Neural networks is the most common type of model for modern AI work. ChatGPT runs on a specific type of neural network called a large language model(LLM)
They can be compared to brains as they strengthen and weaken connections between nodes given the data it receives, just like the brain modifies the connections between neurons based on the experiences you have.
For more detail read here
Below shows how the game aligns with how it works in reality.
- You roll dice to determine if a model was successful, because no matter how much you train it there is still a chance it is wrong. It is a probablity machine, not guarantee list of instructions like traditional programing.
- Input Layer = An animal photo
- Hidden Layer = Where you add and strengthen nodes
- Output Layer = Path that connects an animal photo which is needed for a successful test
![Actual Neural Network Image](assets/actual-neural-network.png)
![Data Brain Image](assets/data-brain-board.png)