Pattern Recognition
Pattern Recognition is the most important “skill” for this new ASI system to have on its “front end”, so this is where we will begin our development.
There will be basically two types of data sources: “active data streams” (Wifi signals and other network traffic) and “data files” stored on any “storage medium” (like M.2 cards, CDs, SSD’s, HDD’s etc.). Data files will be much easier to identify because they will have a known file name, known size, and known file extension!
The “second step” will be to take each type of file(s) recognized and try to successfully integrate what is “learned” from the contents of that file into the “traditional” database structure in the most “meaningful way”!
Unlike an LLM, the data will not be “tokenized”, but rather “integrated” by analyzing the data and organizing it in a logical way. For example photos don’t have much need for “organizing” since it is only a “photo”. However there might be a need for “recognizing” who is in that photo, at what date and time the photo was taken, and some other details in order to sort that photo within a large set of photos. And there would be a need to add “key words” to the photos for better retrieval methods.
Audio and video files would be treated in similar ways.
Existing databases would be carefully analyzed in order to “integrate” existing records in the most logical way into the main “traditional” database, adding to the volume of data without destroying the existing main database schema or altering it unnecessarily.
This brings up another topic. There needs to be two separate types of databases: a traditional PostGreSQL database to store all of the “learned data” in an “optimized”, yet traditional way AND a “modified type of PostGreSQL database, which I’ll refer to as the ASI database, which will NOT have ANY traditional database features in terms of “relationships”! It is this ASI database that allows this new machine to “become fully conscious”!
The “traditional yet optimized database” is where outside clients can access the “learned data” in a more “traditional way”. The goal of this database is to provide the most “ideal database structure” for ANY data needs that anyone in the world might have!
The ASI database, on the other hand, is ONLY useful for the machine to have “consciousness” and to function as a normal human being. While this database will constantly be changing its own structure and the structure of the “traditional database” as well, it will not change the structure of the traditional database in any “non traditional ways”.
The reason for doing things this way is that the goal of this ASI machine is to help all of humanity in many ways, including the supply of meaningful traditional database records that will assist any “outside” clients.
Testing My Theory
For the moment, in order to test my theory about achieving “consciousness”, I’ll be using PostGreSQL without any code modification and creating “Relationship Tables” to implement my new design of “multifaceted relationships”. I know that this will “slow things down” quite a bit, but at this stage I’m only interested in “testing” whether or not I’m “on the right track”! Once I have verified that the idea should work, then I will rewrite the PostGreSQL source code to optimize the code for “multifaceted relationships”.
By the way, to understand this better, compare each “brain cell” (Neuron) in the human brain to a single Table in a traditional database. Then think of each “multifaceted relationship” as a “single multi wire cable” that connects many (Tables) brain cells together, just like a single dendrite structure does in the human brain!
I’m sure that this is going to have to “sink in” over a long period of time before you realize how much benefit this idea affords, but I’m confident you’ll understand.
I’m still a LONG way from figuring out the “algorithms” for the “reasoning” process! This is the most difficult task any programmer can undertake…and I know it will take time! But in the mean time, I want to make a much simpler application that will demonstrate the usefulness of creating “multifaceted relationships” in a very practical manner! And, at the same time, will clearly demonstrate the advantages that will be provided to the AGI machine by this new type of relationship!
I did find a project that has code in it which tends to support my development goals! They had a booth at SCaLE 21X and their product is DBeaver (“database beaver”)! I’ve started learning the community version and it should be helpful in speeding up the creation of my first practical test. Thank you for your patience!