WHERE NO DERIVATIVES COMPANY HAS GONE BEFORE: RUMI MORALES AND THE FUTURE OF FINTECH

POLITICAL FORECASTING AND PREDICTING THE 2016 PRESIDENTIAL ELECTION

GREENER FUTURES: HONEST DOLLAR, FINTECH, AND THE FISH THAT’S EATING THE WHALE The world's first fully cognitive anti-malware system leverages machine learning, natural language processing and A.I. algorithms to analyze files and provide signature-free security

ANALYZES THE DNA OF FILES TO IDENTIFY THREATS

SUB-SECOND MALWARE DETECTION

SIGNATURE-FREE SECURITY

SELF-LEARNS AND RETAINS KNOWLEDGE

COMBINES STRUCTURED + UNSTRUCTURED DATA (INCLUDING NATURAL LANGUAGE) TO RESEARCH THREATS

REDUCES FALSE ALERTS

WWW.SPARKCOGNITION.COM @SPARKCOGNITION FEATURES 26 Maintaining Century-Old Trust with Cutting-Edge Cognition

Chris Corrado is working with the leading innovators in fintech, like SparkCognition, to ensure LSEG remains one of the most trusted financial organizations in the world. By John King

34 36 Where No Derivatives Company Political Forecasting and Predicting Has Gone Before: Rumi Morales the 2016 Presidential Election and the Future of Fintech by Ashvin Govil by John King and Caroline Lee

COGNITIVE TIMES Oct. 2016 1 contents

TECHNOLOGY MEDIA CONSUMER

12 Events 04 Important Dates From 17 Atlas Wearables: The History & Today Personal Trainer That Fits On Your Wrist

13 A.I. in Social Media 20 Digitization of Banks

06 The Next Generation of SECURITY Financial Data is Coming FINANCIAL

TRENDS 21 Machines Are Predicting What You’ll Buy

14 Finding Financial Exfiltration Using Evolving Culture in the 08 SparkSecure® Modern Workplace 24 Four Ways A.I. is Being Used in the Financial Industry

Greener Futures: Honest DeepArmor™ and the 31 Four Financial Technology 16 Dollar, Fintech, and the 09 Future of Cyber Trends to Watch in 2016 Fish That’s Eating the Whale

2 COGNITIVE TIMES Oct. 2016 FROM THE EDITOR by Amir Husain

AS I WRITE THIS, I am flying home to Austin, Texas it is shaping other areas, such as energy and defense. Traders from Abu Dhabi, capital of the United Arab Emirates. will leverage AI platforms to discover counter-intuitive but Wherever my travels take me in the Middle East, Asia, highly profitable strategies, banks will use machine learning Europe or America, it is hard to ignore the transformative to ensure compliance and lower costs, hedge funds will use effect AI is having across the world. At this very moment, predictive capabilities to gauge event outcomes and large the news item flashing before me is about President Obama's exchanges will surveil trades and make their clients' assets statement on how our next president will have to lead us more secure with cognitive techniques. And this range of in a world increasingly run by AI. A second piece of news applications presents just the tip of the proverbial iceberg. explains that United Airlines has delayed a large number of flights due to a suspected cyber attack. A discomfiting notion We hope you enjoy this edition of the magazine, and–as when one is 35,000 feet in the air! Security from cyber threats always–I look forward to your feedback. is a critical protection which nation states and organizations that bank on trust will have to deliver to their citizens and Best wishes, customers in future. And this safeguard will only be available Amir with the leverage of AI capabilities. Humans can't manage the size or scale of the burgeoning and increasingly automated cyber threat.

If there's one industry that runs almost entirely on the fuel of protection and trust, it is financial services. Financial services is a field ripe for AI leverage, as companies seek to make information more insightful and processes more efficient and secure. The impact of AI on finance is what we're focused on in this edition of Cognitive Times.

We have some great stories for you in the pages that follow, including an enlightening interview with one of the world's top Financial Technology executives, Chris Corrado of the London Stock Exchange Group. In all my interactions with him, I've found Chris keenly focused on making LSEG the world's most secure trading platform. He's been a leader in leveraging Artificial Intelligence for Security.

There's also a fantastic article on Rumi Morales, Managing Director at Chicago Mercantile Exchange's Venture Fund, CME Ventures. Rumi has been nothing less than inspirational in her advocacy for AI in Fintech. She doesn't just talk the talk, but backs it up by presenting a number of success stories she and her fund have been involved with.

