By Mushfiq M, MehtA+ AI/Machine Learning Research Bootcamp alum
In part 1 of a three part series on Artificial Intelligence (AI), we focus on what is artificial intelligence and identify common research areas in AI. If you would like to learn more about artificial intelligence, check out the AI camps MehtA+ offers at https://mehtaplus.com/.
Artificial intelligence (AI), which seemed a distant future only some decades ago, has started to become a part of our daily lives.
So how did this happen? As people live a large portion of their lives in the digital world, we now have access to unimaginable amounts of data. Computers also have become very powerful and can store and process this data very efficiently.
What is Artificial Intelligence?
Artificial Intelligence is a field of computer science dedicated to building algorithms by which computers can exhibit human intelligence. In a subset of artificial intelligence, deep learning, scientists use neural networks modelled after the human brain to learn patterns from large amounts of data. Using neural networks, computers can ‘understand’ words, make decisions and ‘recognize’ objects.
Research in artificial intelligence has been focused on learning, logical thinking, problem solving, perception and understanding language.
Learning Computers learn how to play chess, for example, by trial and error. The computer knows the rules of the game. When the computer first starts playing the game, it makes random moves. As it plays several games of chess, it learns the strategies to win the game.
Reasoning Computers can infer facts from existing data or reason in multiple ways. Via deductive reasoning, the computer can use facts to deduce a conclusion. For example, if we knew that “Sam is a computer scientist. ” and “A computer scientist programs”, then the computer will conclude “Sam programs”.
On the other hand, in inductive reasoning, the computer can arrive at a conclusion by generalizing facts. For example, if we knew that “All cats in our neighborhood have fur”, the computer might (incorrectly) generalize “All cats in the world have fur”. (We see you Sphynx cats!)
Problem Solving There are many different methods by which computers can solve problems. The obvious method is solving a problem step-by-step. However, that is not always feasible in a short amount of time. In the heuristic method of problem solving, the computer finds an approximate solution instead of an exact one in a short amount of time.
Perception We take in the world through our five senses. Our interpretation of the data we take in is what we call perception. Similarly, autonomous vehicles or self-driving cars take in the world through various sensors and a 360 degree camera. These cars then interpret this data using artificial intelligence algorithms.
Understanding Language One of the most popular applications in artificial intelligence is understanding and processing language. Chatbots, such as ChatGPT, and voice assistants such as Alexa or Siri can respond to humans for example. In addition to conversing with humans, computers can also summarize text, analyze sentiment and translate text from one language to another.
While the development of new artificial intelligence models that were unimaginable only a few decades ago is very exciting, at the end of the day, it is important to remember that it is humans and human intelligence that is crucial to artificial intelligence. Whether it is the choices we make in selecting training data, labeling data or crafting an algorithm, human input is needed in each step of the process.
It will be up to us how we choose to use and regulate artificial intelligence. Only time will tell how useful (or detrimental) artificial intelligence will be for human beings.
