Can computers think like humans?
In 1950, Alan Turing posed the question “Can machines think?” in a paper. There he went on to describe a task, now known as the Turing Test, where the test of machine intelligence lies in its ability to answer questions in the way a human would. Which means that a human shouldn’t be able to tell the difference between the answers generated by a computer and another human being.
However, it wasn’t until 1956 that John McCarthy actually coined the term “artificial intelligence” (AI) at a conference. During this period, early work with neural networks stirred much excitement for “thinking machines.”
Then, the period during 1980–2010 saw a surge in machine learning (ML) and a new facet of AI. In 1997, IBM’s Deep Blue beat the renowned chess champion Gary Kasparov and in 2011, the company’s question-answering computer system Watson won the quiz show Jeopardy.
Present Day
Today, we’re seeing much excitement around deep learning (DL), a subset of AI, with applications in autonomous vehicles, computer vision, natural language, robotics, and many more.
Autonomous Vehicles
Multiple models work together to make sure an autonomous car functions properly. Some specialize in the classification of pedestrians while others track weather and other anomalies as the car navigates the roads.
Computer vision
Deep neural nets help computers with object detection, image classification, image segmentation, and image restoration. DL models are continuously automating human visual tasks.
Translations
Google translate relies heavily on the progress made in the field of natural language processing. Today, AI can not only understand human speech, but it can also generate it.
Robotics
Robotics is increasingly trying to optimize its autonomy through AI. Robots are being trained and tested to carry out specialized autonomous tasks such as carrying objects, driving vehicles, and performing small surgeries.
In short, AI is expanding to various fields and becoming a part of our everyday life through its diverse applications. In automating and optimizing human tasks, AI does indeed “think” like humans. And just like humans are hungry for new knowledge, AI thrives on data but this data generation is costly — both in time and money.
Procedural.AI offers an AI-first design solution to solve AI’s data needs by creating physically accurate 3D scenes from a natural language description. This means tons of synthetic data for training autonomous systems! Such data allows flexibility (multiple iterations!), ease of use (text to a scene), and focuses on the inherent problem of design in most workflows.
