Understanding Artificial Intelligence, Machine Learning, and Deep Learning

Artificial intelligence (AI) and its subsets Machine Learning (ML) and Deep Learning (DL) are playing an important role in data science. Data science is a comprehensive process that involves pre-processing, analysis, visualization, and prediction. Let’s dive into AI and its subsets.

Artificial intelligence (AI) It is a branch of computer science that deals with the construction of intelligent machines capable of performing tasks that normally require human intelligence. AI is mainly divided into three categories as shown below

• Narrow Artificial Intelligence (ANI)

• General artificial intelligence (AGI)

• Super artificial intelligence (ASI).

Narrow AI, sometimes referred to as “weak AI”, performs a single task in a particular way in the best way. For example, an automated coffee machine steals that performs a well-defined sequence of actions to make coffee. While AGI, which is also known as ‘strong AI’, performs a wide range of tasks that involve thinking and reasoning like a human being. An example is Google Assist, Alexa, Chatbots, which uses natural language processing (NPL). Artificial Super Intelligence (ASI) is the advanced version that exceeds human capabilities. You can engage in creative activities such as art, decision-making, and emotional relationships.

Now let’s see Machine Learning (ML). It is a subset of AI that involves algorithm modeling that helps make predictions based on the recognition of complex data sets and patterns. Machine learning focuses on allowing algorithms to learn from provided data, gather insights, and make predictions on previously unanalyzed data using collected information. The different methods of machine learning are

• supervised learning (weak AI – task-driven)

• unsupervised learning (strong AI, data driven)

• Semi-supervised learning (strong AI, profitable)

• reinforced machine learning. (Strong AI: learn from mistakes)

Supervised machine learning uses historical data to understand behavior and formulate future forecasts. Here the system consists of a designated data set. It is labeled with parameters for input and output. And as the new data arrives, the ML algorithm analyzes the new data and provides the exact output based on the fixed parameters. Supervised learning can perform classification or regression tasks. Examples of classification tasks are image classification, facial recognition, spam classification, identification fraud detection, etc. and for the regression tasks there are weather forecasts, population growth prediction, and so on.

Unsupervised machine learning does not use any classified or tagged parameters. It focuses on discovering hidden structures from unlabeled data to help systems infer a function correctly. They use techniques such as grouping or dimensionality reduction. Grouping involves grouping data points with similar metrics. It is based on data and some examples of grouping are the recommendation of movies for the user on Netflix, customer segmentation, shopping habits, etc. Some of the examples of dimensionality reduction are obtaining features and visualizing big data.

Semi-supervised machine learning works by using tagged and unlabeled data to improve the accuracy of learning. Semi-supervised learning can be a cost-effective solution when data labeling is expensive.

Reinforcement learning is quite different compared to supervised and unsupervised learning. It can be defined as a trial and error process that ultimately yields results. This is accomplished through the beginning of the iterative improvement cycle (learning from past mistakes). Reinforcement learning has also been used to teach officers to drive autonomously in simulated environments. Q-learning is an example of reinforcement learning algorithms.

Moving towards Deep learning (DL), is a subset of machine learning in which algorithms are created that follow a layered architecture. DL uses multiple layers to progressively extract higher-level characteristics from the raw input. For example, in image processing, lower layers can identify edges, while upper layers can identify concepts relevant to a human, such as digits, letters, or faces. DL generally refers to a deep artificial neural network and these are the sets of algorithms that are extremely accurate for problems like sound recognition, image recognition, natural language processing, etc.

To summarize, data science covers artificial intelligence, which includes machine learning. However, machine learning itself covers another sub-technology, which is deep learning. Thanks to AI, it is able to solve increasingly difficult problems (such as detecting cancer better than oncologists) better than humans.