Show understanding of Deep Learning, Machine Learning and Reinforcement Learning and the reasons for using these methods

1 Foundations – AS‑Level Content (Syllabus 1‑12)

1.1 Information Representation

• Number systems – binary, octal, decimal, hexadecimal.
• Binary arithmetic: addition, subtraction, multiplication, division.
• Two’s‑complement for signed integers; overflow detection.
• Floating‑point representation (IEEE‑754 single precision).
• BCD and packed‑BCD for decimal data.
• Prefixes – kibi (Ki), mebi (Mi), gibi (Gi) etc.

Example

Convert the decimal number -23 to an 8‑bit two’s‑complement binary value:

23  = 00010111
Invert bits → 11101000
Add 1 → 11101001  (‑23)

1.2 Data Structures & Algorithms

1.3 Algorithm Design & Problem Solving

• Algorithm notation – pseudocode, flowcharts, N‑step notation.
• Complexity analysis – time (Big‑O) and space.
• Common patterns: search (linear, binary), sort (bubble, selection, insertion, merge, quick).
• Recursion – base case, recursive step, stack‑frame illustration.

Example: Binary Search (pseudocode)

function binarySearch(A, target):
low ← 0; high ← length(A)‑1
while low ≤ high:
mid ← (low+high) // 2
if A[mid] = target return mid
else if A[mid] < target low ← mid+1
else high ← mid‑1
return NOT_FOUND

1.4 Computer Architecture & Processor Fundamentals

• CPU components – ALU, control unit, registers, cache.
• Fetch‑decode‑execute cycle; instruction formats (R‑type, I‑type, J‑type).
• Machine language vs. assembly language; basic MIPS/ARM examples.
• Interrupts, pipelining basics, hazards.

1.5 System Software

• Operating system functions – process management, memory management, file system, I/O control.
• System calls and API examples (open, read, write, close).
• Virtual memory – paging, page tables, page‑fault handling.

1.6 Translators

• Assembler, compiler, interpreter – purpose and workflow.
• Phases of compilation – lexical analysis, parsing, semantic analysis, optimisation, code generation.
• Intermediate code (three‑address code) and its advantages.

1.7 Communication & Networks

• Network topologies – star, bus, ring, mesh, hybrid.
• OSI model – 7 layers, key functions of each layer.
• Common protocols – HTTP, FTP, SMTP, TCP, UDP, IP.
• Data transmission concepts – bandwidth, latency, throughput, error detection (parity, checksum, CRC).

1.8 Security & Ethics

• Confidentiality, integrity, availability (CIA triad).
• Symmetric encryption (DES, AES) and asymmetric encryption (RSA, ECC).
• Hash functions (MD5, SHA‑1/2) and digital signatures.
• Ethical issues – privacy, bias, intellectual property, professional conduct.

1.9 Databases

• Relational model – tables, primary keys, foreign keys.
• SQL – DDL (CREATE, ALTER, DROP) and DML (SELECT, INSERT, UPDATE, DELETE).
• Normalization – 1NF, 2NF, 3NF with example tables.
• Indexing – B‑tree index basics, impact on query performance.

1.10 Programming Fundamentals (Python/Java‑like pseudocode)

• Variables, data types, operators, control structures (if, switch, loops).
• Procedures / functions – parameters, return values, scope.
• Arrays, strings, simple file I/O.
• Basic debugging techniques – print‑statement, step‑through, breakpoints.

1.11 Software Development Life‑Cycle (SDLC)

• Stages – requirements, design, implementation, testing, deployment, maintenance.
• Modelling tools – UML class diagram, sequence diagram, use‑case diagram.
• Testing levels – unit, integration, system, acceptance; test‑case design (black‑box, white‑box).

1.12 Evaluation of Solutions

• Correctness, efficiency, maintainability, usability, reliability.
• Use of metrics – time/space complexity, cyclomatic complexity, code coverage.

2 Advanced – A‑Level Extensions (Syllabus 13‑20)

2.1 User‑Defined Data Types & Abstract Data Types (ADTs)

• Structures / records – grouping heterogeneous fields.
• Classes (OOP) – encapsulation, inheritance, polymorphism.
• Interface definition – abstract methods, contracts.

2.2 File Organisation & Advanced File Processing

• Sequential, random, indexed, and hashed file organisations.
• File access methods – sequential read/write, random access via seek.
• Binary file formats – fixed‑length vs. variable‑length records.
• Example: Reading a binary record of a student (ID int, GPA float) in C‑like pseudocode.

