AWS Innovate AI & ML Edition

In this post I’ll cover my experience with the recent AWS Innovate online conference, focusing on Machine Learning & AI.

This was something of a departure for me as my job and AWS focus has so far been predominantly databases, but the area of Machine Learning & AI is a rapidly expanding one and with the tools AWS brings to the table this AWS Innovate edition represented an ideal opportunity to get some insight. Plus AWS provide a certificate of attendance for those committing to viewing the conference on the day (although bizarrely this is sent as an email as opposed to a PDF) and, more importantly, $25 AWS Credits which is doubled if you give feedback.

On this occasion AWS offered four learning tracks ranging from new to advanced users, along with a hands-on track on Twitch demoing AWS DeepLens & AWS DeepRacer:

The keynote was provided by Julien Simon, the Principal Technical Evangelist for AI & Machine Learning at AWS. The Innovate keynotes are usually done very well and this was no exception - as someone with no AL/MI experience I wasn’t lost and I reckon that those attending for the advanced sessions wouldn’t have been bored. Broad overviews of some AWS AL/MI services were given followed by an overview of each of the learning tracks, so I would have been clear on where my time was best spent had I not read the agenda beforehand.

I chose the I’m Starting My Career In Machine Learning track (I’m not, but the other tracks are a bit rich for my blood at this point). The first half started with Alex Casalboni’s Introduction To Machine Learning With Python And Scikit-Learn followed by Martin Beeby’s session on Getting Started With Machine Learning Using Amazon SageMaker. Martin’s enthusiasm for SageMaker was evident from the start and, after a well-paced run-through of the stages of building, training and deploying ML models, he presented a substantial (and downloadable) demo of using the SageMaker dashboard and Jupyter notebooks that I’m looking forward to having a go with at some point.

The second half of the conference featured Adrian Hornsby’s Introduction To Deep Learning and Antje Barth’s Getting Started With Deep Learning Using Amazon SageMaker presentations, followed by closing remarks from Julien Simon and AWS Machine Learning Hero Pavlos Mitsoulis-Ntompos. The presenters were clearly very knowledgeable on deep learning and, although some of the points went over my head, all presentations and demos were concise, informative and flowed well. Deep learning has clearly come on over the decades, although judging by one of the images in the slides I guess the machines aren’t taking over anytime soon.

Overall the experience was a good one. I’ve attended a few AWS online conferences now and the experience is always polished and of a high standard. Resources are plentiful, with each presentation including its slides as a minimum and demos where featured (as the presentations themselves are not available for download). On this occasion each presentation also ended with GitHub links for project files that can be used to explore the content discussed.

My personal takeaways from the event are to review the learning content on the SageMaker dashboard, to take a closer look at Jupyter notebooks and to test out Textract with some PDFs from my archive to see if any ideas take flight. If I get time I also want to check the other learning tracks for anything of interest as well as the contents of the Resource Centre, while I’ll write about separately if I find anything of interest. For anyone reading this who’d like to view the content themselves, the sessions are available on-demand at the AWS Innovate Site until the end of November.

Thanks for reading ~~^~~