When it comes to technology in the world of business, AI and machine learning (ML) are among the commonly discussed terms. Technology has been making inroads continuously, and ignoring its effects and implications would be a perilous choice.
Here is what some well-known personalities have had to say about these technologies:
- “AI doesn’t have to be evil to destroy humanity – if AI has a goal and humanity just happens in the way, it will destroy humanity as a matter of course without even thinking about it, no hard feelings.”
– Elon Musk, technology entrepreneur
- “Artificial Intelligence, deep learning, machine learning — whatever you’re doing if you don’t understand it — learn it. Because otherwise, you’re going to be a dinosaur within 3 years.”
– Mark Cuban, American entrepreneur, and television personality
An introduction to AI, ML, and DL
A brief introduction to the three would be the following:
- AI, simply put, covers everything related to making machines smarter. It typically has three types:
- Artificial Narrow Intelligence (ANI): programmed to complete a particular task, oriented toward a goal
- Artificial General Intelligence (AGI): machines learn, understand, and act in a manner that cannot be distinguished from how humans would act.
- Artificial Super Intelligence): presently hypothetical – AI with machines whose intelligence exceeds that of humans
- ML, a subset of AI, refers to an AI system that uses an algorithm to self-learn and get smarter over time, without needing human intervention. It is what helps to build AI-driven applications. ML is seen in action in recommendation systems in streaming services for music and movies, for instance. ML algorithms typically comprise supervised learning, unsupervised learning, and reinforcement learning.
- DL, or deep learning, refers to the application of ML to large data sets, where complex algorithms work to train a model. Inspired by how a human brain filters information, DL systems use neural networks to help computer models to predict and classify information by filtering input data through layers. Driverless cars are a common example of DL, whose types include convolutional neural networks, recurrent neural networks, and recursive neural networks.
Here is a diagram that shows how closely AI and machine learning are related, and where DL falls within these:
AI essentially works to develop machines that are self-reliant and can think and act like humans. Examples of AI are machine translation such as Google Translate, speech recognition apps such as Google Assistant or Siri, and AI robots such as Aibo and Sophia.
Types of AI
- Reactive machines: systems that do not form memories or leverage past experiences to make decisions; only react
- Limited memory: machines referencing the past over a period of time, retaining that information for a brief period
- Theory of mind: systems capable of understanding human emotions and their effect on decision-making, and accordingly adjusting their behavior
- Self-awareness: systems designed to be self-aware, understand own internal states, predict feelings of other people, and act accordingly
ML looks to solve business problems through predictive models built on analytics and computer models. The work of a machine learning engineer is seen in sales forecasting, stock price predictions, and banking fraud analysis, among others.
Types of ML
- Supervised learning: systems use labeled, past data to predict future outcomes, and require at least an input and an output variable to be provided to train the model.
- Unsupervised learning: systems identify hidden features and patterns from unlabeled input data on their own, making patterns and similarities clearer once the data is more readable.
- Reinforcement learning: seeks to train the machine to complete a task in an uncertain environment . The machine sends actions to the environment and receives observations and a reward from the environment, with the reward measuring the success of the action with regard to meeting the goal.
A subset of ML, DL works with artificial neural networks employing algorithms inspired by the structure and working of the human brain. DL algorithms can work with huge amounts of both structured and unstructured data; ML, in comparison, typically requires structured data. Use cases include the detection of cancerous tumors and other objects and the coloring of images.
Types of DL
- Convolutional Neural Network (CNN): a type of deep neural network commonly used to analyze images
- Recurrent Neural Network (RNN): builds a model by using sequential information; works better for models that must memorize past data
- Generative Adversarial Network (GAN): an algorithmic architecture that uses two neural networks to create new, synthetic data that passes for real data
- Deep Belief Network (DBN): a generative graphical model comprising several interconnected layers of hidden units, or latent variables; the units themselves are not connected