PROFESSIONAL DIPLOMA IN AI WITH PYTHON

This course teaches the state-of-the art Artificial Intelligence techniques that are used in contemporary software applications such as Data Prediction, Recommender Systems, Image Detection and Chat Bots. The course will mainly focus on teaching a conceptual understanding of Machine Learning and the necessary Python programming skills to build your own AI/ML applications. This course is suited for students who are keen to take a practical step into the field of AI by learning how to utilise Data in creating powerful Prediction tools.

AI & Machine learning is changing the way we use and develop software. Prediction models have shown to be highly valuable in all industries such as Healthcare, Finance, Insurance, Manufacturing, Agriculture and Entertainment. The ability to effectively develop these models is more accessible now than ever. An understanding of this subject can provide business-value to your company by making profitable use of company data, enable you to work on your own ideas or help you join one of the thousands of startups and businesses working on AI solutions.

In application to AI, Python is the most formidable and popular programming language. It is also widely used in other fields of computing such as data analytics, web development and back-end programming.

Week 1

Introduction to Machine Learning and Linear Regression

• Introduction to Machine Learning
• Linear Regression Model
• Gradient Descent
• Linear Regression in one variable
• Linear Regression in multiple variables

Week 2

Classification Models and Feature Engineering

• Logistic Regression
• Logistic Regression with Multiple Labels
• Support Vector Machines (SVM)
• Introduction to Feature Engineering
• Underfitting & Overfitting Data

Week3

Cross Validation and Performance Evaluation

• Cross Validation
• Hyperparameter Tuning
• Model Selection
• Regularisation
• Using a Baseline Predictor
• Classification Performance
• ROC & AUC Curves
• Model assessment

Week 4

Gradient-Free ML Algorithms and Neural Nets

• K-Nearest Neighbour
• Decision Trees
• Neural Nets
• Stochastic Gradient Descent

Week 5

Text-Based models and Time Series

• Text Features
• Tokenizing
• Word Importance: TF-IDF
• Bag of Words Model
• Time Series

Week 6

Convolutional-Neural-Networks

• Introduction to CNN’s
• What is Convolution
• Edge Detection
• Bells and whistles
• CNN’s: Deep Learning
• Convolutional Layer
• Pooling and FC layers
• CNN’s in actions

Week 7

Convolutional-Neural-Network Architecture and State-of-the-art

• ConvNet architecture
• Why Convolutionals
• LeNet and AlexNet
• VGG-16, VGG-19 and ResNet
• Transfer Learning
• ConvNet applications

Week 8

Recommender Systems

• Simplified Setup
• Content and Item-based
• Session based recommendations
• Matrix Factorisation

Week 9

Large Language Models

• History of Large Language Models
• Architecture of Large Language Models
• Creating ChatGPT: The Process
• Pre-training and Fine-tuning techniques

Week 10

Large Language Models (Continued)

• Embeddings in Large Language Models
• Reinforcement Learning from Human Feedback (RLHF)
• Emergence of Capabilities in LLMs Based on Complexity
• Popular LLM Models: GPT, BERT, T5
• Application and Use Cases of LLMs
• Prompting techniques and interacting with LLMs
• Limitations, Ethical Considerations and Future Development

During this course you will learn:

• The necessary Python skills to design and train AI models.
• Cutting-edge frameworks for AI in Python such as Scikit-Learn, PyTorch & Tensorflow.
• How programs can learn from data to make predictions.
• Learn and practice cornerstone Machine Learning Algorithms. for both Classification and Regression.
• How to evaluate the performance of AI Models.
• How contemporary Convolutional-Neural-Networks are designed and used for image-Classification tasks.
• How to use Language Models based on text features and their usage in the development of Recommender Systems.
• Make predictions and spot patterns given different trends and seasonal behaviour.
• How Large Language Models like GPT-3, LLaMA and Bert are developed, how to fine-tune them and their ethical considerations.