It is the third post of the series in AI vs. Humans. This post will look into the algorithm of Artificial Neural networks ANN and Evolutionary Computation. If you have missed the first two posts, I suggest you read them.
Artificial Neural Network and Evolutionary Computation – AI vs. Humans
- How the human brain interprets data – AI vs. Humans
- What makes an Artificial Intelligence Intelligent – AI vs. Humans
- Artificial Neural Network and Evolutionary Computation – AI vs. Humans
Let’s start with the algorithms straight away, be patient as it might be a little geeky. However, I won’t be including any mathematical explanation. It will be simple, plain English with examples.
Artificial Neural Network
The Artificial Neural Network (ANN) is a set of Artificial Neurons connected and interacting as a group. The interaction is based on several mathematical and sophisticated algorithms. It is an extension of a Neural Network, which is biological, meaning it relates to humans, animals, or any carbon-based organic creature.
Neural Network models in Artificial Intelligence are ANN. ANN implies more on maths, which in turn is used for information processing, which depends on how the Neurons are connected and what environmental factors make them respond.
The above figure explains how ANN works. The artificial neurons are connected like human neurons. The inspiration for making ANN was from biology. Every neuron could call specific functions or forward the request to underlying neurons with the information sent from the previous neuron. It makes the work of ANN simple. Looking into the figure, we can see how each neuron(a circle in the chart imitates one neuron) contacts the other, which in turn to others.
Input is given to neurons, which call others for underlying hidden functions and output the result.
Due to the availability of millions of neurons and high connectivity, it is possible for neurons to connect or call the required function in another way. It is called routing. If you are traveling on the road, and there is considerable traffic ahead, you detour to another method. Similarly, neurons choose to change the behavior and give the same result even if there is high information processing. It might become a little slow as it is not an optimal path.
Evolutionary Computation
Evolutionary Computation Algorithms involve things related to AI’s evolution, similar to human algorithms, which are written to make AI more human. It includes reproduction, recombination, mutation, the survival of the fittest, etc.
It might not be possible for a robot or an AI program to give birth biologically, but they can always make another. The real question is how they will evolve as humans evolve with time. For example, a program can build another program, depending on his ability to program and his efficiency in producing another code, “Program Creating Another Program.”
They will survive based on their efficiency and computational speed. We can imagine a program changing its code to work better or even producing another program as a child.
Learning can be another basis for their evolution. In Ai based Games, they change the strategy as the user changes his playing style. They take into consideration previous data and the algorithm to predict what’s the next move. Thus, making the program brighter as it meets more new conditions.
It is predicted that after AI and robots reach a stable state, they will have an equivalent place to a human. We will have small programs or even a humanoid with similar rights.
The next post will be the last post of this series; I will cover the Current and Future states of AI, including some good live examples and links.
