CS191 2019-02-19
Zhu, Justin

CS191 2019-02-19
Tue, Feb 19, 2019

Gefter Context Paper

Pitts grew up in a rather poor area with an abusive father. McCulloch meanwhile grew up in a well-to-do family with ample resources and a well-rounded education. They are the Prince and the Pauper. Their share connection is the early exposure to Russell’s Principia Mathematica, which proved to have formidable effect on their conception of neuroscience using the language of logic.

In great discoveries, we often see a pairing of individual inventors. For example, Pitts and McCulloch, Leibniz and Newton. What is significant about duo discovery? Are there greater truths to what lies in the horizon?

McCulloch & Pitts Paper

McCulloch and Pitts formed an extraordinary and tragic partnership. Apparently McCulloch and Pitts had a fallout after the publication of the paper. Pitts became insane and died an early, tragic death. It’s worth pointing out that McCullouch was 25 years older than Pitts. If anything, that really illustrated the extremely intense brilliance Pitts possesses at his age and given his upbringing.

Logic is all about “all-or-none” character, leaving out the ambiguity. For any logical expression satisfying certain conditions, there is a net that behaves in a certain fashion.

There are rules that govern this all-or-none relationship characterizing neurons, one of which being that a certain fixed number of synapses must be excercised with the period of latent addition in order to excite a neuron at any time, and number is independent of previous activity and position on the neuron.

This latency is a key factor because it is what defines “synaptic delay” that is the activity of any inhibitory synapse absolutely preventing excitation of that neuron at that time

This latency is a key factor because it is what defines “synaptic delay” that is the activity of any inhibitory synapse absolutely preventing excitation of that neuron at that time.

Moreover, the sturcture of the net does not change with time, illustrating a constancy of the neuron that is not sensitive to wild, rampant change.

Here, we see the very origins of the “neural network”, where Pitts and McCulloch systemized $N$ being the neurons of N with no axons synapsing upon them. Moreover, we let $N_1$ to $N_p$ denote the actions of such neurons.

We say that it is readable in the narrow sense if there exists a net N and a series of $N-i$ such that there exists a realizable nature in the extended sense.

Moreover, we are also formalizing the notion of cyclic if there is a circular chain $c_i$,… $c_p$.

These circles, in graph theory, really just are shortest connected components, SCC, which create alterable synapses. Meanwhile, nets with circles are usually singular systems.

Rosenblatt Paper

In Rosenblatt’s paper, we are introduced to the perceptron, which is a probabilistic model for information storage and information recollection.

Whatever information is retained must be stored as a preference for a particular response, contained in connections or associations rather than topographic representations.

The neurons that most similar will form pathways to the same sets of responding cells. Those that are dissimilar will develop connections to different responding cells.

Positive and/or negative reinforcement is chiefly achieved by the successive saturation of stimuli.

George Miller, mathematical psychologist.

People were programming things to see if we could measure reaction time. Getting interested in the substance. One of these people told Harry Lewis the location of Warren McCulloch.

“Piercing eyes of this guy”. Looks like the ancient mariner. That was what McCulloch was like.

Pitts had decided in this paper that the brain operated like a computer, a logical system. In this sense, you would completely understand how the brain would work. Pitts thought this was the finite state machine. Nothing that happened in the past could have affected it. The paper said the problem was completely solved.

Neural computing is pretty important. The automaton theory in Turing 36. Pitts describes logical formula. Church is the one who takes away the abstraction away from Pitts and McCulloch

You can do something completely wrong but as long as you work hard to achieve that idea, its importance will appear. There were no fancy electronics at the time and such.