Artificial intelligence (AI) was first proposed in 1956 and is considered one of the three cutting-edge technologies of the 21st century; it simulates the information processing involved in human thought and consciousness. The concept of AI is quite broad and can be understood, at a superficial level, as enabling machines to think and solve problems like humans. In fact, core AI technologies encompass many aspects, including reasoning, knowledge, planning, learning, communication, perception, mobility, and the ability to manipulate objects.

Generally, we classify AI into the following three major categories based on its capabilities:

Weak AI: AI that excels in a single aspect. For example, there is AI capable of defeating a world chess champion, but it only knows how to play chess; if you ask it how to better store data on a hard drive, it would not know how to answer.

Strong AI: Human-level artificial intelligence. Strong AI refers to systems that can match humans in every respect and perform any mental task that humans can. Creating strong AI is far more difficult than creating weak AI.

Superintelligence: Superintelligence could be systems that are slightly superior to humans in every aspect, or systems that are several times more capable than humans in every respect; however, there is no consensus within the industry on whether this can ultimately be achieved.

There are two different approaches to implementing artificial intelligence on computers. One is to use traditional programming techniques to make the system exhibit intelligent behavior, without considering whether its specific structure resembles or matches that of the human brain. The other is the simulation approach, which not only focuses on the final results but also requires the system’s structure to resemble or match that of the human brain.

Machine learning examines how computers learn from data. One of its primary areas of research is how computer programs automatically learn to recognize complex patterns based on data and make intelligent decisions. Machine learning is a rapidly growing field. Here, we introduce some classic machine learning problems related to data mining.

Supervised learning: This is essentially synonymous with classification. The supervision in this learning process comes from labeled instances in the training dataset.

Unsupervised learning: This is essentially synonymous with clustering. The learning process is unsupervised because the input instances lack class labels.

Semi-supervised learning: A class of machine learning techniques that uses both labeled and unlabeled instances when training a model. In one approach, labeled instances are used to learn class models, while unlabeled instances are used to refine class boundaries.

Active learning: A machine learning method that allows users to play an active role in the learning process. Active learning methods may require users to label instances, which may come from a set of unlabeled instances or be generated by the learning program. Given constraints on the number of instances that can be labeled, the goal is to improve model quality by actively acquiring knowledge from users.

Deep learning is a new field in machine learning research, motivated by the desire to build neural networks that simulate the human brain’s analytical learning processes. It mimics the mechanisms of the human brain to interpret data such as images, sounds, and text. The concept of deep learning originated from research on artificial neural networks. A multi-layer perceptron with multiple hidden layers is one type of deep learning architecture. Deep learning discovers distributed feature representations of data by combining low-level features to form more abstract high-level representations of attribute categories or features. The concept may seem obscure and somewhat difficult to grasp, but behind its seemingly esoteric nature lie far-reaching applications and a promising future.

Today, machines trained using deep learning technology outperform humans in image recognition—for example, in identifying cats, detecting cancer cell characteristics in blood, and identifying tumors in MRI scans.