Artificial Intelligence Fundamentals with Python - PCEI

Prerequisites

There are no formal prerequisites. Participants should have general digital literacy, basic Python programming skills, familiarity with variables, conditions, loops, functions and collections and introductory data-analysis knowledge. Knowledge equivalent to PCEP Python Certified Entry Level Programmer and PCED Python for Data Analytics level is desirable.

Who should attend

• Individuals beginning a career in Artificial Intelligence • Aspiring AI specialists and junior data professionals • Students and career changers • Developers seeking an introduction to applied AI concepts • Business professionals who work with AI-enabled systems • Automation, operations and technical-support personnel • Educators introducing AI literacy and computational thinking • Candidates intending to progress to the PCEI-30-01 examination

Learning Objectives

At the conclusion of this course, attendees will be able to: • Explain the fundamental concepts and terminology of Artificial Intelligence • Distinguish narrow AI from the concept of general AI • Identify the principal components and subfields of AI • Explain how AI systems learn from data • Distinguish training, testing and inference • Compare supervised, unsupervised and reinforcement learning • Recognise common machine-learning algorithms and applications • Implement simple rule-, distance- and grouping-based logic in Python • Interpret accuracy, precision, recall and confusion matrices • Load, clean, organise and analyse small datasets • Prepare features and normalise numerical values • Create and interpret basic data visualisations • Explain the foundations of neural networks and deep learning • Describe introductory NLP and computer-vision concepts • Explain how generative AI and large language models operate • Construct and refine effective prompts • Recognise hallucinations, bias, privacy risks and unsafe AI interactions • Apply responsible, transparent and human-centred AI principles • Identify feasible opportunities for small AI projects • Communicate AI results and limitations to different audiences

Day 1 - Artificial Intelligence and Machine Learning Fundamentals

Core AI terminology, applications, learning concepts, capabilities, limitations and the principal types of machine learning.

Understanding Artificial Intelligence

Defining Artificial Intelligence Agents, environments and inference Input, processing and output Data, algorithms and conceptual models Narrow AI and General AI Common AI capabilities Classification, prediction and pattern recognition Everyday AI applications Search, navigation and recommendation systems Assistants, image processing and text processing

AI Subfields

Machine learning Deep learning Natural language processing Computer vision Robotics Generative AI Large language models Matching business problems to AI subfields

How AI Systems Learn

Learning from examples Features and labels Labelled and unlabelled data Training and inference Predictions and feedback loops Generalisation and robustness The effect of noise, bias and missing data

Capabilities and Limitations

Tasks AI performs effectively Lack of context and common failure cases Misclassification Hallucinations Dependence on data quality and diversity Appropriate and inappropriate uses of AI Situations requiring human judgement

Planning an AI Solution

Defining the problem Identifying objectives Data availability and suitability Stakeholder requirements Project constraints Success criteria and evaluation measures Deciding whether AI is appropriate

Types of Machine Learning

Supervised learning Classification and prediction Unsupervised learning Clustering and pattern discovery Reinforcement learning Agents, actions, environments and rewards Selecting a learning type for a problem Practical work: analyse real-world problems, determine whether AI is appropriate and identify the most suitable AI subfield and learning approach.

Day 2 - Machine Learning Logic and Evaluation with Python

Machine-learning workflows, simple Python implementations and the evaluation of classification results.

The Machine Learning Workflow

Data collection Data cleaning and preparation Feature selection Training Evaluation Testing Inference Training and testing datasets Reasons for separating training and testing data Recognising missing values, outliers and inconsistent types

Common Machine Learning Algorithms

Linear models Decision trees k-nearest neighbours k-means clustering Naive Bayes Rule-based systems Classification and clustering Matching algorithms to problems How models use rules, distances and patterns

Simple AI Logic in Python

Representing examples with lists and dictionaries Implementing rule-based classifications Combining conditions into decision rules Creating a basic sentiment classifier Grouping values using Python Calculating Euclidean distance Finding the nearest data point Implementing introductory k-NN-style logic Writing simple AI-style programs without an external framework

Evaluating Classification Results

Correct and incorrect predictions Accuracy True positives and true negatives False positives and false negatives Confusion matrices Precision Recall Choosing suitable evaluation measures Recognising good and poor model performance

Model Behaviour

Overfitting Underfitting Generalisation Training performance compared with testing performance Model limitations When additional or better data is required Practical work: develop and evaluate a small Python classifier using rules and distance calculations. Produce a confusion matrix and calculate accuracy, precision and recall.

Day 3 - Data Handling, Analysis and Visualisation

Loading, cleaning, analysing, preparing and visualising data for introductory AI workflows.

Loading and Organising Data

Reading text, CSV and JSON data Lists and dictionaries Lists of dictionaries Nested data structures Accessing and modifying nested values Converting between strings and numbers Preparing consistent records

Cleaning Data

Identifying missing values Correcting type mismatches Resolving formatting inconsistencies Detecting duplicate records Detecting incorrect or inconsistent labels Recognising noise and outliers Assessing dataset suitability The relationship between data quality and model reliability

Analysing Data with Python

Mean and median Minimum and maximum values Frequency counts Sorting by attributes Grouping and summarising data Identifying common categories Detecting patterns and trends Using the math module

Vectors, Distances and Similarity

Representing features as vectors Euclidean distance Manhattan distance Basic similarity measures Normalising values to a common range Scaling numerical features Understanding when normalisation is required Preparing feature sets for simple AI tasks

Preparing Data for Machine Learning

Introductory pandas operations Loading tabular data Selecting columns Filtering rows Combining and restructuring data Flattening nested structures Feature selection Feature extraction Preparing data before model training

Visualising Data

Matplotlib fundamentals Line charts Bar charts Histograms Titles, labels and legends Selecting an appropriate chart Interpreting trends and distributions Identifying possible outliers Using visualisation to assess data quality and readiness Practical work: load, clean and explore a dataset; prepare numerical features; calculate distances; and create visualisations that assess whether the data is suitable for an AI task.

