Swedish web site
Neuromorphic Computing2026HT, 4.0 credits
|
|
Course plan
No of lectures
Ca. 3 introductory lectures, and research paper presentations by participants.
Recommended for
PhD students with an interest in neural networks, advanced computing hardware and software techniques, energy-efficient computing, and parallel and distributed computing. The course is mandatory for new PhD students in the ELLIIT F project "ScEENeC" starting in 2026, which organizes the course.
The course was last given
This is a new course.
Goals
Understanding the fundamentals, opportunities, challenges and existing hardware and software techniques for neuromorphic computing.
Prerequisites
A general background in computer science and engineering (about master level) is expected. Some familiarity with artificial neural networks and with parallel/distributed computing is useful.
Organization
Hybrid (zoom) multi-site PhD course format with guest lectures and remote
participants from BTH and Lund University.
Ca. 3 introductory lectures with mandatory attendance.
Research paper presentation and opposition by each participant.
Written or oral exam. All moments must be passed in order to pass the course.
Content
Energy consumption is one of the main challenges of today's AI systems. Most AI
systems today are designed using various forms of neural networks and deep
learning. However, training and even inference on conventional digital
hardware, even with GPUs or NPUs, are very costly in terms of time,
computational demands, and energy consumption. One promising alternative is
spiking neural networks (SNNs) executed on neuromorphic hardware. Neuromorphic
computing tries to mimic how the brain works by relying on changes in signals,
i.e., "spikes", rather than continuously recalculating numerical values as in
traditional artificial neural networks, offering a promising and more
energy-efficient alternative to traditional neural network-based systems for
machine learning. First hardware realizations of neuromorphic hardware are
available (e.g. Intel’s Loihi2 or SpiNNcloud’s SpiNNaker2), as well as software
support. However some technical challenges remain before neuromorphic computing
can widely replace ANNs and conventional digital hardware for machine learning
workloads.
This course will give a general introduction to fundamental concepts of
neuromorphic computing, including SNNs, processing in memory (PIM),
neuromorphic hardware and software, and application domains. We will discuss
existing solutions and remaining challenges, and read selected recent research
papers in neuromorphic computing.
Literature
Selected survey articles about Neuromorphic Computing, and research papers for student presentations, to be announced on the course web page.
Lecturers
Christoph Kessler (LiU), Håkan Grahn (BTH), Flavius Gruian (Lund), possibly also invited guest lecturers.
Examination
Presentation and opposition of a research paper, 1.5hp. Written or oral exam, 2.5hp. Both examination moments must be passed in order to pass the course and receive any credits.
Examiners
Christoph Kessler (LiU), Håkan Grahn (BTH), Flavius Gruian (Lund)
Credits
4 hp. It is not possible to take out partial credits, all examination moments must be passed.
Comments
Likely to be given in early autumn 2026 (late August - October).
Page responsible: Anne Moe
