The first ever Data Intelligence Conference hosted by CapitalOne went from Friday (6/23) to Sunday (6/25). Hosted at the beautiful Capital One McLean Conference Center, the conference featured three days of exciting panels, over 300 attendees to network with, and lots and lots of food.
The Talks: Starting with Technical
Unlike most conferences where the keynote ends up being either a sales pitch or a warm kumbaya inspiration talk, Data Intelligence kicked off with a talk by Professor V.S. Subrahmanian on BEEF: Balanced English Explanation of Forecasts. In addressing the growing challenge of explaining blackbox machine learning models to a non-technical audience, Prof Subrahmanian proposed an extensible framework to cluster predicted outcomes into explainable bins and an algorithm to optimize those bins based on overlap and inclusiveness.
Dr. Subrahamian wasn't the only one tackling explainability of complex models. Patrick Hall from H2O.ai also gave an excellent talk on Ideas for Interpreting Machine Learning, filled with stunning visualization, insightful anecdotes and hilarious analogies.
Throughout the these two days, I've been treated to panels like Building a Gigaword Corpus, Playing Detective with CNNs, or Identifying Language Communities. The full schedule can be viewed here. Not only did the conference include heavy hitting technical talks, it also brought in experts from around the data space, legal practitioners to talk about privacy and engineers to talk about complexity and integration.
The Vendors: No Salespeople In Sight
One thing that I've always noticed about technical conferences (and conferences in general) is the army of vendors lurking about, sometimes in a specific area. While they're sometimes nice for chats about potential work solutions and picking up some swag, they also reflect a reality that many conferences are balance-sheet focused and signing up for conferences with a real email address might be a suboptimal strategy. I'm looking at you ODSC West!
Data-Intelligence had only one lonely vendor booth and it was pretty much just a recruiter from CapitalOne. As they're the ones hosting the conference, that is A-Ok in my book. The conference swag also came with a very comfortable t-shirt and a FIDGET SPINNER! As you can imagine, the fidget spinner was definitely a highlight and massively popular.
The Peers: Facilitating Networking
As people spent less time talking with vendors, they spent more time talking to each other. Whether it was the overabundance of food and coffee or the central location of the atrium, I spent more time pleasantly interacting with peers and fellow practitioners than at any other conference. It was a real joy to share experiences with exceptionally talented people from every corner of the profession.
As another attendee pointed out, the leisurely pacing of the sessions also contributed to promoting networking and discussion among peers. Generous 25 minute breaks (and snacks) after every two sessions allowed attendees to stop and reflect upon the best elements the sessions they just came from.
Off to another day!
I'm writing this as I get ready to go to Day 3 of the conference. So far, this has been one of the best data conferences I've been to. I'll likely do a full run down once I've had a chance to go through all my notes from this event!
Just an FYI, at the current time many of Steemit users posts do not appear on Google naturally, and if they do they take a fair amount of time before they are listed (at least several days after being originally posted). Because of this it is important that you try to manually add all your new posts to Google yourself to greatly speed up the indexing process, which is something that only takes a few seconds to do but can GREATLY increase your posts changes of success by helping to get it in front of more people who may be doing various searches using the Google search engine. To manually add your posts to Google, simply create or log into your Google account and once done go to the following link https://www.google.com/webmasters/tools/submit-url?continue=/addurl, and from there simply input your posts url in the empty field and push "Submit Request". Once done Google will index your posts url within seconds, instead of waiting days/weeks/months to find it naturally. I know it can be tedious to do this, especially if you have 100 posts already done and need to loop back and do this for all of them, but trust me it is worth the time/effort as doing this can really help deliver a lot more results/visitors to each of your posts. Anyways once done, to check if it worked simply go to the Google homepage and input your posts url in the search field and see what happens (before pinging to Google you will see "no results found", while after pinging a post to Google you will see that post listed at the top of the search results, which tells you that your content is now available for the public to find via Google going forward).
Sidenote: I have already ping this specific post to Google for you to help out!
For some more helpful Steemit tips make sure to check out my detailed post below.
9 Detailed Tips for Steemit Users Who Want To Get The Most Value From Their Steemit Experience!
https://steemit.com/steemit/@budz82/8-helpful-tips-for-new-steemit-users-who-want-to-get-the-most-value-from-their-steemit-exerpience
Anyways, I hope this helps!