There is little doubt that the future of AI and the future of Finance are entwined. AI will transform this space, much as

The Next Generation of Financial Data is Coming SparkCognition says A.I. has financial market applications.

by Peter Shadbolt *First published on openmarkets.com on July 29, 2016

ig Data mines don’t come much bigger or deeper David Rich, senior vice president of Global Part- than MasterCard. The financial services corporation nerships & Business Development at Master- Bhas some 2.3 billion cards in circulation, accepted Card, told CME Group’s “Tech Talk 5.0: Next in 205 countries at 38 million merchant locations gener- Generation Financial Technology” audience in ating an average of 160 million transactions an hour. July that making sense of the numbers requires an equally staggering number of equations. “Each time those transactions come in we process them,” Rich says. “We apply about 1.9 million rules and algorithms to make sense of those transactions. MasterCard has been doing this for so long that it has been able to cleanse and normalize data from a longitudinal perspective going back as far as 40 years. Central to the conversation, however, Rich emphasises, is privacy. “Privacy by design is a key component of how we look at data,” he says. “A significant amount of data analytics are actionable with regards to privacy. No matter what level of privacy you have, you can still provide a very significant level of analytical capability.” The data is used by just about every industry vertical on earth: MasterCard works with 44 of the top 100 retailers in the world, 10 of the top 15 restaurants, 13 of the top 25 hotel chains. The way the company aggregates data, he says, allows analysts to see consumer trends in real time. “This is very important for financial forecast- ing when you’re looking at it from macroeco- nomic perspective,” he says, adding that the U.S. Department of Commerce figures come out with a significant time lag often in the order of months. “When we look at the gasoline spend, for example, we can see that discretionary spend – the money that was being saved as gasoline prices went down – was not necessarily translating into discretionary retail spending. “We could see this much earlier than market analytics could.” AI models, he says, have the ability to outperform human models by an order of magnitude and can now accommodate the nuances of language to process meaningful data.

AI Applied to Finance

Drilling even deeper into next genera- five hours, using human-generated meaningful data. tion data with automated model creation models, to five days using its auto- “What it means is that soon AI is the work of the cognitive analytics mated models. He says it’s now time to services will be able to understand startup from Austin, Texas—Spark- begin applying these models to finan- almost at that palpable level. Cognition. CEO Amir Husain told Tech cial services. “When you see a good trader they Talk that its financial platforms are able “At first blush, it may appear there’s have a feel for the market, but what is to spot anomalous patterns in a veritable a huge difference between a machine that feel? It’s a collection of patterns sea of data. on one hand and financial markets on in their biological neural network that “We’ve proven that it works and it the other, but there are a lot of similar- get triggered when they see something works very well,” Husain says. “Today we ities from a mathematical perspective,” that smells, tastes, looks like something have close to 40 clients, some of them Husain says. they’ve seen before. the best-known companies in the world.” AI models, he says, have the ability to “We are now bringing these aspects, With applications for industry that outperform human models by an order the numeracy and language understand- range from wind turbines to pumps, of magnitude and can now accommo- ing, and we think this will democratise SparkCognition’s platforms have been date the nuances of language to process the way value is extracted from financial data.” able to extend failure predictions from COGNITIVE TIMES Oct. 2016 7 Evolving Culture in the Modern Workplace by Kimberly Erler

e’ve all heard the stories of some of perks offered at dot coms around the country: Slides, foosball, nap pods, yoga, free food – the lists go on and on. While these things certainly go far in driving a company’s culture, are they the things truly driving W today’s workforce to choose Company A over Company B?