2.3 Floating‑Point Arithmetic (Extended)

• Normalised scientific notation, exponent bias, rounding modes.
• Common errors – catastrophic cancellation, overflow/underflow.

2.4 Network Protocols & Data Link Layer

• TCP vs. UDP – connection‑oriented vs. connection‑less, flow control, congestion control.
• Ethernet framing, MAC addressing, ARP, VLAN tagging.
• Protocol stacks – client‑server model, socket programming basics.

2.5 Virtual Machines & Runtime Environments

• Purpose – platform independence, sandboxing, JIT compilation.
• Examples – Java Virtual Machine (JVM), .NET CLR, Python interpreter.
• Bytecode vs. native code – trade‑offs.

2.6 Advanced Security Topics

• Public‑key infrastructure (PKI), certificate authorities.
• Secure protocols – TLS/SSL handshake, SSH.
• Authentication methods – passwords, biometrics, multi‑factor.
• Common attacks – Man‑in‑the‑Middle, SQL injection, buffer overflow.

2.7 Artificial Intelligence (AI) – Complete Overview

2.7.1 Graph‑Based AI

• Search graphs – state‑space representation, uninformed (BFS, DFS, Dijkstra) and informed (A*) search.
• Knowledge graphs – triples (subject‑predicate‑object), RDF, SPARQL queries.
• Decision trees – also a basic ML classifier; entropy, information gain.

2.7.2 Machine Learning (ML)

Key algorithms (A‑Level depth)

2.7.3 Deep Learning (DL)

• Artificial Neural Networks (ANNs) – layers of neurons, weights, biases, activation functions (sigmoid, ReLU, tanh).
• Training – loss function L, gradient descent, back‑propagation:

  
  w ← w – η · ∂L/∂w

• Architectures

  • Convolutional Neural Networks (CNNs) – convolution + pooling for images.
  • Recurrent Neural Networks (RNN) / LSTM – memory cells for sequences.
  • Transformers – self‑attention, dominant in NLP and multimodal tasks.

2.7.4 Reinforcement Learning (RL)

• Modelled as a Markov Decision Process (MDP): set of states S, actions A, transition probability P(s'|s,a), reward R(s,a).
• Policy π(s) → a; objective is to maximise expected return Gₜ = Σγⁿ Rₜ₊ₙ.

Core RL algorithms

2.8 Recursion & Advanced Programming Paradigms

• Recursive problem solving – divide‑and‑conquer, backtracking (e.g., N‑Queens).
• Functional programming concepts – pure functions, higher‑order functions, immutability.
• Event‑driven programming – callbacks, listeners (GUI, network).
• Exception handling – try/catch/finally, custom exception classes.

2.9 Advanced File Processing

• Random‑access files – seek(), tell(), record locking.
• Binary serialization – struct packing, endianness considerations.
• XML/JSON parsing – DOM vs. SAX, schema validation.

3 Assessment Objectives (AO) Alignment

4 Examination Overview (Papers 1‑4)

• Paper 1 – Theory (2 hrs) – AO1 & AO2. Includes short‑answer, extended‑response, and data‑analysis questions covering all syllabus blocks.
• Paper 2 – Problem Solving (2 hrs) – AO2. Requires algorithm design, pseudocode, flowcharts, and analysis of given scenarios.
• Paper 3 – Programming (2 hrs) – AO3. Practical coding in a high‑level language (Python/Java). Tasks may involve data structures, file I/O, recursion, and a small AI component (e.g., implementing a decision tree).
• Paper 4 – Practical (2 hrs, optional A‑Level) – AO3. Hands‑on programming on a computer, testing and debugging a supplied program.

Paper‑specific checklist

5 Evaluation Metrics for AI & Machine Learning Models

6 Why Use These Methods? – Comparative Summary

7 Suggested Diagrams (for classroom use)

• Flowchart of the AI hierarchy: AI → Graph‑based techniques → Machine Learning → Deep Learning (ANNs) → Reinforcement Learning, with example applications attached to each node.
• CPU fetch‑decode‑execute pipeline diagram.
• OSI model stack with protocol examples at each layer.
• Decision‑tree split illustration with entropy calculation.
• Q‑learning update equation visualised on a simple grid world.