Artificial intelligence (AI) is intelligence exhibited by machines. This is a cluster of topics in AI or related to AI.
Types of AI
Artificial Intelligence (AI) is classified into types based on the degree to which an AI system can replicate or go beyond human capabilities. One classification system uses four types: reactive machines, limited memory machines, theory of mind and self-aware AI. Another classification divides AI into two divisions: Weak AI or Narrow AI and Strong AI or General AI or Artificial General Intelligence. Different branches of AI are referred to by the method used to achieve AI.
Main classifications
- Weak or narrow AI or artificial narrow intelligence
- Strong or general AI (AGI)
- Artificial superintelligence (ASI)
- Reactive machines: These do not have past memory and cannot use past information for future actions
- Limited memory machines: These can use past experiences to inform future decisions
- Theory of Mind: In humans it is the ability to infer other people’s thoughts, desires and beliefs in others and to understand that they may be different from your own. Theory of mind level AI would be able to interact socially with people
- Self-aware AI (hypothetical, not realized)
Other classifications
- Cognitive computing: A term favored by IBM for computing designed to augment human intelligence rather than replace it such as helping doctors diagnose disease, guiding financial managers in making recommendations and allowing lawyers to search caselaw more quickly.
Branches of AI
- Artificial emotional intelligence: Integrating emotional intelligence with AI is one path towards achieving Theory of Mind AI
- Machine learning
- Deep learning (branch of machine learning)
- Computer vision
- Natural language processing (NLP)
Machine learning
Machine learning is a technique for realizing AI and it is an application of AI where machines are given access to data from which they learn form themselves.
Machine learning tools
Tools, algorithms, libraries and interfaces for machine learning
Neural networks
Artificial neural network (ANN) processing devices can be algorithms or actual hardware that are loosely modeled after the neuronal structure of the mammalian cerebral cortex. Neural networks are used in the branch of machine learning called deep learning. The following are types of neural networks used in machine learning as well as topics associated with neural networks.
- Convolutional neural network
- Recurrent neural network
- Recursive neural network
- Multilayer perceptron
- Hybrid neural network
- Generative adversarial network
- Variational autoencoder
- Highway network
- Image Style Transfer
- Deep belief network (DBN)
- Deep neural network
- Deep residual network
- Dropout
- Exploding gradient problem
- Vanishing gradient problem
- Google Inception
- Long short-term memory network
- Gated recurrent unit
- Paired Opened-Ended Trailblazer (POET)
- Boston Dynamics
Deep learning frameworks
A Deep Learning Framework is an interface, library or a tool which allows users to build deep learning models more easily and quickly, without getting into the details of underlying algorithms. Libraries are useful for individuals who want to implement Deep Learning techniques but don’t have robust fluency in back-propagation, linear algebra or computer math. These libraries provide pre-written code for functions and modules that can be reused for deep learning training for different purposes.
- Tensorflow
- Seq2Seq (framework for Tensorflow)
- Caffe/ Caffe2
- Keras
- Lasagne
- Torch
- MXNet
- DL4J
- OpenNN
- PyTorch
- Sonnet
- deeplearn.js
- MIL WebDNN
- Theano
- Apache SINGA
- Dlib
- Microsoft Cognitive Toolkit
- Swift for Tensorflow
- Deep Q-Learning : Algorithm in deep reinforcement learning
- Deep voice 1: Trains deep neural networks to learn from large amounts of data and simple features
- Deep voice 2
- Deep voice 3
- Google AutoML
Reinforcement learning
Reinforcement learning is an area of machine learning focusing on how machines and software agents react in a specific context to maximize performance and achieve reward known as reinforcement signal. The following are algorithms, tools and research topics related to reinforcement learning.
- Reinforcement learning
- OpenAI Baselines
- Tensorflow Agents
- Unity ML Agents
- Nervana systems coach
- ELF
- Pycolab
- Geek.AI MAgent
- Q-Learning
- SARSA
- Temporal difference (TD)
Supervised learning
Supervised learning is a type of machine learning in which data is fully labelled and algorithms learn to approximate a mapping function well enough that they can accurately predict output variables given new input data. This section contains supervised learning techniques. For example, Gradient Descent is a technique to optimize neural networks in supervised machine learning. Gradient descent optimization algorithms are used to speed up the learning process of deep neural networks. Another example, Support Vector Machine (SVM), is a type of algorithm that is a discriminative classifier formally defined by a separating hyperplane used for regression and classification tasks.
