Artificial Intelligence is the creation of machines to mimic human capabilities, such as teaching a machine to see (recognise objects in an image) and listen (interpret and analyse sounds). In a traditional computer program, our algorithm tells the machine exactly what to do (step-by-step), whereas, in an AI program, the machine is programmed to learn and make its own decisions. Artificial Intelligence is a broad term that covers a range of specialisations and subsets, such as computer vision and natural language processing.

Examples of AI

• To see(Transport- autonomous cars, Health–Cancer treatments, equipments for disabled, Space- Photo of blackhole, Art- to make robot paint
• To Move- Robot arms, Drone helping in farming
• To communicate- to translate language, Virtual assistants, Debating
• To think –Patterns-For predictions and data- Netflix,You tube and gaming

AI Techniques

Supervised and unsupervised learning

• Supervised learningis the process of the human providing the program with many examples of what it is we are wanting it to learn, along with a label that helps the machine classify or identify an object. By providing many examples, potentially millions, the machine is learning about how to identify an object by seeing many examples of the features that represent it. This then enables the machine to make judgements on its own when it receives new data and has to identify something without a label.

  Unsupervised learning involves providing the machine with a large amount of data and letting it find patterns in the data on its own, by trying to identify patterns in the features included. The machine then determines its own set of categories or labels by grouping the data. This process is known as “clustering” – the grouping of similar objects into one category.

Classification, Clustering and Regression

Classification

is a supervised learning technique used to group data based on attributes or features. Humans can provide labels on the data (e.g. for images or text) that tell the machine about the attributes or features (e.g. colour, size, shape, measurements) in each data and how to group the data. The machine then matches future data based on the similarity of the new data to pre-defined groups. For example, sorting images of kangaroos and wombats based on their number of legs, whether they have a tail or no tail and the size of their ears.

In addition to classification, you may also come across some other supervised learning techniques such as regression and unsupervised learning techniques such as clustering within the AI context.

A label is a meaningful tag that can be attached to data from similar categories. Humans can provide labels on the data that tell the machine about the attributes or features in each data and how to group that data. This process of labeling data is also called as annotation. For example, ‘cats’ and ‘birds’ are meaningful labels attached to categorise images of cats and birds. When considering filtering spam emails, we can have labels such as ‘spam’ and ‘no spam’.

Features are the attributes that describe what is in the data (such as features found in images, audio, text). For example, in the images of cats and birds (below), we consider attributes that describe the appearance of cats and birds such as the number of legs, the presence (or not) of a tail, fur, features, ears, height and weight.

Areas of AI

AI & Vision

The field of AI that relates to human vision is called Computer vision. Computer vision is the ability for machines to recognise objects in images or videos. Computer vision aims to mimic human vision by teaching the machine using many examples of images either labelled or unlabelled, where it discovers patterns and learns to recognise objects on its own. Examples of computer vision include face tagging on social media photos, automatic recognition of number plates, and vision used by self-driving cars.

AI & Language (Text & Speech)

The field of AI that relates to human communication and language is called Natural Language Processing.

Natural Language Processing (NLP) is the ability for machines to interpret and analyse forms of human communication, such as text and speech. NLP aims to mimic human communication by teaching the machine to read, write, speak and listen by providing it with many examples of communication data – either labelled or unlabelled, where it discovers patterns and learns to classify data on its own. NLP is used in technologies such as Google auto-complete, virtual assistants, language translators, news recommendation services and more.

Activity- Option 1

Unpacking key AI concept(Labelling , features and how AI sees an image)