After building the philosophical foundation behind a concept, the next step in creating an Alchemy Symbol was to gain Enlightenment through its central philosophy and use that understanding to draw out the Symbol.
Ezra continued focusing his mind.
He went deeper into his thoughts, considering what the Learning Formula Symbol should look like as an Alchemy Symbol.
There was something important he needed to keep in mind.
He already knew that Process Symbols were the foundation upon which Advanced Formula Symbols were developed.
The latter were essentially upgraded versions of the former.
It was that...
Unlike Material Symbols and Fundamental Symbols, Process Symbols and Advanced Formula Symbols involve a series of Symbols working together to derive an outcome.
Just like an equation.
A single Symbol could represent a particular instruction or property, but a process required several Symbols to work together.
Each had its own function, and the arrangement of those Symbols determined the final result.
That was also why an Alchemy Circle containing Process Symbols or Advanced Formula Symbols tended to have a high number of Alchemy Symbols.
The number wasn’t simply for appearance.
Each Symbol contributed to the overall operation, and the Alchemy Lines connected them into a structure that allowed the entire Circle to function as intended.
That was also why Symbols in those categories could only be used and formed by Alchemists whose Alchemist ranks provided them with the necessary Cognis reserves to construct all the required Alchemy Symbols.
The more complex the arrangement, the greater the demand on the Alchemist.
Now...
Unlike Process Symbols, which tended to use more Symbols to drive a particular process, Advanced Formula Symbols used fewer.
However, their Alchemy Symbol arrangements were much more complex.
Another point was that some of the types of Symbols listed within those categories actually had Fundamental Symbols mixed into them.
So...
If someone wanted to create an Advanced Formula Symbol, it would basically be like writing an entire equation.
And Ezra knew all of that.
If I want to get this right, I have to view the philosophies surrounding the Learning Formula as an equation.
Like...
Information + Knowledge + Experience + Change + Development = AI Brain.
That’s what I’ll call it for now.
Ezra thought to himself as he processed his thoughts calmly and analytically.
He wasn’t trying to create the final structure of the Symbol just yet.
First, he needed to understand how the different elements of the Learning Formula could work together.
The philosophical meaning gave him the foundation, but he still needed to determine how that foundation could be expressed through an Advanced Formula Symbol.
An equation was the most suitable way to approach it.
After all, he was trying to create something that could process information, retain knowledge, recognize patterns, and make changes based on what it had learnt.
That was more than a simple process.
At the same time...
Ezra activated his Book of Memories within his Mind Realm.
The familiar book appeared in his mental space.
It contained the experiences of his past self, along with information about the things his Past-self had seen, heard, and studied during his life.
Ezra opened the book.
The pages turned until he reached the section containing information about Machine Learning.
He had already read through this material when he first began researching Artificial Intelligence, particularly when he started looking into Nanotechnology.
Among the many fields of study Ezra had explored after deciding to bring AI into his new world, Machine Learning was one of the subjects he had wanted to understand better.
A few studies from the Intermediate Biology Academy courses had also opened his eyes.
They made him realize that there were subjects in his world that he could study to gain considerably more knowledge than the information already available in his Book of Memories.
His past-self had been a soldier and a weapons engineer.
So, he didn’t possess deep knowledge of AI or the studies behind it.
However, that didn’t mean the Book of Memories lacked information about those subjects.
It simply meant that there were gaps in the understanding he would gain from the Book.
Like...
Neuroscience.
Artificial Neural Networks.
And Machine Learning.
Among the three, the studies Ezra had focused on most during the beginning of his research were Neuroscience and Artificial Neural Networks.
They were important to what he wanted to achieve.
He had first needed to understand how the human brain processed information and how an artificial system could imitate certain aspects of that process.
But now...
He was revisiting the study of Machine Learning.
The information in his Book of Memories came into focus.
***
╔════════════════════════════════════╗
[Machine Learning]
Entry Recorded — Book of Memories
╚════════════════════════════════════╝
Classification: Artificial Intelligence and Computational Science
Primary Focus: Learning algorithms, pattern recognition, and adaptive information processing.
Fundamental Principle
Machine Learning is a branch of Artificial Intelligence concerned with developing systems capable of improving their performance through experience and acquired information, rather than relying entirely on manually written instructions.
Instead of programming every possible response, a learning system processes information, identifies patterns, evaluates its results, and modifies its internal parameters to improve future performance.
Its fundamental principle is the ability to learn from data and use acquired knowledge when encountering new situations.
Neural Network Parameters
Artificial learning systems commonly use interconnected mathematical units to process information.
Each connection possesses a Weight, which determines the influence of incoming information, while a Bias adjusts the unit’s response.
Together, these parameters determine how information is processed and what output is produced.
A simplified artificial neuron works by multiplying incoming information by its respective connection weights, adding the Bias, and then passing the result through an Activation Function to produce an output.
Learning Algorithms
Learning algorithms determine how an artificial system acquires information and modifies its internal parameters.
Supervised Learning uses labelled examples, allowing a system to compare its predictions against known results.
Unsupervised Learning identifies patterns and relationships within information without predefined labels.
Reinforcement Learning involves learning through interaction, feedback, rewards, and penalties, allowing a system to improve its decisions through repeated experience.
Error Evaluation and Correction
A learning system requires a method of evaluating its performance.
A Loss Function measures the difference between its predicted output and the expected result.
