February 12, 2018 Issue
Amir Husain, Founder & CEO of SparkCognition, discusses the impact of AI solutions
Amir Husain Founder & CEO
Interview conducted by:
Bud Wayne, Editorial Executive, CEOCFO Magazine, Published – February 12, 2018
CEOCFO: Mr. Husain, your bio indicates you wanted to build a company that would be at the forefront of the “AI 3.0” revolution. What was the impetus for this desire and do you feel you have accomplished this with SparkCognition?
Mr. Husain: I have been interested in AI from a very young age. Since I was four years old I have been into computers, and since I was a teenager I have been actively researching in the AI space. Therefore, this was a company that I waited many decades to create. Ultimately the process of delivering value to customers, users, and people through computers requires them to take on a level of intelligence that they have not yet had. In many ways, this is the ultimate pursuit for any computer scientist. When I formally incorporated and started SparkCognition in the summer of 2013, it was not quite clear that AI was to become, over the next few years, as big as it has gotten. However, I felt that the underlying components and the industry were ready, data was available, sensors had been deployed, and compute was cheap. All these things that were necessary ingredients were driving the success. What we saw over the subsequent years was AI becoming so popular. Yes, I have been wanting and waiting to found this company for many years. Have I achieved the objective? I think it is too early to say. Certainly, this is a very successful company from a commercial standpoint. We have the largest customers in the world from Boeing to Honeywell, Invenergy to Duke Energy, to the Department of Defense to the Ministry of Defense in the UK. Those are huge customers and from that point of view very successful. The overall mission is a mission for the rest of the century. I call it the AI century, and I think there is a lot more work to do. I think AI’s fundamental drivers are commerce and industry, policy in government, and in our own daily lives, so there is a lot more to do.
CEOCFO: Artificial Intelligence plays a very important role in so many industries today. Why have you chosen the sectors that you are in such as energy, manufacturing, finance, aerospace defense, and security?
Mr. Husain: These are areas that can really derive a lot of benefit from current AI technology. There is real business value to deliver in all of these areas. For example, in clean energy, we work with Invenergy, the largest private wind energy company in the U.S. We do failure prediction based on sensor data gathered from turbines. For that kind of an application, you are using our AI to make wind and clean energy more viable. That is a huge driver for the future of the energy industry, and we can do that with the groundbreaking technology we developed. In the finance space, there is a huge push for AI-based systems that can do everything from trade automatically to identify potential fraud, spoofing, and other types of transactions that you see on a financial exchange. We work with the London Stock Exchange and State Street Corporation, one of the largest banks in the world. That is another area which has a lot of unsolved problems where it does not make sense to have human beings look through reams and terabytes of data. It is not possible for them to do that. Finally, on the national security side, Vladimir Putin recently said that “he who rules AI will rule the world.” The stakes on the national security and defense side could not be higher. That, we think, is one of the most critical application areas for AI, and we are committed to doing everything possible to create an edge in that space.
CEOCFO: Before we discuss your product offering, what do you understand about AI and machine learning that sets you apart from other companies in this area?
Mr. Husain: If I was to talk about just one thing, and there are many, it is that it’s very important to look at transformative technology. That will change everything. It will change many applications, workflows, and functions. It is not a silo capability; it is a broad capability. When you realize that is the nature of how AI will add value to the enterprise and that is a broad capability that will amplify everything, then what that means is all kinds of people will have to use it. Today, AI is a lot of rocket science. One of the things that we have been working on heavily is what we call automated model building, an idea that AI can build AI and you do not have to have the domain experts and programmers and high-end rocket science staff for your company to use the technology. We have made great strides in that area and have actually advanced the state of the art. We have contributed to the science of AI in building algorithms and capabilities that really take us forward. We have also translated that scientific progress into tangible, usable products that a non-programmer can use. I think that has been a huge factor in how we have been able to quickly apply AI to so many different domains by using AI to build AI and create an experience that non-programmers can leverage.
CEOCFO: If the average person is putting together a wind turbine farm and they do not know much about computers or maybe they are not that knowledgeable about computers and AI. What would you say to make it easy for them to understand about what AI is?