How did this culture shift even happen? essentially accessible to their employers SHRM/Globoforce, “Companies with Did employees begin demanding more and/or clients 24/7, so it only makes peer-to-peer recognition are 35% more of their employers, or did employers see a sense that technology is playing a role likely to report lower turnover” (www. need and adapt? Kani Sterling, Director, in giving these employees a better work/ globoforce.com/resources). Address- Human Resources at Statesman Media, life balance. ing that truth is another Austin-based had this to say: One such tech company, Austin-based company, YouEarnedIt, which has taken "Part of this shift includes the emerg- Student Loan Genius, finds that 70% of employee rewards and recognition to a ing technology tools, where employees today’s graduates carry an average loan new level. By giving businesses an indi- are always 'on,' highlighting the needs debt of $61,223 and that employers vidualized company platform, employ- for wellness and work life initiatives more who offer to help an employee alleviate ees and management can routinely give than ever before. With so many Baby some of that debt are much more likely kudos and recognition at the moment it’s Boomers retiring, we are left with a gap to attract higher caliber talent. Their due, not just at annual reviews or occa- in our workforce overall. This, combined website, studentloangenius.com, gives sional company gatherings. Employees with an incredibly low unemployment individuals various payment options get to share each other’s successes and be rate, makes this a candidate’s market. recognized for their achievements in real Employees today want more flexibility, Employees and manage- time, and on a public platform. There’s and the opportunity to contribute their even a points system that translates into abilities to society in meaningful ways. ment can routinely give monetary, charitable, or other fun team Combine these things with compa- rewards, incentivizing peers to recognize nies trying to contain costs, more and kudos and recognition each other. more employees are going shorter term at the moment it’s Studies do show that, all other things or even independent and it eventually being equal, potential job candidates leads to what some have referred to as due, not just at annual will choose a company that has a culture the “gig economy.” This perfect storm is reviews or occasional aimed at bettering employees and that making attraction and retention critical shows an interest in their current and metrics for many organizations, and we company gatherings. future well-being over one that doesn’t. are having to get more creative in how And in today’s tight job climate, that we approach both." to help them pay off their debt faster. could make all the difference in a compa- Gone are the days of people work- But, more creatively, they also work with ny’s future. So, while being able to have ing at one job until they retire, being employers to tie payments into payroll, a Pajamas Day might be a hip thing to loyal to the company no matter what. taking the onus off of the employee to do, it’s not going to pull in that program- Today’s workforce is looking for ways to juggle sometimes multiple payments. , , mer three other companies are clamoring make a difference not just at their job And where employers once used 401(k)s or over. But explain to her the rewards and but in the world. They want to know pensions as perks, they can use services recognitions platform you have in place that their contributions are given the like Student Loan Genius and others like or let her know how much you value recognition they deserve and that their it to match an employee’s loan payment, her current and future contribution to employers are invested in their personal or to tie in a matching payment to the the team by helping pay off college debt growth and continued well-being. And employee’s 401(k) when that student or provide continuing education? Now as Ms. Sterling points out, technology contributes to his own loan payment. – you’ve got the competitive edge. now allows employees to be always on According to a 2012 Study by 8 COGNITIVE TIMES Oct. 2016 FOUR FINANCIAL TECHNOLOGY TRENDS TO WATCH IN 2016

The technology behind driverless cars, blockchain and messaging apps will find it’s way into finance this year. by Ari Studnitzer *First published on openmarkets.com on Feb. 19, 2016 2016 has certainly started off with a bang with increased volatility and uncer- tainty of future demand. It’s hard to focus let alone think about key technology enablers for 2016 in the face of the macroeconomic events. But with uncertainty, comes new opportunity for innovation.

Last year, I wrote about the importance of customer experience, security, democrati- zation (of technology), and data science. Like most technology themes, it takes more than a year to fully implement and optimize enabled business value, and those 2015 highlighted themes are no different. As we look ahead to 2016, one theme remains on the list with security and the increased scope of compliance.

One noteworthy trend that continues to expand is mobility. Much has been written over the years around the need for a new mindset of mobile platforms and design- ing first for mobile. While not a new theme, 2016 is projected to be the first year since mobile platforms were introduced where CMEGroup.com will have more than 50 percent of its traffic from mobile platforms. Just like security, mobile is going to continue as a critical, global theme and one that will accelerate in lock step with accelerating innovation in mobile platforms.

SECURITY & COMPLIANCE LEDGER TECHNOLOGY

Last year, I wrote that we’re in an increasingly insecure How could I write about 2016 without referencing block- world when it comes to technology. Unfortunately, that chain and distributed ledger technology? While many hasn’t changed and if anything, has continued its trajectory. became skeptical of the hype cycle that encircled all things CME Group continues to focus on its own technology and ledger-based, I see tremendous opportunity for enhanced infrastructure along with continued maturity of the broader efficiency and the enablement of new capabilities. connected ecosystems. Firms should not look at ledger technology for its own Many new technologies are being introduced that sake, but in the context of use cases and solving broader enhance security with new integrated machine learning ecosystem problems. For example, CME Ventures invested techniques. I talk about machine learning further below, in Digital Asset Holdings who is focusing on syndicated and it’s exciting to see new combinatorial innovations being loans as their first set of use cases. There’s no doubt this too driven. As an example, CME Ventures invested in Forts- is a long-term theme that will see continued innovation and cale. Fortscale uses machine learning for user behavior anal- course corrections for the coming years. ysis. I expect to see continued, especially combinatorial, innovation to further enhance security and compliance at financial firms.

