Introduction: Neurality

Introduction: Neurality

“Deep learning” is a term that’s used to be only for those in the know. Now it’s getting tossed around everywhere:

  • Gigaom had posted back in November 2013 a great “guide” to deep learning
  • Microsoft recently released a cloud-based machine learning platform Azure ML
  • Ersatz Labs has just switched to a public offering of their cloud-based take on it
  • And yes, there is a startup called Skymind for all of you who remember Terminator (and catch the obvious reference)

Of course, those are just a sampling, but, for the layman (and I include myself in this bucket), what is deep learning?

Simply put: algorithms that once applied to sets of data, structured or unstructured, are capable of a) figuring out what factors are creating specific results, b) adjusting itself to work better, and therefore c) can provide insight which a data scientist, while theoretically is capable of getting to the same conclusion remains limited in capacity to digest and interpret at the same pace.

Are you still with me?

That’s okay if you’re not quite there yet – I’m going to give an example later on that shifts theoretical definition into concrete applicability. However, per title, I’m going to focus on the startup Neurality, and their initial product, NeoPredict, for three specific reasons:

1)      I find the offering they have typifies the immediate value that can be gained from utilizing “deep learning” because they’ve structured it in a way that makes it accessible, easily integratable, and at a reasonable price point for a new product

2)      They have a real product roadmap and a consistent, solid plan to realize that roadmap

3)      The proof is in the pudding, and they have the pudding


[Disclaimer: I have spoken directly to the CEO and co-founders, James Costantini, but only in the aspect of making sure I properly understand what I’m talking about to convey it to you. This posting is by no means represents any specific sort of endorsement.]


Let’s Talk About Pudding

I like to shop, how about you? While I do my best to control my retail habits, on occasion I let myself loose and depending on where I am, what card I have in my wallet, if I don’t happen to have cash, and what I’m purchasing, there are days where my card companies have a field day in trying to figure out a) is this really Cassandra or b) was her card stolen.

One specific incident comes to mind. I had purchased a mid-price watch at a well-known retailer and decided to use a debit card. As I was stepping out of the store, I received a phone call from my bank: “Please call and confirm, etc., etc.,” Now, I’d already had the purchase in hand, but I called and did the verification process on the transaction mainly to avoid getting a continuing series of annoying automated messages. (Mind you, they also called my house phone, too). Irritating? Mildly. I recognized it was a card I didn’t use often, but the amount was not excessive and I would have preferred that my bank simply knew it was a me and didn’t put me in the position to have to do an affirmation.

How does this relate to deep learning and Neurality? Take a look at this case study. Integrated at the point of authorization, that entire experience could have been better. Theoretically, I could have been identified as the obvious authorized user and never had to make a phone call and do a verification. Or, if it hadn’t been me, the goods never would have left the store, which is a win for both the retailer and the bank.

Let’s not kid ourselves: we know that there are a number of institutions, particularly financial ones, that have significant amounts of data on you – spending habits and locales, demographics, et al. – and the combination of that data, historical, static, and dynamic, structured and unstructured, with the right set of algorithms, can quickly become incredibly effective in areas like fraud and transaction analysis.

Off the top of my head, I can think of three other areas of finance that could benefit from a tool like NeoPredict applied to existing data sets: FX trading, commissions management, and the AML/KYC process.


Now and Soon

Have you ever funded a Kickstarter project? I love Kickstarter: great ideas and an opportunity to participate at the formation of something real. However, there have been some complaints that products don’t get shipped or they take forever and a day and it’s a bit of a downer.

That’s what I call the case of “now” versus “soon”. The space isn’t yet crowded for deep learning, but you name something a trend (thanks, Gigaom) and adherents will come, and all of sudden there will be quite a bit of “coming soon”.

This is not the case with Neurality. They have proven case studies, I believe they have already signed up a client, they have an active Beta Program which is still accepting companies to effectively co-invest in making this a reality, and best of all, NeoPredict is out of the box ready to integrate Predictive Analytics with BI tools a number of enterprise already have (Tableu, Spotfire, Qlik). Icing on the cake is that they do it fast, and, at least for right now, the pricing won’t strain your business.

Regarding the “soon” part of this, in speaking with Mr. Costantini, he highlighted two things which are worth sharing:

1)      The offering they already have is not just the beginning because they have dreams + a pie in the sky; they have a definitive product roadmap on which they’re continuing their research and build with set release time periods and add-on functionality including additional heuristic models and out of the box integrations with platforms far beyond the reach of BI tools

2)      “This has the potential to change the world for better.” – That sounds big, I know, but remember when we talked about wicked problems? Neurality was able to hit off a >95% identification of breast cancer when applied to laboratory datasets. If that’s not changing the world for better then I don’t know what is. And this is just the beginning.


Why this matters

Because now is the time and place to take advantage of an opportunity which on the small-scale could immediately impact your understanding of data plus your ability to make better decisions off of it, but on a larger scale could allow strategic decision making that can alter the competitive landscape.

What do I mean by that? Let me tell you exactly what my parents used to tell us: when the blind man in the village says he’s going to pelt you with the stone, he already has the stone underneath his foot.

To bring that home, right now, we’re all blind men. We’re feeling about and guessing at business decisions. We have data scientists but, with deep respect, sometimes they’re sometimes more like soothsayers than fully reliable. For now, that’s an acceptable risk, but that time is passing. Therefore, the company that wraps its head around and integrates deep learning and predictive analytics faster and better than anybody else is going to going to pick up quite the sizeable stone.

Neurality is the best undercover bet right now. It’s going to be a serious player and it’s worth engaging now. That is, if you don’t want to get pelted by that stone…


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s