Artificial Intelligence (AI) and its subsets are playing a major role in machine learning (ML) and deep learning (DL) data science. Data science is a broad process that involves pre-processing, analysis, visualization and prediction. Gives deep dives into AI and its subsets.
Artificial Intelligence (AI) is a branch of computer science that is capable of performing tasks to create smart machines that typically require human intelligence. AI is basically divided into three sections as follows:
Artificial Narrow Intelligence (ANI)
Artificial General Intelligence (AGI)
Artificial Super Intelligence (ASI).
Narrow AI is sometimes known as ‘weak AI’, it performs a particular task in a best way. For example, snatch an automatic coffee machine that gives a precise definition of the activity for making coffee. Although AGI, also known as ‘Strong AI’, performs a wide range of tasks that involve human-like thinking and reasoning. Examples of Google Assist, Alexa, Chatbots that use Natural Language Processing (NPL). Artificial Super Intelligence (ASI) is an advanced version that performs human capabilities. It can perform creative activities such as art, decision making and sensitive relationships.
Now let’s look at Machine Learning (ML). It is a subset of AI that involves modeling algorithms that help make predictions based on the recognition of complex data patterns and sets. Machine learning emphasizes enabling algorithms to learn from the data provided, to gather insights, and to make predictions about previously unrelated data using the collected data. There are different methods of machine learning
Learning to supervise (weak AI – task driven)
Non-supervised study (Strong AI – data driven)
Semi-supervised study (Strong AI-Cost effective)
Strong machine learning. (Powerful AI – Learn From Mistakes)
Supervised machine learning uses historical data to understand behavior and make predictions for the future. Here the system consists of a designated dataset. It is labeled with parameters for input and output. And analyzes the ML algorithm as new data, new data arrives and gives accurate output based on specific parameters. Study under supervision can perform classification or regression functions. Examples of classification functions are image classification, facial recognition, email spam classification, fraud detection, etc. and weather forecasting for prevention work, population growth forecasting, etc.
Unsupervised machine learning does not use any classified or labeled parameters. It focuses on discovering hidden structures from labeled data to help systems discover a function correctly. They use techniques such as clustering or reducing dimensionality. Data points are grouped with metrics similar to clustering. Examples of data driven and clustering are a few dimensional reductions in movie offerings, customer segmentation, buying habits, etc. for Netflix users, such as feature description, big data visualization.
Semi-supervised machine learning works using both labeled and labeled data to improve learning accuracy. Semi-supervised reading can be an affordable solution when labeling data becomes expensive.
Reinforcement learning is quite different when compared to supervised and unsearched learning. This can be defined as an error in the delivery of results in the testing process and results. T is achieved by the principle of repetitive improvement cycles (must learn from past mistakes). Reinforcement learning is used to teach agents autonomous driving in a simulated environment. Examples of key-learning reinforcement learning algorithms.
Moving on to Deep Learning (DL), it is a subset of machine learning where you build algorithms where layered architecture follows. DL uses multiple layers to sequentially extract high-level properties from raw input. For example, in image processing, the lower layers can detect edges, while the higher layers can detect human-related concepts such as numbers or letters or faces. DLs are commonly referred to as deep artificial neural networks and are set of algorithms that are perfect for problems such as word recognition, image recognition, natural language processing, etc.
Briefly, Data Science covers AI, which includes machine learning. However, machine learning itself covers another sub-technology, which is deep learning. Thanks to AI since it is able to solve harder and harder problems than humans (better than cancer detection than oncologists).
Data science & machine learning service [forwordit.digital]