIdentity theft (misusing another person's password or electronic signature)
Section 66DCheating by personation using a computer resource
Section 66FCyber terrorism
Section 69APower of the government to block public access to online information
Section 70BCERT-In, the national agency for responding to cyber security incidents
Section 79'Safe harbour' protection for intermediaries that observe due diligence

Aadhaar - identity - UIDAI. UPI - payments - NPCI. DigiLocker - documents - MeitY. IT Act 2000: e-records + digital signatures + cyber offences; 66A struck down (2015). DPDP Act 2023: data principal, data fiduciary, consent.

Key takeaways: link each DPI system to its layer and the body that runs it; know the IT Act's purpose and its frequently asked sections; and keep the DPDP Act's three core words - data principal, data fiduciary and consent - clearly apart.

Read the full lesson

Privacy · Terms · Refunds · Delivery