Mr. Husain: At a high level, AI is basically the science concerned with building smart programs that think algorithmically. The computer comes up with ideas and evaluates them at machine speed and scale—things that would not occur to a human being because we’re constrained by common sense, our limited experience, and our ability to process limited amounts of data. All those constraints are removed with the current technology that is out there today, and AI is very good at perceiving things. It can find patterns where human beings cannot. In many areas now, AI is outperforming humans at pattern recognition. For example, we have a product in the cybersecurity category that detects malware that there are no signatures for. It looks at new things that have never been seen before and finds the kinds of patterns that would suggest if this is bad. That is very hard for humans to do, given the amount of malware being created. In another area, AI is now being used to identify handwriting samples with greater accuracy than human beings. These are all examples of perception data, and AI is doing very well in these areas. You asked about the wind energy industry. I think that’s a valid and powerful question. Many people will use this technology but won’t really understand how it works. Other than that high-level description I gave, what we have been focusing on is the value it delivers, and how it achieves that value is interesting but not critical for a user to understand. In the wind space, would you want to reduce the operating cost of your infrastructure? Would you want to identify failures before they happen so you can proactively group maintenance activities together and save yourselves huge amounts of money every time an unforeseen breakdown occurs? The answer to those questions is “Of course. That is a no-brainer.” How well are you able to do that today? Well, here is how we do it today, and here is the process, and we can forecast a failure six hours in advance. Moreover, how would you like us to do that six days in advance? Does that make a difference? Yes, it makes a difference. That is the value of AI for that specific industry. With Flowserve, the largest manufacturer of pumps in the U.S., and Dover, the second largest manufacturer, both of those companies are our clients. And again, being able to forecast when equipment will fail and what it will do are really valuable applications. That is the value that you are speaking—you are not trying to explain to them the kinds of AI.
CEOCFO: Would you tell us about your DeepArmor, SparkPredict and DeepNLP solutions and how they apply to the various industries you are selling them into?
Mr. Husain: At the heart of it, SparkCognition is about AI building AI—automated model building. And these are three products that encapsulate and represent different capabilities in individual products. DeepArmor is an endpoint defense product that identifies zero-day threats with no pre-engineered signature. The way cybersecurity has functioned up until this point in time is that threats like malware are found in the wild, researchers engineer signatures to identify those threats in the future, and send a list of those signatures to your software running on your laptops and desktops. That software then matches the signature of every file on your system with the engineered signature. The problem is if you do not have the signature for a new threat, that threat will not be identified. That is where the world is now, and that’s why endpoint security is so compromised and weak. With machine learning, we have created an incredibly accurate way to identify whether a file is a threat without requiring signatures. And DeepArmor is a leading-edge capability for that kind of cybersecurity. That is cybersecurity. However, SparkPredict is cyber-physical safety and security. SparkPredict does not look at files. It looks at sensor data coming not just from virtual items but from physical systems such as generators, turbines, pumps, ships, and so forth. It automatically builds models of the underlying asset and then projects whether the asset is about to experience a natural failure (for example, if it is going to have a safety-related incident or whether it’s under malevolent threat such as a hacker targeting that asset and causing a failure). That kind of threat can also be detected. Therefore, SparkPredict and DeepArmor come together to deliver cyber-physical safety and security. The third product, DeepNLP, is a natural language understanding product. The other two products build models out of structured sensor data and then data in files, whereas DeepNLP is focused on human language. It’s focused on large sets of documents designed for human consumption, and it can ingest all of them and do a lot of things with that kind of content. It answers questions about it, identifies similarities, and comes up with hierarchies of elements that were referenced in the document. For example, you might have a document about some kind of machinery, and then DeepNLP implies what the components of that machinery might be. It summarizes across multiple documents so you might have part of an answer in one manufacturer’s manual and the other part in another manual, and it can stitch those two together and answer questions in a cross-document way. Therefore, it does many things with language that was designed for human consumption. We have had great success with this in industrial and defense use cases where there are huge amounts of information that no human can consume, so this is a great system to be able to do that.
CEOCFO: Is there any contract that you are most proud of and why?
Mr. Husain: Our relationship with Boeing is very special to us. We are extending and advancing the state of the art. We are heading into the realm of using automation model building for autonomy not just to solve perception problems where you recognize a pattern but also using AI to take action. Our relationship with Boeing is one that is very strategic and important. Our relationships with our energy partners— such as Invenergy, Duke Energy, and Dover Energy have been seminal in getting the company focused on industrial use cases and solving a lot of problems. I would also like to mention Flowserve, which I talked about earlier, as another relationship that is key to us. At this point, we are approaching a hundred relationships. So to me, all of them are very important, but these are a few where we can give you an answer to your question.
