Open Source SNN

Transcript

1. Open Source SNN cataska @ AGIT #2
2. About Me 翥嘨ށ / ਎ৼᇁ / ᑀଝ蝈
4. What is SNN
5. Spiking Neural Network Spiking Neural Network
8. Neuron models • Hodgkin-Huxley • Leaky Integrate-and-Fire • Izhikevich
9. Learning Algorithms
10. Spike-time dependency plasticity
11. Why SNN?
12. Why SNN? • Dynamic
13. Why SNN? • Dynamic • Time sensitive
14. Why SNN? • Dynamic • Time sensitive • Mimics neuron in real-world
15. Open source SNN projects
20. PyNN
21. Why another one?
22. Why another one? • Eating our own Dogfood!
23. Why another one? • Eating our own Dogfood! • Reinventing the wheel!
24. Why another one? • Eating our own Dogfood! • Reinventing the wheel! • Common Lisp & Rust!
25. Programming Languages
26. Created by Mozilla
27. Why Rust?
28. Why Rust? • Performance
29. Why Rust? • Performance • Reliability
30. Why Rust? • Performance • Reliability • Productivity
31. Created by John McCarthy
32. Why Lisp?
33. Why Lisp? • Simple
34. Why Lisp? • Simple • Interactive
35. Why Lisp? • Simple • Interactive • Programmable language
36. WheatNNLeek
40. https://github.com/libgirlenterprise/ WheatNNLeek
41. wheatnnleek-mnist
42. Future plans
43. Future plans • More models and synapses
44. Future plans • More models and synapses • Support GPUs
45. Future plans • More models and synapses • Support GPUs • Dynamic module
48. Under the paving-stones, the beach! — Graffito, Paris, May 1968
50. Thank you