Democratizing AI of the past, present and future - a quick swing at the AI ball
by Mountain Computers Inc., Publication Date: Sunday, March 20, 2022
View Count: 182, Keywords: AI, Past, Present, Future, Expectations, Democratization, Hashtags: #AI #Past #Present #Future #Expectations #Democratization
The article (ref 2 by Priya Dialani, "Democratize AI: Exploring Future of Technology" has some interesting thoughts that I would like to comment and explore briefly.
1. the "first organizations that ring a bell are typically the FAANGs — Facebook, Apple, Amazon, Netflix and Google" are actually rather new and young in comparison to those involved pre-2000s.
2. the "These organizations entered new business sectors, giving AI-led services
and products, frequently outcompeting the occupants. They were
remunerated with amazing development" statement is about a generation of companies with Internet pre-massive cloud data-centric pre-hybrid data collection pools who were vastly collecting data about people via skipping and overlooking privacy ethics that today are considered taboo to touch without fear of litigation.
3. the "To use the full power of AI and address these difficulties, the
technology needs to get open to a more extensive range of businesses, in
other words, tackle the Democratization of AI." statement is really about basic statistical math applied to large data sets, where questions are asked in the old way first, and analysis is applied to confirm or deny, versus, AI which says, with this data, what patterns do you see and can arrive at which will form the basis of our questions.
with those three items in the first few paragraphs of the article, one can summarize that AI and the hop skip and jump (hopscotch approach) can lead to machine learning (ML) and vice versa; when in fact neural networks and the such have long existed before we had huge amounts of pooled data on heavy big iron in the sky / data center aka cloud.
Dialani goes on about the advantages and hopeful wishes and expectations of those in the deep of AI versus those just tippy-toeing in the shallow end of the AI swimming pool. Those in the deep end are extensively in a huge opportunity cost growth analysis which is where infomatics and the deep dive of data analysis, by whomever, with some basic statistical knowledge with access to huge data schemas and data sets can contrive some semblance of AI out of it; for some value; benevolent or not.
In Dialani's conclusion; I would conclude his approach to what I call risk-reward analysis and gap analysis is simple yet his ending thought on embedding AI throughout the business is good. For me, apply AI where it makes sense and those involved in manufacturing and in-line quality assurance have been doing this for decades; especially since we moved from manual labor to automation across all things.
For me, this simple article reflects a long overdue simplification that it is not hard to get into AI or ML and it not be a full time job. It just needs to be in your business plan in a case by case basis wherever automation and optimization can be streamlined without massive overhead to maintain and manage.
overall, I liked Dialini's rough cut on subject that many over-complicate rather than just simplify and say, "hey, we have these processes that have lots of data points, and we need some input/output processing continuous improvement and feedback loops throughout to do all the things we need to make sure stays in controls and boundaries we have set and can easily adjust. sure we can add or ignore a few data points without much cost and let's make sure our efforts for each data point is worth it - otherwise - its data waste alike carbon waste"™ me, andy flagg.
The fact that the post came from an Green Software AI initiative angle also begs the question; over analysis at cost vs benefit risk is always a balancing act; especially if we need to modify and change without major tear down and reconstruction; and if so, fast, easy and within control.
my quote of the article and day, "statistics can be fun if applied appropriately and wisely used timely for good improvements. remember, do your homework, know your math, and apply it with an ethical human touch."
more to come...
if you found this article helpful, consider contributing $10, 20 an Andrew Jackson or so..to the author. more authors coming soon
FYI we use paypal or patreon, patreon has 3x the transaction fees, so we don't, not yet.
© 2022 myBlog™ v1.1 All rights reserved. We count views as reads, so let's not over think it.