COMP 6001 Neuromorphic Algorithms and Computation
Credit Points 10
Legacy Code 800232
Coordinator Saeed Afshar Opens in new window
Description Designing and implementing processing pipelines for event-based sensory data is a crucial skill for neuromorphic engineers to test novel hardware platforms or to develop new algorithms and learning mechanisms. This project-based subject focuses on principles of neuromorphic algorithm design and hardware-friendly neural architecture design for neuromorphic information processors. This subject consists of two streams of research: applied event-based algorithms and bio-inspired spiking networks. Through solving increasingly challenging tasks using distributed, event-based competitive processing elements, students will learn the differences between conventional and neuromorphic algorithm design, critically assessing real-world problems in a structured manner.
School Graduate Research School
Discipline Algorithms
Student Contribution Band HECS Band 2 10cp
Check your fees via the Fees page.
Level Postgraduate Coursework Level 6 subject
Restrictions
Must be enrolled in 8124 Master of Applied Neuromorphic Engineering
Learning Outcomes
On successful completion of this subject, students should be able to:
- Critically evaluate the advantages and disadvantages of event-based data processing in comparison to Conventional Frame-based data
- Assess the fundamental building blocks of neural computation in biology and Neuromorphic Systems
- Design and evaluate event-based algorithms on standard von Neumann architectures
- Propose novel neuromorphic processing methods relevant to distributed neuromorphic processors
- Develop a solution-oriented way of critically assessing real-world problems using Neuromorphic algorithms
- Effectively communicate the significance and impact of a specific Neuromorphic system to an audience consisting of both specialist and non-specialists
Subject Content
- Encoding and Processing Conventional and Event-based data
- Architectures of Neural Computation
- Spiking Neural Networks in Biology, Software Simulation and Neuromorphic Hardware
- Event-based Classification
- Event-based Tracking
- Event-based Feature Extraction
- Designing a Novel Event-based Algorithm
Assessment
The following table summarises the standard assessment tasks for this subject. Please note this is a guide only. Assessment tasks are regularly updated, where there is a difference your Learning Guide takes precedence.
| Type | Length | Percent | Threshold | Individual/Group Task |
|---|---|---|---|---|
| Practical | Maximum 1000 lines of code | 30 | N | Individual |
| Practical | Maximum 1000 lines of code | 30 | N | Individual |
| Applied Project | 2 weeks | 20 | N | Individual |
| Applied Project | 2 weeks | 20 | N | Individual |
Teaching Periods
Spring (2023)
Parramatta City - Macquarie St
On-site
Subject Contact Saeed Afshar Opens in new window
View timetable Opens in new window
Spring (2024)
Parramatta City - Macquarie St
On-site
Subject Contact Saeed Afshar Opens in new window