Artificial Intelligence (AI) is the overarching concept that covers anything relating to making machines smart or at least seem smart. Automated Speech Recognition (ASR) and natural language processing are examples of AI that make interacting with computers more natural.
Machine Learning (ML) is a sub-set of AI and is typically defined as systems that teach themselves and are designed to get smarter over time the more they are trained. Machine learning requires large datasets and training criteria to learn how to identify patterns and behaviours, which can be used in calculate predictions.
Neural Network or Artificial Neural Networks (ANNs) are computing systems that are loosely modelled after the neurons of a biological brain.
Some ML systems use Neural Networks to learn patterns by providing known inputs and results to teach the system how to identify data. Image classification is an example of this where the ML trained model can automatically determine if an image contains a certain type of object. Or object identification that can identify specific objects in an image.
AI and Machine Learning systems require training to create a model, a definition of how to perform & understand a task and interpret data.
You can use sample datasets with known values to train to AI or machine model. You can help the system accurately predict the results and train it to correctly identify the data. The more training that is given the more accurate the results are.
Once your model is trained you can validate the model and test it against new values. The you can use the trained model in your application or systems.
Pre-trained models are available that have been created for a variety of purposes including object identification, image classification and speed recognition. These can then be integrated into your application without the need to train it yourselves.