Major Subfields of AI - Computer Vision (CV)

 Computer Vision (CV) is a very significant branch of AI that can help machines to see and analyze visual data in form of pictures and videos. Its subfields are object recognition, image classification, and object detection which find application in self-driving cars, medical imaging and security systems. Others include video processing, which includes motion estimation and event detection and video tracking. 

The major fields in computer vision.

• Image and Object Recognition: Object, face and other visual image recognition and classification.
• Object Detection: Moving past classification and locating objects within an image, in addition to classifying them, which may be achieved by drawing bounding boxes around them.
• Image and Video Segmentation: The process of splitting an image or a video into parts in order to identify various objects or areas. This is semantic (labeling pixels with class names) and instance (labeling individual instances of an object) segmentation.
• Motion and Tracking: The process of studying how Objects move over the timeframe of a video and estimating the movement direction of objects, generally applied in tracking objects, or decoding actions.
• 3D and Pose Estimation: Recalling a 3D representation of a scene or 3D locating an object or a person.
• Image Restoration: Restoration or repairing damaged or noisy images to enhance the quality of an image. 

Common applications

• Autonomous Vehicles: The idea of autonomous cars is to allow cars to sense the world around them by identifying other cars, pedestrians, and road signs.
• Healthcare: Helping with the analysis of medical imaging, e.g. by detecting tumors in CT or diseases in X-rays.
• Robotics: Giving robots visual abilities in performing such tasks as inventory management, defects and route-finding.
• Security and Surveillance: Easily the facial recognition feature is provided, the anomalous behavior is detected and monitored over a large area.
• Manufacturing: It is used to automatically check the products to help to check defects and assume product quality.
• Augmented and Virtual Reality (AR/VR): following and placing the virtual objects in the real world.

Types of AI: Narrow AI vs General AI vs Super AI

Key components of AI

Intelligence has a broader context that reflects a deeper capability to comprehend the surroundings. However, for it to qualify as AI, all its components need to work in conjunction with each other. Let’s understand the key components of AI.

Machine learning: Machine learning is an AI application that automatically learns and improves from previous sets of experiences without the requirement for explicit programming.

Deep learning: Deep learning is a subset of ML that learns by processing data with the help of artificial neural networks.

Neural network: Neural networks are computer systems that are loosely modeled on neural connections in the human brain and enable deep learning.

Cognitive computing: Cognitive computing aims to recreate the human thought process in a computer model. It seeks to imitate and improve the interaction between humans and machines by understanding human language and the meaning of images.

Natural language processing (NLP): NLP is a tool that allows computers to comprehend, recognize, interpret, and produce human language and speech.

Computer vision: Computer vision employs deep learning and pattern identification to interpret image content (graphs, tables, PDF pictures, and videos).

Types of AI

Artificial Intelligence can be broadly divided into two categories: AI based on capability and AI based on functionality. Let’s understand each type in detail.

1. Narrow AI

Narrow AI is a goal-oriented AI trained to perform a specific task. The machine intelligence that we witness all around us today is a form of narrow AI. Examples of narrow AI include Apple’s Siri and IBM’s Watson supercomputer.

Narrow AI is also referred to as weak AI as it operates within a limited and pre-defined set of parameters, constraints, and contexts. For example, use cases such as Netflix recommendations, purchase suggestions on ecommerce sites, autonomous cars, and speech & image recognition fall under the narrow AI category.

2. General AI

General AI is an AI version that performs any intellectual task with a human-like efficiency. The objective of general AI is to design a system capable of thinking for itself just like humans do. Currently, general AI is still under research, and efforts are being made to develop machines that have enhanced cognitive capabilities.

3. Super AI

Super AI is the AI version that surpasses human intelligence and can perform any task better than a human. Capabilities of a machine with super AI include thinking, reasoning, solving a puzzle, making judgments, learning, and communicating on its own. Today, super AI is a hypothetical concept but represents the future of AI.

Now, let’s understand the types of AI based on functionality.

4. Reactive machines

Reactive machines are basic AI types that do not store past experiences or memories for future actions. Such systems zero in on current scenarios and react to them based on the best possible action. Popular examples of reactive machines include IBM’s Deep Blue system and Google’s AlphaGo.

5. Limited memory machines

Limited memory machines can store and use past experiences or data for a short period of time. For example, a self-driving car can store the speeds of vehicles in its vicinity, their respective distances, speed limits, and other relevant information for it to navigate through the traffic.

6. Theory of mind

Theory of mind refers to the type of AI that can understand human emotions and beliefs and socially interact like humans. This AI type has not yet been developed but is in contention for the future.

7. Self-aware AI

Self-aware AI deals with super-intelligent machines with their consciousness, sentiments, emotions, and beliefs. Such systems are expected to be smarter than a human mind and may outperform us in assigned tasks. Self-aware AI is still a distant reality, but efforts are being made in this direction.

History of Artificial Intelligence: From Turing to ChatGPT

 The history of AI has been a progression of conceptual work by Alan Turing in the 1950s, to modern systems such as ChatGPT, which have been characterized by stages of advancement and AI winters. Before 1956 The term artificial intelligence was coined in a workshop at Dartmouth in 1956, followed by the implementation of the so-called expert systems and early chatbots such as ELIZA in the 1960s and the resulting AI winter because of over-promising and under-delivering. The renaissance was accompanied by the development of machine learning and deep learning, and in 2022 the ChatGPT large language model was released, which proves to be able to converse on many subjects.

Early foundations (1950s)

• Alan Turing: He was viewed as a father of modern computing and artificial intelligence, and in 1950, he developed the so-called Turing Test to determine the power of a machine to behave in an intelligent way that would not be perceived as that of a human being.
• Dartmouth Workshop: In 1956, this workshop was the first workshop that used the term artificial intelligence which brought together the main researchers.
• First AI and Machine Learning: Scholars started to consider the ideas of artificial neural networks and what would come to be known as machine learning with researchers like the Shopper program on the EDSAC computer showing that its past search history could be used to learn. 

The period of AI winter and rule-based (1960s-1980s). 

• Expert Systems: The earliest so-called expert systems were developed in 1965, and they were aimed at mimicking the work of a human expert.
• Chatbots: Joseph Weizenbaum wrote a chatbot in 1966, called ELIZA, one of the earliest chatbots, which could use natural language processing to imitate a psychotherapist.
• First AI winter Research funding declined in the 1970s and 1980s because of a report on the lack of progress and the constraint of computational power, which became the first AI winter. 

Revival and modern AI (1990s-present)

Modern AI (1990s-present) Revival Revival art typically takes the form of art derived from an original artwork, with its distinctive methodological advances aligning with the revival era more closely.Revival and modern AI (1990s-present) Revival Art Revival art usually represents an art work based on an existing piece of art, and its new methodological developments can be more closely associated with the revival period.

Machine Learning and Deep Learning: The resurgence of AI research in the 1980s was due to the advancement of machine learning algorithms and the creation of deep learning.

Growth of Data: The proliferation of digital data to be used in training became a main point of improvement.

Deep Blue: this was the first milestone when in 1997 the Deep Blue computer played and won against the world chess champion Garry Kasparov.

AlphaGo: In 2016, AI succeeded the master of Go, Lee Sedol, and this feat was achieved by the AlphaGo created by Google, which led to the idea that AI is able to master a game that is believed to be much more complicated than chess.

ChatGPT: ChatGPT is a 2022 large language model that utilizes deep learning and the Generative Pre-trained Transformer (GPT) architecture to chat with humans and complete several text-based tasks, or ChatGPT.