Elastic Compute for Deep Learning

Cloud service providers such as Amazon Web Services are offering data scientists and machine learning specialists out there new compute options designed to accelerate specialized applications like training deep learning models. These compute resources are more cost-effective than other options, and if you’re currently using GPU-enhanced instances to train your models in PyTorch or TensorFlow then you should consider trying these new accelerated EC instances like the DL1. To get the word out, Amazon is sponsoring a hackathon next month to attract attention from the developer community, and it comes with a $200 AWS credit to help you develop your eligible computer vision or natural language project idea.

Powered by 8 Gaudi Accelerators

Amazon claims their DL1 instances are up to 40% more cost-effective than GPU instances for training deep learning models, and looking under the hood it’s easy to see why. Each instance is powered by 8 Habana Labs’ Gaudi accelerators cross-connected to each other with 100 Gb/s networking, and each with a programmable Tensor Processing Core (TPC). Their integration into the AWS ecosystem also makes it a snap for you to install supporting software (Habana Labs’ SynapseAI software stack), frameworks and drivers from an AMI or container. Habana Labs developer resources site can jump start your existing model migration, leveraging their model garden examples or developing your own custom kernels.

What is this about a Hackathon?

The AWS Deep Learning Challenge is scheduled for 5. January, 2022, through 14. February, 2022. You can sign-up on the hackathon’s devpost.com website. Be sure to read the terms and conditions carefully, particularly about having the intellectual property rights to any data your deep learning model needs to train itself on.

They are looking for computer vision or natural language processing project ideas; two areas for which there are no shortage of training materials out there for students looking to get more experience in the field of machine learning.

Developers and teams with an eligible project idea can request AWS credits, which should help defray the cost of training your deep learning model on DL1 instances.

Tip to Reduce DL1 Training Costs Further

AWS EC instance spot pricing constantly changes, but while DL1 instances are still relatively new there is going to be a surplus of instances available early on. You may be able to stretch your AWS credits in this hackathon by checking DL1 spot pricing availability during off-hours. I was surprised to see them priced at about a two-thirds discount.

Looking for a Better Result

Some of you who attended my 2019 Seattle Code Camp demonstration of an OpenCV-based supervised machine learning model for identifying user emotional sentiment from facial expression recognition. Using 160 training images of actor Patrick Stewart; could I identify the emotion of another 40 unseen images of the actor? Disappointingly, the results were little more accurate than what would be expected by chance. I chalk it up to irregularity across my training data set (not all of the head shots were taken from a fully-frontal perspective, and also lighting conditions and age varied significantly).

This time I’m going into this hackathon with an even better project idea (one that overlaps both computer vision and natural language processing categories, actually). After the hackathon is over, I will be sure to describe my experience with using PyTorch on the DL1 instance here for you all. Hopefully, I have convinced some of you to sign-up for the hackathon yourselves.

Make it so.

Groundhog Day for Missouri Football Fans

Groundhog Day for Missouri Football Fans

It’s Groundhog Day in the state of Missouri. Much like the Bill Murray film of the same name, events have repeated themselves for the state’s football franchises. It is remarkable then, that the key to a Lombardi trophy for two teams on opposite sides of the state across two generations arose not from a midwest wheat field, but the streets of Philadelphia. Continue reading

Preventing Spear Phishing Attacks in 2020

With the world preoccupied by retrospectives on the calendar year gone by, a particularly problematic news story slipped out in the waning days of December. This Microsoft announcement smacks down domains associated with yet another state-backed cyberthreat. It’s one more warning to users of the continued threat posed by so-called spear-“phishing” attacks. Could 2019 be the last year we have to worry about this?

Continue reading

Music Assisted Language Learning

If you have been through the American educational system, you have undoubtedly encountered the foreign language requirement in most college-preparatory programs.  Young people, most often during their high school years, are introduced to two years of basic language instruction.  How can we best learn a foreign language, particularly if we have been raised in a monolingual setting?

