Understanding Neural Networks and Deep Learning Training Course
Course Overview
Technical staff with some Python and maths learn how neural networks work and how to build feedforward, convolutional and recurrent models for business problems. Day one covers perceptrons, forward propagation, loss and backpropagation, with a first network in Python. Day two goes deeper with activation functions, SGD, Adam, RMSprop, dropout and L1/L2 regularization in TensorFlow or PyTorch. Day three trains CNNs for image classification and uses transfer learning with VGG and ResNet. Day four covers RNNs, LSTMs, GRUs, GANs and autoencoders. Day five deals with overfitting, interpretability, ethics and deployment, then ends with a group capstone.
Format of Training
- Interactive sessions that work through backpropagation and optimizer math on the board
- Labs in Python, TensorFlow and PyTorch, from a simple network to CNN image classifiers and RNN text models
- Case studies of deep learning in business settings
- Group Q&A and a team capstone on a business problem
Course Objectives
- Explain the core ideas behind neural networks and deep learning.
- Compare deep learning architectures such as CNNs and RNNs.
- Apply backpropagation and optimization methods when training networks.
- Build, train and test neural network models in common frameworks.
- Analyze how healthcare, finance and marketing use deep learning.
- Tackle common problems such as overfitting and poor interpretability.
- Tune deep learning models for better performance and scale.
Prerequisites
- A basic grounding in ML ideas
- Some programming, ideally in Python
- Basic linear algebra, calculus and statistics
- An interest in data science and AI uses
Course Outline
Day 1
Session 1: Introduction to Neural Networks and Deep Learning
- How neural networks and deep learning developed
- Core ideas: neurons, layers and activation functions
- Where businesses use deep learning
Session 2: Fundamentals of Neural Networks
- Perceptrons and multilayer networks explained
- Forward propagation and loss functions
- A first look at backpropagation
Session 3: Hands-on Lab: Building a Simple Neural Network
- Coding a basic neural network in Python
- Training it on sample data
- Plotting how the model performs
Day 2
Session 1: Deep Learning Architectures
- What deep neural networks (DNNs) are
- Depth, width and non-linearity explained
- Activation functions: ReLU, sigmoid and softmax
Session 2: Optimization Techniques in Deep Learning
- Gradient descent and stochastic gradient descent (SGD)
- Learning rate schedules and optimizers such as Adam and RMSprop
- Regularization methods: dropout and L1/L2 regularization
Session 3: Hands-on Lab: Training Deep Neural Networks
- Building deep networks in TensorFlow/PyTorch
- Trying out different optimization algorithms
- Scoring the model and fine-tuning it
Day 3
Session 1: Convolutional Neural Networks (CNNs)
- CNNs and how they process images
- Convolution, pooling and feature extraction
- Where CNNs are used in computer vision
Session 2: Hands-on Lab: Image Classification with CNNs
- Coding CNNs for image recognition
- Training on real image datasets
- Checking model accuracy and performance
Session 3: Advanced CNN Architectures
- Transfer learning from pre-trained models such as VGG and ResNet
- Data augmentation methods
- Fine-tuning models for custom uses
Day 4
Session 1: Recurrent Neural Networks (RNNs) and Sequence Models
- RNNs and why they suit sequence data
- Long Short-Term Memory (LSTM) and Gated Recurrent Units (GRUs)
- Uses in natural language processing (NLP) and time series forecasting
Session 2: Hands-on Lab: Text Classification with RNNs
- Coding RNNs to analyze text
- Sentiment analysis and sequence prediction
- Scoring the model and dealing with vanishing gradients
Session 3: Generative Models and Autoencoders
- What generative adversarial networks (GANs) are
- Autoencoders for dimensionality reduction and anomaly detection
- Practical uses of generative models
Day 5
Session 1: Challenges in Deep Learning
- Overfitting and underfitting in neural networks
- Methods for interpreting and explaining models
- Ethical questions in deep learning work
Session 2: Hands-on Lab: Model Optimization and Deployment
- Hyperparameter tuning and ways to optimize models
- Deploying deep learning models for practical use
- Good practice for scaling and maintaining models
Session 3: Capstone Project and Presentation
- Group project: designing a deep learning solution to a business problem and presenting it
- Sharing project results and main lessons
- Course close and feedback
Book This Course
Customise This Course for Your Team
Teams with image, text or time series data of their own can use it in the Day 3 and Day 4 labs. If your projects lean toward vision, we can give CNNs more time and trim the generative models session. Tell us your main use case and the framework you prefer, TensorFlow or PyTorch.
Need help with the right course to choose?
Explore more opportunities
- Duration: 1 Day
- 4.5 Ratings
Introduction to Artificial Intelligence for Business Leaders Training Course
- Duration: 2 Days
- 4.5 Ratings
AI Fundamentals: Concepts, Technologies, and Applications Training Course
- Duration: 3 Days
- 4.5 Ratings
AI in Business: Strategies for Digital Transformation Training Course
- Duration: 2 Days
- 4.5 Ratings
Machine Learning Basics for Non-Technical Professionals Training Course
- Duration: 1 Day
- 4.5 Ratings
Foundations of AI Ethics and Responsible AI Practices Training Course
- Duration: 3 Days
- 4.5 Ratings
Applied AI: Real-World Use Cases and Business Applications Training Course