18. Artificial Intelligence

Important points

Extra

Definition

AI is split into 3 categories:

  • Narrow AI: when a machine has superior performance in 1 specific task.

  • General AI: A machine is similiar in it’s performance to a human in any task.

  • Strong AI: machine has superior performance in any task.

AI has 2 subsets: Machine learning and Deep learning.

Machine Learning

Algorithms are ‘trained’ and learn from their past experiences. The system makes predictions or decisions based on previous scenarios.

It offers fast and accurate outcomes and manages/analyse large amounts of data.

Examples: Spam detection, search engines.

Labelled and Unlabelled data:

Labelled data: data where we know the target answer and the data object is fully recognizable.

Unlabelled data: data where objects are undefined and need to be manually recognised.

Types of machine learning:

Supervised Learning

System which is able to predict future outcomes based on past data. It requires input and output values in the training process.

Procedure:

  • The system requires both an input and output to be given for training. The model uses labelled data, so the desired output is known.

  • Algorithms receive a set of inputs and correct outputs for the learning process. The model is them run using labelled data.

  • Results are compared with expected output and the model is run with unlabelled data.


Unsupervised Learning

System which is able to identify hidden patterns from input data, the system is not trained on the ‘right’ answer.

By making data more readable and organised; patterns, similarities and anomalies become evident. Algorithms evaluate the data to find any structures withing the data set. Example: Individuals with similiar purchasing behaviours are regarded as 1 unit.

Reinforcement Learning

System which is given no training, learns on the basis of ‘reward amd punishment’.

It helps to increase the efficiency of the system by making use of optimisation techniques.

Semi-supervised learning

System that interactively queries source data to reach the desired result. It uses both labelled and unlabelled data.

Small amounts of labelled data is used combined with large amounts of unlabelled data. Example: classification of web pages into different categories.


Deep Learning

Machines that think in a way similar to the human brain. They handle huge amounts of data using artificial neural networks.

It structures algorithms into layers (input, output and hidden layers), the hidden layers is where data from the input layer is processed and sent to the output layer. It identifies patterns, for example, face recognition.


Comparison between machine learning & deep learning:

Machine Learning

Deep Learning

Enables machines to make decisions based on past data.

Enables machines to make decisions using an artificial neural system.

Requires small amount of data to carry out the training.

Requires large amounts of data during training.

Most features in the data need to be identified in advance and then manually coded into the system.

Learns the features of data from the data itself and doesn’t need to be identified in advance.

Modular approach is taken to solve a problem.

Problem is solved from the beginning to end as a single entity.

Testing of the system takes a long time to carry out.

Testing the system takes less time.

There are clear rules explaining why each stage was made.

Reasoning behind the decisions may be difficult to understand as the system makes decisions on it’s own logic.