Day 4 - Neural Networks, Deep Learning and Generative AI

Neural-network foundations, NLP, computer vision, generative AI, prompt engineering and pre-trained models.

Neural Network Fundamentals

Artificial neurons Inputs and outputs Weights and biases Layers Activation functions Feedforward processing Errors and loss Backpropagation at a conceptual level Adjusting weights to reduce errors Shallow and deep networks

Machine Learning and Deep Learning

Classical machine learning Deep learning Differences between traditional algorithms and neural networks Data and computational requirements Advantages and limitations of deep models Image and speech applications Selecting classical or deep-learning approaches

Natural Language Processing

Representing text for computers Tokens and sequences Introductory embeddings Sentiment analysis Translation Summarisation Text classification Rule-based and machine-learned NLP The importance of context

Computer Vision

Images as numerical arrays Pixels and colour channels Image classification Object detection Image segmentation Convolutional neural networks at a conceptual level Face recognition and document scanning Autonomous-system applications

Generative AI and Large Language Models

Generative and predictive AI Text, image, audio and code generation Large language models Next-token prediction Context and context limitations Model strengths and weaknesses Creativity, inconsistency, bias and hallucinations Generative AI within the wider AI ecosystem

Prompt Engineering

Context-and-task prompt structure Providing clear instructions Specifying audiences, formats and constraints Refining prompts iteratively Prompts for summarisation and explanation Prompts for classification and transformation Evaluating prompt results Prompt injection and unsafe requests Recognising when human verification is required

Pre-Trained Models and Deployment

Training and inference Using pre-trained models Transfer learning Model reuse Deployment concepts OCR, recommendation systems and chatbots Applying trained models without training from scratch Practical work: compare prompts for a set of AI tasks, evaluate outputs for accuracy and safety, and document improvements made through iterative refinement.

Day 5 - Responsible AI and the Integrated AI Project

Ethics, safety, critical evaluation, project planning, collaboration, communication and an integrated AI project.

Ethical Risks and AI Safety

Bias and discrimination Unfair outcomes Harmful stereotypes Privacy and data-protection concerns Hallucinations and misinformation Inappropriate or harmful outputs Safe and unsafe AI uses Identifying potentially high-risk applications

Safe and Secure AI Use

Personal and sensitive information Passwords and confidential documents Data minimisation Responsible input handling Adhering to organisational policies and terms of use Attempts to manipulate or bypass system controls Recognising unexpected or harmful behaviour Knowing when to stop or escalate

Social and Economic Impact

AI in business, education and public services Productivity and innovation Changes to occupations and workflows Reskilling requirements Job displacement Inequality and the digital divide Evaluating positive and negative effects

Responsible and Human-Centred AI

Fairness Transparency Accountability Explainability Human-in-the-loop processes Human oversight Verifying outputs and sources Applying human judgement Trustworthy AI deployment

Critical Evaluation of AI Outputs

Detecting incorrect and misleading statements Hallucinations and unsupported claims Logical errors and contradictions Cross-checking information Evaluating consistency and reliability Recognising bias Deciding when an automated output should be rejected

Planning an AI Project

Identifying suitable AI opportunities Project goals, inputs and outputs Data and domain-knowledge requirements Constraints and feasibility Data volume and model complexity Compute, storage and time requirements AI service and API costs Usage limits Comparing AI with simpler rule-based solutions Cost-benefit considerations

Collaboration and Communication

Analyst, developer, domain-expert and reviewer roles Shared responsibilities Version-control concepts Pair programming and code review Peer feedback Documenting decisions Reproducibility Explaining methods, outputs, errors and limitations Communicating with technical and non-technical audiences

Integrated AI Project

Selecting an appropriate AI use case Defining objectives and success measures Loading and cleaning a small dataset Selecting and preparing features Implementing rule- or distance-based AI logic in Python Evaluating results with basic metrics Visualising relevant findings Identifying ethical risks and limitations Considering costs and resource requirements Producing a concise project report Presenting recommendations to a non-technical audience

Optional Day 6 - PCEI Examination Preparation Workshop

A separate one-day instructor-led workshop for candidates who have completed the five-day course or possess equivalent knowledge.

PCEI Examination Overview

Examination structure and current syllabus The six examination blocks and their weightings Single-select and multiple-select questions Scenario-based questions Interactive items Managing the available examination time Understanding the 75% passing requirement

Structured Syllabus Review

Artificial Intelligence fundamentals Machine-learning fundamentals Data handling, analysis and visualisation Neural networks, deep learning and generative AI Responsible AI, ethics and critical thinking AI projects, collaboration and communication

Question-Answering Techniques

Reading questions precisely Identifying the concept being tested Evaluating Python logic without executing it Calculating distances and basic model metrics Interpreting confusion matrices Matching algorithms to problems Identifying unsafe or unethical scenarios Eliminating implausible answers Handling questions with multiple correct answers Recognising common distractors

Guided Practice

AI terminology and concept questions Python logic exercises Data-quality scenarios Accuracy, precision and recall calculations Neural-network and deep-learning questions Prompt-engineering scenarios Generative AI and LLM evaluation Responsible-AI case studies Project-planning and communication questions

Mock Examination

Timed 36-question mock examination Simulation of examination conditions Review of questions and answers Explanation of incorrect alternatives Performance analysis by syllabus block Identification of final revision priorities

Personal Preparation Plan

Individual instructor feedback Targeted revision recommendations Examination-day preparation Time-management strategy Follow-up practice and study resources