I was struggling to find inspiration for a topic to write about today, but then I remembered this guy.
When I saw the Google’s AlphaGo, I realized something really serious is happening here [..] That was the trigger for me to start developing the cucumber sorter with deep learning technology.
Makoto Koike’s family own a cucumber farm and he was so impressed with the idea of deep learning, after seeing AlphaGo’s famous victory over Lee Sedol, that he decided to create an algorithm that sorts cucumbers.
It’s an incredibly intricate classification and computer vision problem, and just a cool use of deep learning. No doubt you’ve heard the term ‘deep learning’ and the ‘hype’ that surrounds it.
I’m sure we’re just at the surface of the amazing things that can be accomplished with deep learning, and I just wanted to share some of the (I think) really cool uses of deep learning out there.
Computer vision is a hard problem in itself, and AI research has been trying to solve various image recognition problems for decades. However, there are some offshoots of image-related deep learning projects that I think deserve a mention.
Researchers have come up with an algorithm that can take a sketch drawing and automatically create a “solid” line trace of it. Their results are incredible.
Taking black and white images and drawing them in colour, in cases where a colour version doesn’t even exist.
Artistic Style Transfer
Taking an image and redrawing it in the style of a famous painting.
Deep learning algorithms can also generate realistic-sounding music!
Some of my favourite AI-powered projects that I’ve seen relate to text generation.
Footnote: This is the 24th entry in my 30 day blog challenge.