AWS DeepLens Hackathon: A Machine Learning Newbie’s Journey Into the Abyss

Don’t Stop Learning

I know next to nothing about Machine Learning.  Shoot, I don’t even have a C.S. Degree.  But damn the torpedoes, full speed ahead, I’ve committed to completing a project for the AWS DeepLens Hackathon, currently slated to conclude on this February 14th, 2018.  Technology is advancing at a break-neck pace and this is one way to try to keep up.  Plus, I’ve heard that the first Trillionaires will be minted from the A.I. Industry, so show me the money! I thought it would be cool to blog some of my experience using the DeepLens technology during this hackathon (and I’m actually writing this blog on my DeepLens device).

Actually, on a side note, I’m a little worried about where technology is going these days, especially with such a strong emphasis on A.I. and autonomous machines and all, all driven by profit and power motives, not REAL problems.  If we’re not innovating, we’re dying, right?  But as Sun Tzu once said, keep your friends close but your (potential) enemies closer.

Confronting the Beast

I went to AWS re:Invent last November in Vegas and somehow managed to get into one of the last DeepLens Workshops of the Conference.  Competition for these workshops was, shall we say, fierce!  Attending meant I had to miss the re:Play party, but at that point I didn’t really care since the D.J. was not Van Buuren, Garrix or AfroJack.  By attending the DeepLens Workshop, I was able to take a free DeepLens computer home, and even received a voucher for $25 worth of AWS Credits to get started.

I was really psyched to get started on the hackathon upon returning home, but I already had a project in progress I had to complete first.  Fortunately, I finally completed my Android App (my second Android App ever) and got it released in the Google Play Store (CANDLES Tracker) on 1/11, so I finally have my evenings ‘free’ to devote to this hackathon.

Progress

So yesterday, 1/12, I cracked open my DeepLens box and unpacked the device.  I realized I needed a new keyboard, mouse and HDMI cable with a micro-HDMI male end.  So last night I ordered these things, off of Amazon…of course, and received them in a Prime Shipment by the time I got home from work this evening.

Tonight, 1/13, my DeepLens is all connected and registered with my AWS Account.  I imagine it’s not going to take me long to burn through the $25 AWS Credits when I start uploading data to train my models, which I will hopefully get a better feel for this weekend.

Once the camera boots-up, you can log into the OS using the password ‘aws_cam’, which is the same as the username.  You can connect to wifi and use Firefox to get on the internet and access your AWS Account from there.  Strangely though, the instructions say to connect to the DeepLens Wifi endpoint from another computer and configure it using a browser pointing to http://192.168.0.1.  I found this strange as I was already logged into the device, but could not get Firefox to connect to http://localhost to connect to the configuration portal from within the device.  But it’s all working now by simply not thinking and following the instructions.

Screenshot of DeepLense Ubuntu Desktop and Video Stream from Camera.

Next Steps

Start deploying some of the Amazon pre-built models to get a better feel for the deploy process and integration possibilities outside of the device.

I’ve recently gotten hooked on this Carl Cox Ibiza set, which is a nice groove for hacking to, I’ve found:

2 thoughts on “AWS DeepLens Hackathon: A Machine Learning Newbie’s Journey Into the Abyss

Comments are closed.