- Class 30
- Practice 30
- Independent work 120
Advanced machine learning
Lecturers and associates
- Assistant Professor PhD Hrvoje Jerković
- Lovro Sindičić, Instructor
- Assistant Professor PhD Vedrana Vidulin
The objectives of this module are to enable students to:
• Interpret the definition of a range of neural network models.
• Be able to derive and implement optimisation algorithms for these models
• Interpret neural implementations of attention mechanisms and sequence embedding models and how these modular components can be combined to build state-of-the-art NLP systems
• Be able to implement and evaluate common neural network models for language
• Interpret model selection process in order to describe a particular type of data
• Evaluate a learned model in practice
• Interpret the mathematics necessary for constructing novel machine learning solutions
• Be able to design and implement various machine learning algorithms in a range of real-world applications
Students learn to build and maintain machine learning models including deep learning models, today the most important machine learning method used in the world's most important production systems for various tasks. Through this module, students will acquire and implement basic deep learning techniques on examples from natural language processing such as machine translation, sentiment analysis, and recognition of named entities. Also, the module will handle deep and awarded learning.
It is important for students to take this module in order to enable students to deepen their understanding of mathematics and algorithms of deep neural architecture and deep learning, as well as acquire practical knowledge to implement deep learning. Students will acquire the skills of designing deep architecture in TensorFlow, as well as hand-made deep neural networks that can be implemented later in any programming language.
1. Skansi, S. (2018) Introduction to Deep Learning, Cham: Springer International Publishing,
2. Goodfellow, I., Bengio, Y., Courville, A. (2016) Deep Learning (Adaptive Computation and Machine Learning series), Cambridge: MIT Press, available at https://arxiv.org/abs/1609.08144
Minimal learning outcomes
- Define steps in most common basic depth learning algorithms
- Explain changes in information during passage through an artificial neuron.
- Explain impact of different components of deep neural architectures
- Describe steps in selected project based on deep learning architecture for specific business problem.
Preferred learning outcomes
- Explain how the basic depth learning algorithms work.
- Critically judge changes in information during passage through an artificial neuron.
- Evaluate the impact of different components of deep neural architectures.
- Apply the chosen deep learning architecture to the problems.