Ranging Far and Wide NLP and Deep Learning

It is Sunday morning which means time to let my thoughts range far and wide, where ever they will go will lead me.

I am plowing through this in order to come up to speed the current state of Natural Language Processing.  Plowing is hardly the term.  For nerds only.

Deep Learning For NLP (Without Magic) Part 1  -  Stanford university presentation.

I hit equations 25 minutes in and digressed to google for an explanation of algebra.  At 30 minutes I began to see the forest for the trees.  The light dawned when I put it all in the perspective of an object oriented conceptual structure.  The "bingo moment" is when he explained words as atomic things.

"The vast majority of rule based and statistical NLP work regards words as atomic symbols."....."called a "one hot" representation.

For example:

Motel [00000000000100000]
Hotel  [00000000010000000]
Motel and Hotel = 0

I can get it because Hotel and Motel are reduced to their unique exclusive  1 to 1 mandatory relationship with a representative  "atomic symbol" that a computer can read at the most elemental level and recognize some relationship implementation value.  That is how NLP is structured from the ground up.  All words have a unique atomic symbol in a neural network.  I might have a 20,000 word symbols represented uniquely in my head (somehow) that relate (somehow) to other words.  Making sense is a function of object class relationships.

State of the art in NLP has a world of implications about where we are, who is doing what with it and where it is going!!!!!

Coulda, woulda, shoulda was often a factor in a recent month in Europe although the entire journey was excellent.

If I coulda have grasped algebra in terms of object oriented system structure in grade school my entire life would have been different.

Shoulda gone to a Jesuit school and studied this. or gone here.

Is there such a thing as an open source standards dictionary that assigns an "atomic value" to every word in our dictionary?  Of course, that would be at the computer binary machine language level just like there are numeric binary representations.  I infer that there probably is such a thing.

While the word itself translates to some binary machine language representation --which of course makes this blog possible.  Does binary representation in the form of if the example only serve for the sake of an example or give some some dimensional relationship information at the syntactic level before going to some stand alone system producing semantic level meaning??.....just wondering.

Note to self:  got to 53 minutes.  Enough for this nice Sunday morning!  Pick up from there later.