| Danial Norjidi |
ARTIFICIAL intelligence (AI) and machine learning can improve processes and increase efficiency in different types of segments, and offers many opportunities around which business can be built.
These were among the points highlighted by Google Developer Expert from the Philippines Ralph Vincent Regalado while delivering a keynote presentation at the Google DevFest Brunei organised by Google Developers Group (GDG) Brunei on November 24.
In his presentation, entitled ‘Machine Learning using Google Technologies’, he began by defining AI, explaining that it means giving computers the ability to think and learn. “So it’s more on putting knowledge or learning capabilities and abilities onto machines themselves.”
On how machines learn, and how this is done, Regalado said, “You do it by thinking about how relatively you must learn, and in a sense, when you create an AI project, you always think about what the task that you need to do is, what the experience is, and how you measure performance. It’s relatively different compared to how regular software development works.
“In AI, in machine learning, you always continuously find ways on how you can actually improve your algorithm, how to improve your system. You measure it in a variety of ways, and you teach the machine based on experience.
“What do I mean by experience? For example, if you’re going to teach a machine to classify apples and oranges you need to provide data, such as pictures of apples and oranges, for it to actually learn. The same thing with how humans learn – when you first learn to write, you need to do it repeatedly so that you can write a smooth letter ‘A’. It’s the same thing when you went to school. If you want to learn mathematics, you need to actually have experience doing mathematics, algebra, trigonometry and the way that you actually measure that you’re learning from it is by taking an exam.”
He said it’s the same with machines. “You need to train them, you need to put experience in them so that they can actually learn.”
In his presentation, Regalado noted that there are two types of AI. “First is what we call ‘Weak AI’, which can only perform one task. For example, if you teach a machine to classify images, it can only perform that specific task only. It cannot actually do other tasks.”
The other type is ‘Strong AI’, which he explained can perform different types of tasks at the same time. “So you can do multi-tasking. You can actually classify data using a camera, for example, while doing another thing.”
He said that across the globe a lot of work and research has already been done with regard to AI.
As an example, he mentioned Google DeepMind’s computer program called AlphaGo, which beat the human world champion player of the board game Go in 2015.
He also shared that AI and machine learning power a lot of the technology that people use in their everyday lives, such as automated spam detection and friend lists on social media.
Regalado highlighted that AI and machine learning have already been there, but that in the last three years they have been getting a lot more attention and more opportunities, especially for AI developers.
“In Strong AI, it’s still at the early stage. It’s only starting to develop what we call Strong AI.”
He went on to share that machine learning is being applied to many different sectors and domains, “because it can actually improve processes and increase efficiency in different types of segments”.