Introduction
Introduction
Machines that think as humans
We train the machine as we train a human.
Its like a student, we give him a class and we give him a test. If a lot of students pass the test, the model of teaching is good. If not, we need to change the model of the class.
The machine is good cause its can take actions without emotions, but it can be bias if the data is bias and the machine is taught with this data.
Types of AI
- General AI: It's a machine that can do anything that a human can do.
- Narrow AI: It's a machine that can do a specific task.
- Machine Learning: It's a machine that can learn from data.
Turing Test
It's a test to see if a machine can think as a human. The people talk with a machine and a human and try to guess who is the machine.
The relation between AI and ML
We can obtain or not a Narrow AI with ML. We use ML to train a machine to do a specific task. If the machine can do the task with a good accuracy, we can say that we have a Narrow AI.
ML
ML is the method to teach a machine to do a specific task. We use data to train the machine.
Narrow AI
Reinforcement learning
We have a Agent and a Environment. The agent take actions in the environment and the environment give a reward to the agent. The agent try to maximize the reward.
Neural Networks
The most used method in ML to build Narrow AI. The neural networks are inspired by the human brain.
Deep Learning
Deep learning is a neural network with a lot of layers (neural).