Around 2001, I started to concentrate on how groups of neurons work
together in temporal synchronization. The classical computer models of
the neural network seem to be a dead end, so I sort of followed
my intuition and committed to studying the temporal analysis
of neural signals. For example, in a synthetic circuit the synapse
of a particular neuron can occur only when a particular set of neurons
have previously fired. This synchronization depends on
when each neuron synapses relative to each other
in the circuit. This fact led me to study temporal sequences
and its correlation coefficients.
I've been greatly influenced by a few people two of whom I've
thought about for years. They are Satosi Watanabe and David Bohm.
I also admire Walter Freeman for his insights into neuroscience,
and Dennis Gabor's work on signal processing.
Within the past couple of years I've focused my work on designing
and writing software using temporal state machines in computer
algorithms like the Aho-Corasick program. These programs fall
into a software class of what I call "large breath, short depth
trees" used in internet search engines. These trees essentially
model the efficient neural pathways in our brain. But there are large
problems in constructing efficient methods of correlating symbols
between trees because of the exponential explosion in complexity
as these trees grow in number. I've been reconsidering some methods
of bitmap pattern matching I had used years ago in the PRISM software
to implement logical queries and reduce coding complexity. Standard
methods of constraint based programming also helps simplify
relationships between these inter-tree dependencies.
I hope you enjoy reading the pages on this site.