- Gradient descent
- Stochastic gradient descent (SGD)
- Support Vector Machine (SVM)
- Momentum / SGD with momentum
- Nesterov momentum / Nesterov Accelerated Gradient (NAG)
- Rprop
- RMSprop
- Adagrad
- Adadelta
- Adam
Decision trees
A decision tree is a simple representation for classifying samples. Decision tree algorithms are used in supervised machine learning where data is continuously split according to a parameter.
- Classification and regression trees (CART)
- Iterative Dichotomiser 3
- Random forest
- Extra-Trees
- Recursive partitioning
- Bagging
Unsupervised learning
Unsupervised learning is a branch of machine learning that tries to make sense of data that has not been labeled, classified, or categorized by extracting features and patterns on its own. The following are methods used in unsupervised machine learning.
- GloVe
- Principal component analysis (PCA)
- Kernel PCA
- Sparse PCA
- Independent component analysis (ICA)
- Non-negative matrix factorization (NMF)
- Archetypal analysis (AA)
- Non-negative matrix factorization via archetypal analysis
- Singular value decomposition
- Nonlinear dimensionality reduction
- Independent component analysis (ICA)
- Google PageRank algorithm
Clustering
In unsupervised machine learning, clustering is the process of grouping similar entities together in order to find similarities in the data points and group similar data points together.
Ensemble learning
Ensemble methods are meta-algorithms that combine several machine learning techniques into one predictive model. The purpose is to decrease variance (bagging), bias (boosting), or improve predictions (stacking).
Dimensionality reduction
In machine learning classification problems when there are too many factors or variables, also called features. When most of the features are correlated or redundant, dimensionality reduction algorithms are used to reduce the number of random variables. Certain features are selected and others are extracted.
Parameterized statistical models
Machine learning models are parameterized to tune their behavior for a given problem. Noise contrastive estimation (NCE) is an estimation principle for parameterized statistical models. NCE is a way of learning a data distribution by comparing it against a defined noise distribution. The technique is used to cast an unsupervised problem as a supervised logistic regression problem. NCE is often used to train neural language models in place of Maximum Likelihood Estimation.
- Noise-contrastive estimation
Computer Vision
Computer vision is the ability of artificially intelligent systems to “see” like humans. In the computer vision field machines are developed that automate tasks that require visual cognition. Deep learning and artificial neural networks are used to develop computer vision. The following are topics related to computer vision as well as tools and libraries. Companies developing or selling computer vision products are under the Computer Vision subheading under the AI applications and companies section.
- Computer vision
- Convolutional neural network
- Biometric scanning
- Facial recognition
- Self-driving vehicle/ autonomous vehicle
- OpenCV
- Matlab
- TensorFlow
- AForge.NET/Accord.NET
- SimpleCV
- GPUImage
- Google Cloud Vision API
- Amazon Rekognition
Natural language processing
Natural language processing is a branch of AI that helps computers understand, interpret and manipulate human language. The following are tools and topics related to NLP. NLP companies developing or selling NLP applications are found in the AI applications and companies section under Natural language processing.
- Natural language processing
- Text-to-speech
- Google Cloud Natural Language API
- Paralleldots
- FriendlyData
- Neural machine translation
- Word2vec
- FastText
- Tacotron
- Tacotron2
Quantum Computing
Advances in deep learning are expected to increase understanding in quantum mechanics. It is thought that quantum computers will accelerate AI. Quantum computers have the potential to surpass conventional ones in machine learning tasks such as data pattern recognition. The following are topics, companies and technologies that link quantum computing and AI.
Semantic Computing
Semantic computing deals with the derivation, description, integration and use of semantics (meaning, context and intention) for resources including data, document, tool, device, process and people. Semantic computing includes analytics, semantics description languages, integration of data and services, interfaces and applications. In AI, semantic computing involves the creation of ontologies that are combined with machine learning to help computers create new knowledge. Semantic technology helps cognitive computing extract useful information from unstructured data in pattern recognition and natural-language processing.
- Semantic Computing
- Semantic Software Asia Pacific
- The IEEE Computer Society Technical Committee on Semantic Computing (TCSC)
- Resource Description Framework (RDF)
IoT (Internet of Things)
The Internet of Things (IoT) refers to objects that connect and transfer data via the internet and the sharing of information between devices. IoT based smart systems generate a large volume of data including sensor data valuable to researchers in healthcare, bioinformatics, information sciences, policy and decision making, government and enterprises. AI can be combined with machine learning for analysis of data and prediction.
