Machine learning is here. Allow me to explain. If you follow this blog, you know I love technology, and have a strong belief that it will control much of our lives in the future. As recently as this last week I gave a lecture where I included this statement “Google cannot see your images” I went on to say, that if you use white hat tactics, everything will be all right.
Convolutional Neural Networks
Machine learning will dramatically effect out future. To begin with I own a SEO company, so the ultimate question for me is how it affects SEO; we will get back to that in a minute. Machine learning is the process of teaching a machine (Google in this case) to understand, or see our images. Although I recently mislead the group I was speaking to (not inferentially) I have come to believe that Google can in fact see our images. I have a great friend named Gary Horsman, he also is involved in SEO for a living. His most recent blog (SEO Notebook) he tried to prove that Google can in fact see out images. He is a deep thinker and would never make a statement that was unproven or required additional research.
Of course I immediately called him and we had a lengthy discussion. He had done his research, believe me. I found it hard to dispute the findings in his blog (you should go read it) Machine learning is no longer futuristic concept. Although I believed it was inevitable. I was unconvinced that it was settled. I was unaware of the ”ImageNet Large Scale Vision Recognition Challenge” (or ILSVRC) It is a competition sponsored by non-other than Google. They have been running it for years, and someone wins every year. However, I have never seen or heard about it on the news, I wonder why?
Google is the Winner
Schools with big abbreviated names Mit, CalTech compete in the challenge alongside Micro-soft and Google own teams. Machine learning or rather Image recognition it the end goal. Surly Google has more to worry about than our Title or Alt tags, and if we are cheating. But they would have a lot to gain if they could see our images (or any image for that matter)
A practical application of Machine Learning, more than just being able to identify our images, what it you owned a self-driving car. You would need to identify images maybe several thousands of images, maybe thousands of times a second, just to be safe. Is that car coming at me? Do I need to apply brakes? Is this my turn? If I owned a self-driving car, I would have to have Machine Learning be can interracial part of it OS. Of course I would need the most extensive map of all of the streets in the world, and processing speed and power that is unmatched like a Convolutional Neural Network functions. “O” wait now Google has all of that. And is developing self-driving technology too! Imagine that.
I will stick to my original statement “If you use white hat tactics, everything will be all right” Google is formidable and smart. In my speeches I frequently compare them to Sun Tzu. Ruthlessly intelligent, clever, shrewd, and cool. Trust Google to keep their cards close to their chest, and just like the tactics of Sun Tzu, only deployed or show as much as is needed. Keeping Machine Learning as a weapon only to be deployed when and as required.
And always have e2-Total Solutions help with your enlightenment and Machine Learning.