10 COGNITIVE TIMES Oct. 2016 PRODUCTIVITY & COLLABORATION TOOLS MACHINE LEARNING 2016 will be the year exciting tools like Jive, Slack, and Yammer become enterprise hardened and available for use Last year I wrote about data science as an emerging tech- in compliance-focused firms. Facebook is also focusing nology with machine and deep learning. The underlying its priorities on the application of its technology for the technology and applicable use cases are exploding where enterprise. As an industry, finance has long been focused 2016 could be the year commercial, externally-facing on email as a means for communication and collaboration. applications with machine learning start to emerge in the I think we all agree any form of communication dependent financial services industry. CME Ventures has invested in on using capital letters to display emotion is likely not an SparkCognition and Nervana who both provide very excit- optimal collaboration forum. ing platforms. Between driverless cars and the need for I see many use cases that can be solved with the new greater automated monitoring, it’s clear machine learning breed of collaboration software. If you’re not already looking is here to stay. at next generation tools, you may find yourself quickly feel- 2015 was a year of many exciting technology innova- ing like you’re using an Apple Newton. More importantly, tions, and we don’t see a slowdown in 2016. These are a few increasingly your colleagues will be looking to leverage these technology themes to keep in mind in the months ahead. next generation tools and prospective hires will be measuring firms by this capability. COGNITIVE TIMES Oct. 2016 11 EVENTS

Tech Talk 5.0 Next-Generation Financial Technology July 5

CEO Amir Husain was invited to speak at CME Group's Tech Talk 5.0 in London. He spoke about the potential applications for Deep Learning in the financial industry and how SparkCognition's technology could help shape the way we trade and invest in the coming years.

BlackHat July 30-Aug 4

Black Hat is the most technical and relevant global information security event series in the world. For more than 18 years, Black Hat has provided attendees with the very latest in information security research, development, and trends in a strictly vendor-neutral environment. These high-profile global events and trainings are driven by the needs of the security community, striving to bring together the best minds in the industry.

N.I. Week Aug 1-Aug 5

N.I. Week brings together the brightest minds in science and engineering, with more than 3,200 leading innovators representing a wide variety of industries. For the N.I. Week Panel, SparkCognition joined other industry experts in the field of machine learning to discuss how advanced software technology is helping Big Analog Data problems and the Industrial Internet of Things (IIoT) N.I. Week Panel

American Assoc. of Drilling Aug 24 Engineers Monthly Lunch and Meeting

The American Association of Drilling Engineers (AADE) is a non-profit, volunteer organization. AADE offers a forum for the exchange of information, among its members and guests, specifically on drilling related topics. AADE chapters generally meet once a month where programs are presented by knowledgeable industry leaders in an informal luncheon or dinner environment. SparkCognition was invited to offer our insights into how machine learning can solve complex O&G problems and improve efficiency. 12 COGNITIVE TIMES Oct. 2016 PANcomm 03:30 PM - 16 Aug 2016 Kirk Borne 10:53 AM - 16 Aug 2016 @PANcomm @KirkDBorne #Millennials #FinTech How do you reach with ? What's driving the amazing rise of #fintech? @Cybzy shares how: http://bit.ly/2bkOfz0 via http://shar.es/1ZMM5d #IoT #blockchain @pancomm #BigData #Analytics #Cloud

CMEGroup 3:13 PM - 14 Jul 2016 @CMEGroup Rumi Morales, our Exec. Director of Strategic Investments speaks at the @MarketsWikiEdu intern series today

IMFahad 12:49 PM - 22 Aug 2016 SparkCognition 10:57 AM - 13 July 2016 World Economic Forum @thefadkhan @wef @SparkCognition 2:33 AM - 23 Aug 2016 The smartest minds in #FinTech talk about how Our CEO, @amirhusain_tx speaking @CMEGroup #WallStreet will change http://read.bi/2b1RGzo #CMETechTalk 5.0 on Cognitive Analytics for These are the world's #fintech hubs http://wef. #blockchain #AI #Tech #fintech #Techchilla Financial Platforms ch/2buGg31

