Current state of AI and machine learning technology

In the Computer world, at present, the most speaking topic is Artificial intelligence which is a marvelous Computer Engineering design known so far, such that the transistor made a revolution in the whole computer world. Today, applications of artificial intelligence are in a wide range, which involves most computational activities. It is not a thing, that happened in one night. It took years to develop AI and train those models carefully because there are so many issues in the scope of artificial intelligence.

The artificial intelligence concept originated in the mid-20th century when researchers began to explore the possibility of creating machines that could perform tasks that need human interference and intelligence such as reasoning, critical thinking, problem-solving, and learning. In 1980-the 1990s this concept was rapidly growing in the world. Machine learning, a subfield of AI also began to develop during this period. After 2000, the machine learning concept was extensively developed that allows computers to learn from data rather than pre-programmed algorithms. Machine learning is a type of artificial intelligence that enables computers to learn from data instead of a program. This is done by some algorithms specially written for improving their performance by analyzing fetched data. There are 3 main types of machine learning.

1. Supervised learning: This involves training a model using labeled data, which means the correct output for each input is provided to the model in advance.
2. Unsupervised learning: In here, no labeled data are available. The algorithm analyzes data to find and discover patterns in the datasets provided to it.
3. Reinforcement learning: Reinforcement learning is a kind of learning from mistakes. It is a trial-and-error method of learning, which allows the computer to label data on its own experience.

Applications of machine learning

Machine learning is applied in a vast range of applications in today’s world. Some of them are healthcare, finance, retail, e-commerce, energy and utilities, and transportation. Machine learning algorithms are used to diagnose diseases from symptoms, analyze medical reports, and detect cancers as well. It is also very useful in making financial reports and doing calculations easily and efficiently.

Machine learning is about to be used and is already being used extensively in transportation. The major factor that involves machine learning in transportation is autonomous vehicles. It is also used in traffic predictions and route optimization in applications like Google Maps.

Machine learning algorithms are used for chatbots, voice recognition, language translation, and image generation according to text. At present, there are applications available that can provide the most relevant and appropriate information on a given topic, applications that control according to voice commands, and dictating applications as well. Still, there are some fields about to incorporate artificial intelligence and machine learning, and in the near future, almost all human activities may incorporate AI.

Natural Language Processing

Natural Language Processing (NLP) is another subfield of artificial intelligence that deals with recognizing, understanding, and generating human language as the words imply. NLP is mostly used in chatbots, language translation, converting text to speech, and vice-versa. NLP uses some techniques to analyze data.

1. Tokenization: It breaks a sentence into words and embeds tokens for each word or phrase to identify it.
2. Part-of-speech (POS) tagging: This means NLP assigns a grammatical label to each word, like a noun, verb, adjective, etc.
3. Named Entity Recognition: Identifying named entities in a text, such as organizations and people. Some social media platforms use this technique to identify people in an image or tag them with text.

Autonomous vehicles, computer vision, fraud detection, gaming, recommendation systems, and medical diagnosis are the most trending machine learning applications in the current world. Autonomous vehicles use image recognition to detect objects, humans, and traffic signals to make decisions and drive safely on the road without any human intervention. A special kind of OS known as RTOS (Real Time Operating System) is used in these autonomous cars and can timely respond to critical events such as auto-braking when an object is detected.

Computer vision is another trending field of AI. It allows machines to interpret and understand visual data and provide useful outputs based on them. One of the major uses of computer vision is object detection. This is used in social media platforms, face recognition in cellular phones, autonomous vehicles to detect obstacles, etc.

The medical sector widely uses AI nowadays to facilitate their treatment activities as well as detect some diseases. AI is being used to improve the accuracy of mammography screening for breast cancer by helping radiologists identify potential tumors more quickly and accurately. Artificial intelligence is not only used in these areas but could be also used in many other sectors about which we have not thought so far.

Therefore, the current situation of artificial intelligence as well as machine learning is being frequently updating and changing with a wide variety of new applications.

Kalhara Batangala
Faculty of Engineering
University of Sri Jayewardenepura