Knowledge Gaps And Chat Bots

January 10th 2019 – By: William Larsen – Civilians News – “News For All Views” 

– Knowledge Gaps And Chat Bots –

I’ve almost completed my first Ai software, a chat-bot with, “human like,” personality traits.

Furthermore, in doing so… I’ve come to find a number of things fulfilling, in terms of the science…. behind Artificial Intelligence.

Notably, how simple it can be.

Take, “chat-bots,” for example.

Chat-bots themselves… are really just simple forms of Ai. Now, take that concept 1 step further…. and one of the simplest designs for a chat-bot, is on the home page of ArtificialIntelligenceProject.com. That design takes inputs…. and then treats those inputs as, “information the machine has heard or learned,” which it then repeats back at a ratio of 1:1, mimicking human conversation.

That’s 1 design for a chat-bot, albeit a simplistic illustration of how Ai could work, quite literally in it’s simplest form. (But what if bugs actually think that way?)

Nevertheless, as I finish building a much more complex *2.0 model of that Ai software, I’ve come to realiz there are a great number of variances, as to how to build a chat-bot…

That is to say, there are varying complexities and types of chat-bots (be it primitive Ai or not) that one could build……

Mitsuku Chat-Bot, winner of the 2018 Loebner Prize (a Turing Test competition), is an Ai software that’s programmed to directly respond to questions, with a previously programmed response…… That is to say, the creator has gone over every possible question, or phrase, in which a person can input and assigned it to a set response, then Mitsuku’s programmer has refined those responses, over the course of 10-15 years of programming…..

That’s another way of programming a chat bot……..

Now, let’s examine another simple, yet slightly more complex means for going about this task……

Google, Microsoft….. companies like this…… probably started building their chat-bots and Ai software, in a more mathematically based Ai model…. (I think it’s fairly safe to assume).

For example; every inputted word…. is attributed to a numeral and those numerals are expressed in equations, giving you simple, mathematically based Ai.

Word 1…. + Word 2 + Word 3 =   X % / or the probability of Word (x), or Word (z), or Word (q).

In fact, Google has so much search data, that this method in itself, could create a very human like Ai (or chat-bot) software….simply based off of the algorithms and probabilities, of words people connect through searches everyday.

And that equation, could create a very powerful Ai software, based on one hinging equation……..

Now, take that concept and make it more complex, say subtract a word (or 2) and check the probabilities again…. then begin creating algorithms….. Would (word1) + (word2) have a greater probability of eliciting (word3) from the respondent if (wordX) was inserted into the equation or (wordY)?

In this way, you could create a very powerful Ai software, relatively quickly….. (however, also note, that that version of Ai….. would be very powerful but also non-human)…… Now, also take a moment to consider how Ai can grow in accuracy, or complexity, by “branching off of,” pillars within the software, or psychological equations, so to speak (in this case 1 hinging equation).

My personal designs are actually more complicated than that, but to some extent, using a formula like the one above in order to code Ai, is sensible, for some computer learning models…….. Also note, how similar chat-bots are to, “search engines.” Furthermore, I’ll explain how I’m coding my own Ai chat bot / software, after I finish it, later on this Summer, but that’s 2-3 ways of programming a chat bot / Ai.

These are some of the utmost primitive forms of Ai, but Ai none-the-less and semi-powerful forms of it……Nevertheless, I also find it fascinating to think of psychology through the lens of computer science, in this way. Furthermore, these are simple, surface level complexities of Ai…. whereas a deeper personality has; added equations, filters, algorithms and things of that nature, coded into it……

Which then leads me to question the, “knowledge gaps,” in our education systems.

Nevertheless, as the founder of an Ai startup myself (PDX Larsen LLC), in a field that’s just beginning to reach it’s relatively primitive peaks and valleys, with some what arbitrary science….. this field raises many questions concerning the business of science itself. Cognitive dissonance is an interesting topic as well.

Long story short, because of the arbitrary nature of psychology combined with computer science, which has led to early stages of Ai, I’ve personally decided to make my Ai designs, for my next Ai software and chat-bots, public information, right here on CiviliansNews.com (be the change).

So, let’s see what the future holds.

Design specs for my *Ai 2.0 software (still un-named) will be printed, right here on CiviliansNews.com, later this year. (Or in early 2020.)

-William Larsen, Civilians News “News For All Views.”