Conclusion
- You can learn categorisers with biologically plausible spiking
neurons and STDP.
- A lot of standard machine learning tricks work, like co-variance,
feature combination, and
Word2Vec.
- The layered topology and timing presentation mechanism is not
biologically plausible.
- This is not a very advanced form of machine learning. There is
a lot left to explore.
- The current results are sound, but not leading edge.