Continue reading

Earth Science and Continuous Inspection

Earth showing one quadrant breaking-up into disassembled bits.

Earth as the product of the continuous integration of its bits.

On this Earth Day I’m preparing to start a new job, one full of IT management buzzwords like continuous inspection (CI), continuous integration (my other brother, CI) and continuous delivery (CD). When making policy towards our homeworld though, it strikes me that these proven management practices fall by the wayside of many “pro-business” policymakers. What does CI for earth science really mean? Continue reading

Shellvarnia: Shell Variables Down Unix

Political map showing Shellvarnia on the left side of the continent, Flapjackia, separated from Userland (on the east coast of Flapjackia) by Environmentopia in the middle. No direct connection between Shellvarnia and Userland is shown.

Political map of continent, Flapjackia, showing countries of Shellvarnia, Environmentopia, and Userland.

This is a fun short story about the fictional land of Shellvarnia, located Down Unix, that occurred to me one night while I was reading Brian Ward‘s excellent book, How Linux Works: What Every Superuser Should Know. Mixing one dash of political satire with a heaping serving of Unix lore, it will entertain and inform novices approaching a Linux VM for the first time.  If you’ve ever wondered about distinctions between local shell variables and environment variables, /bin/sh from /bin/csh, and who built that wall between Shellvarnia and Environmentopia, then this story was written just for you.  Continue reading

Birthday Paradox

M.C Escher inspired birthday balloons overlapping each other paradoxically. Watercolor illustration by Allen West.

M.C. Escher inspired birthday balloon paradox watercolor by Allen West.

A math magician appears before you on the street in a puff of smoke, and he will make you his apprentice if you can but answer one question. You fantasize how you could make your investment portfolio grow if only you knew magic, or how you could star in your own series of blockbuster films–move aside Harry Potter! But before you get carried away, the magician posits his question to you: how many people off of the street do I need to invite to my magic show to have a 50% chance that any two of them celebrate their birthday on the same day? You sweat a little, think a little, then sweat a little more before answering: 183.

You won’t be attending Hogwarts anytime soon, my friend. You’ve been hit by, you’ve been struck by, a Birthday Paradox! Continue reading

When Averages Attack: For the Love of Means

What is not to love about Averages? (Candy heart image with the word Averages generated by: ImageChef.com)

What is not to love about Averages?
(Courtesy: ImageChef.com)

Americans love the average. Ask the average american on the street what he knows of statistics, and they will probably answer in so many words about something relating to an average (arithmetic mean). An average describes for us the central tendency of some data; the whole distribution of whose values we find it easier not to remember. Yet averages have a darkside to them, beyond sunny days on a baseball diamond figuring your favorite batter’s batting average. Let us look at statistics… when averages attack! Continue reading

What was Good for General Motors?

Charlie E. Wilson, Secretary of Defense (1953-1957) as shown on cover illustration of Time Magazine.

Charlie E. Wilson, Secretary of Defense (1953-1957) as shown on cover illustration of Time Magazine.

This Groundhog’s Day I revisit the Senate Armed Forces Committee 1953 confirmation hearing of highly-paid General Motors CEO, “Engine” Charles E. Wilson.  How different are today’s governmental ethics concerns from those of our grandparents?  Continue reading

Subresource Integrity – The Cyber Defense of 2016 You Haven’t Heard Of

Padlock over circuit board layoutAs we make plans this New Year’s Eve to bid farewell to 2016, we’re continually beleaguered by headlines concerning cyber attacks. The stories range from the latest allegations of Russian malware found on a Vermont utility’s computer, to leaked e-mails ahead of the U.S. Presidential Election, and you could even include Apple’s refusal earlier this year to furnish U.S. Government law enforcement agencies with a backdoor they insisted upon to enable encrypted data access. Cybersecurity must go down as one of the leading themes of 2016, and likely will go on to concern us for years to come.

That’s why we should take a moment to recognize a most important development within the information security community during 2016: the publication of the Subresource Integrity (SRI) recommendation by the W3C. Continue reading