Artificial Neural Networks, Machine Learning, Deep Thinking Training Course

DAY 1 - ARTIFICIAL NEURAL NETWORKS

Introduction and ANN Structure.

• Comparison of biological and artificial neurons.
• The underlying model of an ANN.
• Activation functions employed in ANNs.
• Common categories of network architectures.

Mathematical Foundations and Learning mechanisms.

• Review of vector and matrix algebra.
• Concepts related to state-space.
• Principles of optimization.
• Error-correction learning methods.
• Memory-based learning techniques.
• Hebbian learning principles.
• Competitive learning approaches.

Single layer perceptrons.

• Structure and learning processes of perceptrons.
• Introduction to pattern classifiers and Bayes' classifiers.
• Utilizing the perceptron as a pattern classifier.
• Analysis of perceptron convergence.
• Limitations inherent to perceptrons.

Feedforward ANN.

• Structure of Multi-layer feedforward networks.
• The Back propagation algorithm.
• Back propagation: training dynamics and convergence.
• Functional approximation via back propagation.
• Practical considerations and design challenges in back propagation learning.

Radial Basis Function Networks.

• Pattern separability and interpolation concepts.
• Overview of Regularization Theory.
• The relationship between Regularization and RBF networks.
• Design and training procedures for RBF networks.
• Approximation capabilities of RBF.

Competitive Learning and Self organizing ANN.

• General clustering methodologies.
• Learning Vector Quantization (LVQ).
• Competitive learning algorithms and associated architectures.
• Self organizing feature maps.
• Key properties of feature maps.

Fuzzy Neural Networks.

• Introduction to Neuro-fuzzy systems.
• Foundations of fuzzy sets and logic.
• Design principles for fuzzy systems.
• Design of fuzzy ANNs.

Applications

• Discussion of selected Neural Network applications, highlighting their benefits and associated challenges.

DAY 2 - MACHINE LEARNING

• The PAC Learning Framework
  • Guarantees for finite hypothesis sets in the consistent case
  • Guarantees for finite hypothesis sets in the inconsistent case
  • General considerations
    • Deterministic vs. Stochastic scenarios
    • Bayes error noise
    • Estimation and approximation errors
    • Model selection strategies
• Rademacher Complexity and VC Dimension
• The Bias-Variance tradeoff
• Regularisation techniques
• Managing Over-fitting
• Validation methods
• Support Vector Machines
• Kriging (Gaussian Process regression)
• PCA and Kernel PCA
• Self Organisation Maps (SOM)
• Kernel induced vector space
  • Mercer Kernels and Kernel-induced similarity metrics
• Reinforcement Learning

DAY 3 - DEEP LEARNING

Content will be taught in relation to the topics covered on Day 1 and Day 2

• Logistic and Softmax Regression
• Sparse Autoencoders
• Vectorization, PCA, and Whitening
• Self-Taught Learning
• Deep Networks
• Linear Decoders
• Convolution and Pooling
• Sparse Coding
• Independent Component Analysis
• Canonical Correlation Analysis
• Demos and Applications