Computers Sense Our Absurd Cat Video Obsession
Every once in a while someone will upload a video on Youtube enjoying a disgusting bowl of Cheetos with milk or choking on a peanut butter and pickles sandwich. There are just some things that don’t pair well together, or do they?
While deep learning had its beginnings in the early 1960s, it only started to catch up speed by 2011 and 2012. This is when the big moguls, Google, Facebook and Microsoft started to experiment with this technology and began adopting into its practices. Just when scientists had innovated and played along enough with these new discoveries, these big companies became early adopters. Early adopters generally have large financial power, are able to influence big groups and often are very discrete in their adoption. Do we ever know what Google is scheming?
What was Google doing in 2012?
Watching funny cat videos. Well, actually 16,000 computers did.
Funny cat videos have become a cultural and viral phenomenon, enticing users to share these videos at the speed of light every day, to friends and colleagues. Contrastingly, not many people really know what deep learning is all about. This new science is exploring how human brain patterns and algorithms can be replicated through technology. Cat videos and deep learning don’t seem very appetizing, right?
Google was creating one of the biggest neural networks through the application of deep learning concepts and testing it. Researchers from the X laboratory at Google decided to have 16,000 computers scanning 10 million images on Youtube videos and see if the computers could recognize cat images. What was different this time is that researchers didn’t provide any input to help computers identify a cat. No one told a computer what a cat looked like. Instead, similarly to the human brain, the computer was learning from repetition. Similar image after similar image the computer learned what a cat looks like.
This application can already be found in face recognition software used by police bodies.
Who else was an early adopter of Deep Learning in 2012?
If we are talking about computers, it would be a big miss not to mention computer royalty Microsoft. Right about the same year, the company’s chief research officer introduced Microsoft’s Deep Neural Net (DNN) translation technology. As captured in video, his English speech was translated on the spot and it was spoken by a machine in another language. Not only did the machine found similar words in Mandarin but it also rearranged them to reflect how the Chinese language works. Also, machine actually simulated his voice. Today, speech recognition software is popular in mobile technology. We have Siri, Kinect, Alexa and a few others.
This application can already be found in voice recognition software used for elderly care.
Last but not least, who could be more interested in having computers understand the human brain and work like it? Facebook. Right around 2013, the social media leader hired a renowned NYU professor Yann Lecun to lead their AI developments. Some of the company’s projects include Torch and Caffe2 which include machine learning and computer frameworks that can be used and applied by the general public. Surprisingly, or not, it may not be that hard to steal some of Facebook’s secrets.
This application can already be found in GPU technology used in self-driving cars.
— —
There is not a day that we don’t hear people getting frustrated with technology and complaining about how unresponsive it is or how little user friendly it is at times. Computers may not be humans and they might not think how the human brain thinks yet. However, what we can’t deny is that scientists started working on bringing computer intelligence to the level of human intelligence years ago and that big technology power figures are helping with this.
Deep learning is now widely spread in photography through the regeneration of coloring, in the gaming industry through virtual reality and in the health industry through medical imaging just to name a few.
It is possible that computers will be smarter than humans one day; but I don’t think they will be spending most of their time watching funny cat videos as we do.
Are you using Deep Learning technology?
— —
References
Deep Learning. (n.d.). Retrieved March 11th, 2018, from http://deeplearning.net/
Markoff, J. (2012). How Many Computers to Identify a Cat? 16,000. Retrieved March 11th, 2018, from http://www.nytimes.com/2012/06/26/technology/in-a-big-network-of-computers-evidence-of-machine-learning.html
Meyer, D. (2012). Microsoft’s translation breakthrough: Speak, and hear your voice in Chinese. Retrieved March 9th, 2018, from http://www.zdnet.com/article/microsofts-translation-breakthrough-speak-and-hear-your-voice-in-chinese/
Simonite, T. (2013). Facebook Launches Advanced AI Effort to Find Meaning in Your Posts. Retrieved March 9th, 2018, from https://www.technologyreview.com/s/519411/facebook-launches-advanced-ai-effort-to-find-meaning-in-your-posts/
The 5 Customer Segments of Technology Adoption. (n.d.). Retrieved March 9th, 2018, from https://ondigitalmarketing.com/learn/odm/foundations/5-customer-segments-technology-adoption/