Artificial life and evolutionary computation
While some lines of AI research aim to simulate the human brain. Artificial life or animate approach is concerned with the conception and construction of artificial animals as simulations or actual robots. It aims to explain how certain faculties of the human brain might be inherited from the simplest adaptive abilities of animals. Evolutionary computation is a generic optimization technique that draws inspiration from the theory of evolution by natural selection.
- Artificial Life (ALife)
- Swarm robotics
- Swarm intelligence
- Swarmalator
- Lenia
- Plantoid
- Cellular Automaton
- Conway's Game of Life
- Genetic algorithms
- Quantum artificial life
AI applications and companies
The following are companies using AI to develop products or producing AI software for various applications. AI programs designed for a specific applications are also listed.
Medical, veterinary and pharmaceutical
- Auransa
- Betasight
- CareSkore
- Desktop Genetics
- Doc.AI
- Enlitic (Computer vision for radiology)
- FarmaTrust
- Genetesis
- GTN Ltd
- Janus Genomics
- Juno Bio
- Juniper Medical Computing
- Kheiron Medical Technologies
- Kiroku
- KroniKare
- LifeMine Therapeutics
- macro-eyes
- Medios Technologies (Computer vision software for retinal images)
- Mighty Health
- Neurobit Technologies (Machine learning and sleep technology)
- Machine Medicine Technologies (Parkinson's disease assessment using computer vision)
- Primer
- Recursion Pharmaceuticals
- BenevolentAI
- Remedy Health
- Tyde
- BlueDot
- GYANT
- K Health
- Steth IO
- Cardiogram
- ContinUse Biometrics
- Binah.ai
- Sickweather
- StethoMe
Chemical industry
- Enzbond - prediction programs for enzyme development
Computer vision, image recognition and generation
Computer vision has applications in healthcare, security, manufacturing and transportation.
- Airraiy
- SeizeFace
- Umbo Computer Vision
- Emotech
- ViSenze
- SenseTime
- Iris Automation
- AImotive
- Latent Logic
- Anduril Industries
- GumGum
- AdMobilize
- Loom.ai
- Shazura
- Castle Global
Natural language processing
Voice recognition
Chatbots
- Cleverbot
- ChatFuel
- X.ai
- Destin AI
- Concierge (YouThere)
- Sensay
- OJO Labs
- Octane AI
- Youper
- Obie.ai
- Giant Otter Technologies
- Haptik
- AllyO
- Solvvy
- Kasisto
- Tact.ai
- Mya Systems
- Amazon Lex
- Aylien
Education
Business
Game playing
- Pysc2
- Airsim
- Leela chess (LCZero)
- AlphaGo Zero (Deepmind)
- AlphaZero (Deepmind)
- Master (software)
- DeepStack
- Libratus
Art and music creation
Agriculture
Autonomous navigation
Autonomous driving
- Tesla (autopilot / driver assistance / selfdriving car)
- Drive.ai
- Michelangelo (Uber)
Banking/Finance
Drone navigation/Military
Industry/Factory automation and monitoring
Fitness
Brain-computer interface
Cyber security/Law
Employee Behavior Analytics
Hotel Assistant
Sales/Marketing/Retail/Customer service
Trademark Analysis
Real-estate
Social Media/Human Interaction/Recruitment
Personal assistant
Selfie Apps
Data science
Robotics
AI software and API development
Other AI companies
Investment
Timeline
People
Dragos Grozavu
Founder
Further reading
Documentaries, videos and podcasts
Companies
News
References
Contributors
- 1
Meredith Hanel
Medical Genetics PhD, following life science applications that improve the health of humans and our planet. - 2
Rahul Abhyankar
Writer, travel instructor, and computer engineer. - 3
Daniel Frumkin
Mechanical engineering, cryptocurrencies, AI, and travel. - 4
Melanie Manipula
Researcher, writer and a former librarian - 5
Alex Dean
Healthcare Research Analyst @ Veradigm | Masters in Biostatistics/Epidemiology: UC Berkeley | BA in Biology: Wesleyan University - 6
Niroj Manandhar
Software Engineer - 7
Carla Faraguna
Golden People and Operations. - 8
Arthur Smalley
MSc Radiation Detection and Instrumentation & BSc Physics. - 9
Jude Gomila
Founder and CEO of Golden, Entrepreneur. Investor in 200+ companies. Cantabrigian (Gonville and Caius). Interested in the universe. - 10
Sergiu Miron
I am a person with a strong interest in obtaining knowledge and sharing it with the world.