CEOCFO: SparkCognition has received a number of awards and recognitions, such as Red Herring Top 100. What does that mean for you and is there any one that stands out specifically?
Mr. Husain: Awards are always great. It shows that our vision is becoming real, it’s showing results in the marketplace, and it’s solving problems for our clients. To me, that is the most important thing. We need to deliver value to the people that believe in us, and those people are our clients. In that sense, it is great to see that. One award that we won recently, which I am particularly proud of, is the CNBC Disruptor award, where CNBC publishes this list annually across the U.S. We were named, along with Airbnb and Uber and some of the most successful on that list, as a company that is going to disrupt the industry. We were very proud of that.
CEOCFO: SparkCognition has hosted AI Conferences in the past such as the “Time Machine 2017”. Is there anything on tap for 2018 that our readers should look forward to?
Mr. Husain: Time Machine was an absolute stellar success. It was one of the best AI conferences I did for the year—but of course I am biased—in the sense that it was probably the only conference that brought together real users of the technology with scientists and professors that are building the technology and combined that with policy people and generals as well as the former Secretary General of the United Nations and people of that caliber and stature. Policy, science, and application usage were combined on one stage. It was such a successful formula that 2018 will be an even bigger Time Machine, and we are very excited about that. We have planning which has already started around that. It is going to be an event that could be magnified many times.
CEOCFO: On a personal note, you are an author and CEO. Did you see yourself with such a successful life as you were growing up or was there a turning point for you?
Mr. Husain: I have a long way to go to be successful. To me, the goal was not to start a company or write a book. The goal was to pursue my passion. I have been madly in love with computers since the age of four. I have been obsessed with AI and computer science for many decades. We are at the tip of the iceberg with the potential of AI. I think of success as the outcome of an endeavor. I am very grateful for all that has come my way, and I have been very lucky. However, we still have decades of work to do. Did I think about the fact that I would build a company? Yes, I always wanted to build a company. I dropped out of my PhD program to start my first venture-funded company. I sold that and did another one. I like building things. I would not call myself successful.
CEOCFO: Why did you write the book, “The Sentient Machine: The Coming Age of Artificial Intelligence”? What has been the industry’s response?
“The Sentient Machine” is something I have been thinking about for many
years. I actively worked on the book for three years, and it really reflects
my thinking over a long period of time. The reason I wrote about the science
of AI through the lens of philosophy is to address some of the big questions
that we have around the coming of AI: Will it render us useless, will it
take away our jobs, and will it kill us? These are the fears that a lot of
people raise. I have a framework that I use to address those questions. I
have also combined that with multiple tractable domains where AI is being
used today and how applications such as autonomous weapons go back and tie
in with these existential concerns. I thought there was a big part of the
conversation that has not been covered in this way that I wanted to cover,
and I thought this would be a great way to put those ideas down. Even though
I was running a company and it took me almost three years to finish this,
the idea just started flowing when I picked up my pen because I realized I
was not coming up with those ideas at the spur of the moment. They were
thoughts that were bottled up in my head and flowing out of my fingertips.
Therefore, it was also a cathartic experience. I am pleased with how the
book has come out and grateful it has been doing well. It has been reviewed
the world over and featured on national television. It has made several
best-seller lists. The happiest I have been with this experience is hearing
from readers. I have gotten wonderful feedback which frankly makes my day,
week, and month. That has been very beneficial for me.
CEOCFO: Where will growth come from for SparkCognition over 2018?
We are approaching a hundred different relationships with customers, and we
are in expansion mode with many of them where we proved the application of
our products in one particular area and now that is being scaled, and there
are associated auxiliary applications with that same product. We are well
positioned, and the organic growth from within the community of customers
and clients that we have is going to be huge. The other area is we are
starting to acquire customers at a faster rate, and these customers are not
just domestic—they are international. So 2018 is going to be the year for
SparkCognition to seriously expand internationally. We already have some
footprint. We have several customers in Europe, but I think you will start
seeing us move into Asia and the Middle East.
CEOCFO: Do you have the funds in place to continue to build out your business or will you be looking to partnerships or investors?
Last year, at the time that we raised our last round, it was the largest
round raised in Austin the whole year. We have some additional announcements
coming out. We are one of the best companies in the space.
CEOCFO: In closing, why should potential customers, partners and
investors consider SparkCognition?
“To me, the goal was not to start a company or write a book. The goal was to pursue my passion.”- Amir Husain
Contact: John King
ceocfointerviews.com does not purchase or
recommendation on stocks based on the interviews published.