Neuromorphic Computing index
Course on non-conventional computation on brain-inspired architectures.
Lecturer: Johan Kwisthout.
Topics:
- Traditional computation (Von Neumann architecture)
- Brain-inspired computation
- Training and programming a Neuromorphic chip
- Learning paradigms for spiking networks (local/global)
- Neuroscience applications
- Formal analysis of computation
- Guest lecture on neuromorphic hardware design (Leila Bagheriye)
- Guest Lecture for novel information encoding (Johan Mentink)
- Hands-on experience with Nengo FPGA, Intel Loihi (not commercialy available!), potentially Spinnaker or BrainScales.
Goal of the course: make us think "out-of-the-box" w.r.t. what computation is.
Content
- Lecture 1: Neumann Computing Architecture
- Lecture 2: Encoding Information
- Lecture 3: Learning in Spiking NNs
- Lecture 4: Programming neural nets
- Lecture 5a: Analyzing SNNs
- Lecture 5b: NMC chips in neuroscience
- Lecture 6: Architectures.md
Guest lectures:
Tutorials:
Q&A:
Concepts
- 202009091419-bit_addition_carry_propagation.md
- 202009161133-nengo.md
- 202010071335-Nengo_on_FPGA.md
- 202010231520-STDP_learning_rule.md
Assessment
- Online open book exam about 7 theory lectures
- Hands-on individual assignments
- Week 1: Turing Tumble
- Week 3: Learning
- Week 4: Programming
- Week 5: Nengo
- Report on your activities
- Group project: implementation or replication study
- Try to model/replicate a reported effect from a recent scientific article
- Write a draft paper on your results
Other links
https://logic.ly/demo