One commonly used method is Mean Squared Error (MSE). It calculates the average of the squared differences between the expected results and the system’s predictions.
The calculated error can then be used to adjust the system’s internal parameters.
Gradient Descent is an optimization method that modifies connection weights in an attempt to reduce the error.
It works by calculating how much each connection weight contributes to the error, then adjusting those weights in the direction that reduces it.
The amount of adjustment is determined by the Learning Rate, which controls how large each adjustment should be.
In simplified terms:
New Weight = Previous Weight ? Adjustment
The adjustment is calculated using the error information and the Learning Rate.
Backpropagation calculates how much each parameter contributes to the overall error, allowing the system to determine which connections require adjustment.
Memory and Adaptation
Learning systems can preserve acquired information through internal parameters, stored data, or dedicated memory structures.
When new information is introduced, the system can compare it with existing knowledge, identify patterns, and modify its parameters according to its learning algorithm.
However, changing parameters does not automatically guarantee genuine understanding. The system’s ability to adapt depends on its architecture, available information, and learning mechanisms.
Limitations
Machine Learning systems depend on the quality of their training information, processing resources, mathematical architecture, and learning algorithms.
Insufficient or inaccurate information can produce unreliable results, while excessive adjustments may cause a system to lose previously acquired knowledge, a problem commonly associated with catastrophic forgetting.
Furthermore, a system that improves through training does not necessarily possess independent reasoning or consciousness.
***
Ezra read through the information in his Mind Realm.
The pages remained open before him.
He went through the different sections, paying attention to the connections between them.
As written earlier...
The information to Ezra wasn’t entirely new.
Since he had already studied it during his previous research, but revisiting it now gave him a different perspective.
Back then, he had been trying to understand how Machine Learning worked and how it could be applied to the development of Artificial Intelligence.
Now, he had another purpose.
He wanted to use that knowledge to construct the Learning Formula Symbol.
Seeing this information again made it easier for Ezra to develop a better understanding of how exactly he wanted the equation behind the Learning Formula Symbol to look.
But it was more than that.
This AI study is almost connected to the way the human brain works.
How we see, store, and process information.
Ezra thought to himself as he began piecing things together.
He looked at the information on Machine Learning again.
The first section explained how a learning system acquired information and improved its performance through experience.
The next described the mathematical units and parameters used to process incoming information.
Then came the learning algorithms, followed by the methods used to evaluate errors and adjust the system.
Finally, there were the sections on memory, adaptation, and the limitations of Machine Learning.
Ezra considered the arrangement.
It wasn’t simply about receiving information.
A learning system needed to process what it received, compare it with what it already knew, evaluate its results, and make adjustments.
The more he examined the information, the more he could see how it connected to the six principles he had established for the Learning Formula.
He remembered the very key words that the Adult Top Alchemists usually stated about Alchemists.
An Alchemist’s view of the World should be that of an Equation.
Ezra recalled those words.
And since Learning dealt with information gathering, storing, and processing, the Formula Symbols should also hold a connection to how he could view natural information gathering.
He returned his attention to the equation he had formed earlier.
Information + Knowledge + Experience + Change + Development = AI Brain.
The equation was still a simplified representation.
He knew that.
But it gave him a direction to follow.
The information gathered by a system had to become something it could retain and use.
Its stored knowledge had to be connected with new information, and the results of those connections had to influence how it responded in the future.
That was where the other principles became important.
Adaptation.
Self-Modification.
Application of Knowledge.
They weren’t simply additional functions placed beside the first three principles.
They were part of the same learning process.
Ezra continued examining the information.
This was also why he wanted the Nanites to possess Artificial Intelligence.
Utilizing them to function on their own would greatly increase their ability to multitask.
Instead of having to give every individual Nanite a direct command whenever he needed something done, an artificial learning system could allow them to process information and perform their assigned functions independently.
That was the purpose behind his idea.
Of course...
If he used the Alchemical method behind Illusionary Alchemy, he could utilize his own consciousness to give commands to the Nanites.
But that would be stupidly risky.
He understood the danger.
If he attempted to use his consciousness in that manner, he could end up dividing it.
And there was also the question of what would happen if one of the Nanites was destroyed.
The consequences weren’t something he wanted to gamble with.
Only when he reached the Alchemist General or Alchemist Scholar rank could he truly utilize something like that.
But why even use that method when AI could solve all of this?
Ezra’s mind remained in a state of information processing.
The different pieces of knowledge were gradually coming together.
The web-like structure remained in one frame within his mind.
It contained the six principles and their philosophies, with their connections leading towards the central meaning he had discovered.
In another frame was the information on Machine Learning inside his opened Book of Memories.
He could examine both at the same time.
One represented the philosophical foundation of the Learning Formula.
The other contained the knowledge of artificial learning systems from his previous world.
They were different forms of information, but Ezra could now see where they connected.
The six principles described what a learning entity needed to accomplish.
The Machine Learning study showed him how an artificial system could process information, recognize patterns, evaluate its performance, and modify its internal parameters.
He wasn’t simply trying to copy the technology of his previous world.
He needed to understand the principles behind it and find a way to express them through Alchemy.
And finally...
With all the studies mixing together, the Learning Formula Symbol was beginning to come together in Ezra’s mind.
The web-like structure and the information from the Book of Memories no longer felt like two separate pieces of knowledge.
They were becoming parts of the same idea.
The sensation of creation hit him, as well as the flow of Cognis generating from his brain increased slightly.