A program that can sense, analyse, act and adapt. (Un programme qui peut détecter, analyser, agir et s'adapter.)
Programming is breaking down a task into a series of steps that can be followed. (La programmation consiste à décomposer une tâche en une série d'étapes qui peuvent être suivies.)
Machine Learning is getting the computer to perform a task from a collection of examples. This is done because computers are good at identifying patterns in huge data sets. (L'apprentissage automatique consiste à faire exécuter une tâche par l'ordinateur à partir d'une collection d'exemples. Ceci est fait parce que les ordinateurs sont bons pour identifier les modèles dans les ensembles de données énormes.)
Deep learning is when a computer model is programmed to perform classification tasks directly from images, text, or sound. (L'apprentissage approfondi consiste à programmer un modèle informatique pour effectuer des tâches de classification directement à partir d'images, de textes ou de sons.)
It's objective is to collect more relevant data to train the AI model
The more data an AI model has, the more precise the model will be able to detect patterns and make predictions
Identify the problem (What do you need to solve? What is the desired outcome?)
Prepare the data (Which data is going to be the most relevant? Get it, clean it, check it, organise it)
Choose the algorithms (What does the model need to do? Predict something or predict a quantity? Should it combine objects together? Or find links between objects?)
Train the algorithms (What are the limits of acceptance and / or error?)
Choose a language (e.g. Scratch or Python)
Choose a Machine Learning Platform (e.g. IBM Cloud, Google Cloud Prediction API, Microsoft Azure, Ayasdi, TensorFlow)
spam filters (phishing, mass mailouts)
smart email categorisation (inbox, spam, school, home)
voice to text features (transcribe filmed meetings)
smart personal assistants, such as Siri, Cortana and Google Now, Alexa
smart reply and online customer support to mimic your personal writing style (Thanks a million! Speak soon! Be right back, Welcome! Let's do it! I'm in!)
process automation (repeat often-used tasks)
sales and business forecasting (making predictions)
security surveillance (protect and warn)
smart devices that adjust according to behaviour (detect your regular patterns and offer suggestions you are likely to be interested in)
automated insights, especially for data-driven industries (eg financial services or e-commerce)
smart searches and relevance features
personalisation as a service
suggest new products based on buying habits and what other people have bought
fraud detection and prevention for online transactions
dynamic price optimisation (prices adjust according to the user's purchase probability)
recommendations and content curation based on users' tastes
personalisation of news feeds
pattern and image recognition
language recognition - to digest unstructured data from customers and sales prospects
advertisement targeting and optimised, real-time bidding
defining customer segmentation (high earner/low earner, single, family etc)
social semantics and sentiment analysis (to how people feel and react to social issues)
automated web design (based on users' probable preferences and style)
predictive customer service
Email is now sorted into the following folders: Primary, Social, Promotions, Updates, Forums and Spam
Smart replies to mimic your personal writing style (Thanks a million! Speak soon! Be right back, Welcome! Let's do it! I'm in!)
Nudging reminders (old unread emails might be highlighted in a different colour)
Match people to jobs and identify who would get hired
Lens tools to identify objects in images and find similar ones
Chatbots identify keywords and respond with helpful information e.g. scheduling appointments, taking orders or dealing with billing queries
Mental Health information is sent to individuals who might be behaving in certain ways; local police is informed to provide support
Search keywords are suggested
Search results will be adjusted to the person searching i.e. the algorithm tries to determine the quality of content required by the user.
AI tracks your listening habits
AI suggests similar musical genres and artists
AI could suggest new discoveries, new releases, and old favorites based on your listening habits
AI also suggests types of music based on the day, time of the day and the weather
to remind users when to pay bills
to predict if a person withdraws more money than they have in the account
to suggest payee details when paying bills or making transfers
to prevent fraud you can set up alerts. AI can recognise your unusual spending patterns
Calculate traffic and construction
Find the quickest route
Shows where traffic is slower
Know when to expect an Uber driver or food delivery
For driving and food services, AI will estimate how much a customer is willing to pay
AI can determine if a customer is likely to have been drinking
AI can determine how fast a customer walks and how fast a customer types
AI flies planes mostly - (the pilot manually controls the plane for an average of 7 mins during a flight)
Artificial intelligence is used to determine which visual to store and use in movie trailers.
AI is used to provide text from spoken words
AI is used to determine which audiences will like certain movies
AI is used to scan raw footage and select the best shots before a human edits them to make a film
AI correctly guesses whether a film would make money most of the time based on the visuals in the film and knowing what an audience wants
crimes are relatively predictable; AI is able to sort through a massive volume of data to find patterns
Predicting future crime spots - hotspots are identified and more police allocated
AI gait analysis technology will determine if individuals are walking back and forward on the same spot (e.g. pickpockets) or suspicious changes in behaviour
determining if individuals are low, medium or high risk
AI uses smart city infrastructure to triangulate the location of a gunshot
AI fitted in cameras (and not in the cloud) can better scan for license plates on cars, run facial recognition to search for potential criminals or missing people, and automatically detect suspicious anomalies like unattended bags in crowded venues.
detect, track and investigate infections in an environment
control robotic arms, precise movement and magnetised vision
identify social anxiety or mental illness and offer a course or prescribe solutions
gathering of data through patient interviews and tests
processing data and analysing results
using multiple sources of data to come to an accurate diagnosis
determining an appropriate treatment method (often presenting options)
preparing and administering the chosen treatment method
patient monitoring
aftercare, and follow-up appointments
eliminate one-size fits all and target treatments, personalise therapies,and uniquely composed drugs