AI / AGI
Advanced field of computer science creating machines capable of performing tasks commonly associated with human intelligence, i.e. learning, understanding language, reasoning, solving problems.
Narrow (or weak) AI is good at performing one particular task and is the only form of AI that is achieved so far. Narrow AI examples: self-driving car, weather forecasts, sales predictions, virtual assistants.
General (or strong) AI is the type of AI that can understand and reason its environment as a human would. For now it's hypothetical intelligence and can be found only in science fiction.
(Supervised, Reinforcement, Unsupervised Learning)
Machine learning is a field of study that gives computers the ability to learn from data, identify patterns, and make predictions or decisions without being explicitly programmed to do so.
There are different types of learning: supervised, unsupervised, reinforcement.
The majority of practical machine learning uses supervised learning. In supervised learning, the dataset is the collection of labeled examples, i.e. pictures with object name in it ("dog", "cat", "tiger"). Since we know the correct answers, the algorithm iteratively makes predictions on the training data and is corrected. Learning stops when the algorithm achieves an acceptable level of performance. Examples: Spam detection, disease detection, market price prediction, forecasting.
In unsupervised learning, algorithms are left to their own devices to discover and present an interesting structure in the data. Examples: discover that people that buy A also tend to buy B, group customers by purchasing behavior.
Reinforcement Learning solves a particular kind of problem where decision making is sequential, and the goal is long-term, such as game playing, robotics, resource management, or logistics. These algorithms are penalized when they make the wrong decisions and rewarded when they make the right ones – this is reinforcement.
Artificial Neural Networks
Artificial neural networks (ANN) are one of the main tools used in machine learning. As with the brain, neural networks are made of building blocks called “neurons” that are connected in various ways. Using algorithms, they can recognize hidden patterns and correlations in raw data, cluster and classify it, and – over time – continuously learn and improve.
ANN application examples: Google's translation service, Amazon's Alexa AI-powered assistant, diagnosis of skin and breast cancer, "eyes" for self-driving cars. ANN is composed of an input layer, which receives data from outside sources (data files, images, sensors, microphones…), one or more hidden layers that process the data, and an output layer that depends on the function of the network.
Artificial neural networks (ANN) which consist of more than three layers of neurons (including the input and output layer) are called Deep Neural Networks. And training them is called Deep Learning.
Most AI examples that you hear about today – from chess-playing computers to self-driving cars – rely heavily on deep learning.
Conversational AI, Chatbots, Virtual Assistants, NLP/U
Conversational AI allows us to interact naturally with computers simply by having a conversation. We use conversational AI when our virtual assistant wakes us up in the morning, when asking for directions on our commute, or when communicating with a chatbot while shopping online.
Chatbots are used instead of providing direct contact with a live human being. Many banks, insurers, media companies, e-commerce companies, airlines, and restaurant chains are using chatbots to answer simple questions, increase customer engagement, for promotion, and to offer additional ways to order from them.
\When we talk about a personal agent helping us with commands and questions, we refer to virtual assistants such as Apple Siri, Google Assistant, Amazon Alexa.
The conversational AI is still a developing area, heavily related to artificial intelligence, machine learning, and natural language processing (understanding), NLP/U. NLP/U takes text as input, tries to understand context and intent, and generates an intelligent response.
AI raises many ethical concerns about privacy, discrimination, unemployment, transparency, and many more. Consider as an example the use of facial recognition in public places, which may help to reduce crime, but in ways that entail surveillance activities affecting individual liberty and privacy.
AI Governance / Regulation
AI is improving many areas of life but at the same time entails a number of potential risks, such as opaque decision-making, gender-based or other kinds of discrimination, intrusion in our private lives, or being used for criminal purposes.
The government has an important role to play in clarifying expectations about AI's application, determining frameworks within which AI is developed and deployed, setting regulations, and legal codes.
The General Data Protection Regulation (GDPR) is a regulation in EU law on data protection and privacy in the European Union (EU) and the European Economic Area (EEA).
The GDPR aims primarily to give control to individuals over their personal data. The GDPR defines what data can be processed and under which conditions; for how long can data be kept; what information must be given to individuals whose data is collected.
Blackbox / Explainability / Transparency
When we use AI to make critical decisions such as loan applications, who gets paroled, and who gets hired, we want to understand how do those algorithms work, why do they make the recommendations they do. Sometimes it is difficult to reach such transparency. For example, algorithms based on deep learning are often called "black box" because the reasons for their decisions are not easily accessed or understood by humans, even though they can reach a better performance than other algorithms.