AP Guha 8:41 AM - 13 Jul 2016 @APGuha

@amirhusain_tx, CEO speaks on cognitive analytics for finance...next frontier for SparkCognition, @CMEGroup TT5.0

A.I. IN SOCIAL MEDIA

Tom Grant @thomaslgrant 9:33 AM - 23 Aug 2016 #BOJ may apply ' #fintech ' to operations in future: Kuroda - Greater competition and transparency @Reuters http://goo.gl/eyD3jx #Japan

Wall Street Journal @WSJ 8:08 AM - 23 Aug 2016 BOJ encourages fintech, remains wary of threat from hackers http://on.wsj.com/2bRws5W

CFO Essentials 9:05 AM - 27 Jun 2016 @CFOEssentials #ArtificialIntelligence Finds Its Place in Finance #AI http://buff.ly/298Ki0M

Oliver Bussman 2:15 AM - 9 Jun 2016 IBM Partners 7:20 AM - 24 Aug 2016 @obussman Michael Krigsman 9:37 PM - 20 Aug 2016 @IBMPartners @mkrigsman #LinuxONE News: IBM launches cloud services The #FinTech Threat is Real: 28m Households are Robo-advisors have a $2 trillion opportunity in front for #blockchain on industry’s most secure server Ready to Move to a More Innovative Bank of them #fintech http://ibm.co/2bnd6GE http://americas.nttdata.com/news/news . . .

COGNITIVE TIMES Oct. 2016 13 Finding Financial Exfiltration Using SparkSecure®

by Jerry M. Schirmer, PhD

ne of the scariest scenarios for any company that is exposed to the internet is the lingering threat of data exfiltration. All of your data, from passwords and usernames, to actual trade secrets and market strategies are contained O on computers. If these computers are going to be able to communicate with each other, whether to conduct trade-floor execution, strategic discussion amongst humans, or for myriad other reasons, it is necessary that they be on internet-connected devices. But if a device is on the internet, then it is vulnerable from a security perspective. And when compromised, the data exfiltration may happen silently, with the breach only discovered when the leaked information has been acted upon. This is where SparkSecure® can actively come to work.

1 Put a filter between your device and the outside world

2 Build a list of the possible ways that compromises can happen

3 Generate a set of blocking measures that can be automatically implemented

4 Set your filter to execute this set

5 Hope for the best Knowing the 'normal' behavior through the system, SparkSecure® can look for known instances where the system has been attacked or compromised.

The problems here are obvious. Most notably, they are in a manner consistent with these various profiles, or even static. System administrators are constantly updating in a way that is simply anomolous. And as the exploits their list of vulnerabilities, and the actual people doing change, so do the models. So, when a rogue trader stops the exfiltration just need to be conversant about what is connecting to their normal set of sites, and uploads market- being looked for in order to avoid the known traps. This ing strategies via seldom-used-by-them FTP, the system is where SparkSecure’s Profile Based Threat Detection can identify it, and flag them. system comes in. Rather than look for static signatures For instance, take the recent Swift messaging breach and blocking them, Profile Based Threat Detection looks that affected the Bangladeshi Central Bank, amongst over all of the logs that have come in over the life of the other Southeast Asian institutions. This hack was possible SparkSecure® deployment. From here, it generates a set of because the internal systems that used Swift were config- 1 statistics . Using these statistics, we can split the “normal” ured to trust all traffic once they had gotten through the behavior of each user or IP in the system, dynamically. authentication system. Had SparkSecure® been running Then, knowing the “normal” behavior through the against the outbound firewall logs of the Swift system, system, SparkSecure® can look for known instances where the system would have been able to flag the illicit transfer the system has been attacked or compromised. in near real-time by detecting a new IP connecting to the From here, the system is capable of building up a profile system, initiating new, anomalous traffic. This would have of bad behavior–whether the user is installing malware, more of a chance of success than the pure signature-based visiting compromised sites, or exfiltrating data, SparkSe- methods of the past, because it would be calibrated, specifi- cure® constructs a profile. Then, when new data comes in, cally, to Swift’s traffic, which makes the detection of anom- that new data can be checked against the existing model, alous behavior more consistent and less prone to the false and in close to real-time, users can be flagged as behaving positives that break security experts’ confidence in a tool.

1 i.e., how many times has this user connected to port 80 in the past half hour? How many distinct URIs have they requested per second?

