Overview
Introduction to deep learning methods for computation with artificial neural networks and applications to image/natural language processing.
Artificial Intelligence has been revolutionised by a new generation of neural network based machine learning algorithms. This course will introduce these new algorithms, as they are applied in two important areas of AI: machine vision and natural language processing. The course will cover a range of new architectures and techniques, including deep learning and convolutional networks for vision, and sequence-to-sequence networks and transformers for language processing.
About this paper
| Paper title | Deep Learning |
|---|---|
| Subject | Computer Science |
| EFTS | 0.1667 |
| Points | 20 points |
| Teaching period | Semester 1 (On campus) |
| Domestic Tuition Fees ( NZD ) | $1,535.64 |
| International Tuition Fees | Tuition Fees for international students are elsewhere on this website. |
- Eligibility
There are no formal prerequisites for the 400-level papers, but prior knowledge is assumed.
- Contact
- Teaching staff
- Paper Structure
The paper introduces the neural networks in a lecture-driven format, with some elements of discussion based on prescribed reading of relevant literature. We will cover an introduction to the topic, the practical use and tuning of neural nets (using TensorFlow) and fundamental algorithms and architectures, including (a selection of):
- Multi-layer Perceptron networks
- Backpropagation
- Convolutional networks
- Language models
- Transformers
- Teaching Arrangements
One 2-hour lecture per week.
- Textbooks
Textbooks are not required for this paper.
- Course outline
- View the course outline for COSC 420
- Graduate Attributes Emphasised
- Interdisciplinary perspective, Lifelong learning, Communication, Critical thinking, Research.
View more information about Otago's graduate attributes. - Learning Outcomes
This paper will enable students to:
- Understand and implement a range of artificial neural networks
- Understand the strengths and weaknesses of neural networks compared to traditional symbolic methods
- Perform practical research with a neural network simulation and systematically present the outcomes