DeepArmor and the Future of Cyber by Keith Moore

odern-day antivirus has become the equivalent of the Windows '98 avatar Clippy–you see it all of the time, but you’re not really sure how it helps you. In fact, over 78% of security professionals no longer trust antivirus simply because it is not able to keep M up with the state-of-the-art in malware.

Things like Domain Generated Algo- DeepArmor® is powered by cutting- rithms and obfuscation have far edge technology that is light-years surpassed the detection capabilities of ahead of those used for malware gener- the signature-based tools that we all use ation or propagation. Pulling from today. This is the bad news. many patented, patent-pending, and Even worse news is that cyber crime proprietary SparkCognition automated is growing. According to the Singapore model-building techniques, DeepArmor® Minister of Home Affairs, Law Shan- starts by looking at every unscanned file mugam, an estimated $2 Trillion will on a user’s desktop or laptop. It breaks be lost through cybercrime by 2019. each file into thousands of different This is a recipe for disaster, and the pieces for initial review. It then elevates major reason why both state and federal the features using an advanced feature governments are making cyber security derivation algorithm known as SparkAr- the top priority. temis to turn those thousands of pieces Instead of looking at To combat this growing problem and into tens or even hundreds of thousands technological deficiency, SparkCogni- of individual components. All of these static signatures, or tion is proud to release the first cognitive individual components are then run antivirus solution to the market, Deep- through cognitive and adaptive neural even exploding files in Armor®. DeepArmor® takes a unique networks to find patterns that may be malicious in nature. Because these ® approach to endpoint protection by a sandbox, DeepArmor leveraging neural networks, advanced cognitive neural networks are trained heuristics, and complex data science to on a bevy of threat types, from Worms looks at the DNA of find and remove malicious files. Instead to Ransomware, any threatening patterns of looking at static signatures, or even present will be unearthed and called out every file to identify exploding files in a sandbox, DeepAr- immediately. mor® looks at the DNA of every file to DeepArmor® has been tailored to if any components are identify if any components are suspicious operate seamlessly behind the scenes on or malicious in nature. DeepArmor® uses each endpoint, and to only identify real suspicious or malicious cognitive algorithms to constantly learn threats without calling out false posi- new malware behaviors and recognize tives. This gives any user the freedom in nature. how polymorphic files may try to attack to do what they would like without the in the future! This keeps every endpoint fear that their computer may become safe from malware that leverages infected. domain-generated algorithms, obfusca- tion, packing, minor code tweaks, and many other modern tools. Better yet, this is the perfect defense against terrorizing Zero-Day threats, which can never be captured using existing tools.

16 COGNITIVE TIMES Oct. 2016 Atlas Wearables: The Personal Trainer That Fits On Your Wrist by John King

illian Michaels is one of the leading health and wellness experts in the world. She is a television personality, entrepreneur, life coach, nutrition and wellness consultant, motivator, and a New York Times best-selling author. Jillian is a multi-millionaire personal trainer who is best known for her appearance on different television Jshows such as Losing It with Jillian, The Doctors, and The Biggest Loser. Michaels is an expert who truly knows our attempts to beef up or slim down. degrees. Atlas Wearables’ co-founder, her field. Like Jillian, the best of the The hourly cost of a personal trainer Mike Kasparian, was working at best know their field better than their can vary from around $50 an hour, Philips Healthcare while Peter was peers and colleagues. They do this to upwards of hundreds per hour. researching at Johns Hopkins. by immersing themselves in their In Jillian Michael’s case, you’d prob- It was through those experiences industry, by going deeper than others ably need to pay her thousands to that they were exposed to the core are willing to go. They read libraries work with you. problem of advanced machine learn- of books and reports, they talk to as We pay experts because they know ing for attracting 3D motion and how many industry veterans as are willing, what they’re doing. They have spent it could be applied to the fitness world. they understand the history of their more time understanding the data of At the time, Peter and a friend had industry, and they pay attention to the their industry than anyone else and started a program called the Sound present trends of their trade. they can process that data into action- Body Challenge. It was a three month In essence, they immerse themselves able insights, which produce results. program, very similar to other moti- in the data of their industry. They take Experts are paid at the highest levels vational challenges that are common- in as much data as they can find, and in their field because they understand place today: an individual would come process it. The most successful people and apply the knowledge (data) they in once a week, maybe on a Saturday, in any industry understand that data have acquired in the most effective and a personal trainer would take them and are able to see the future through and efficient methods possible. We through a standard set of exercises and it. Whether Warren Buffet in finance, pay them to utilize their knowledge, routines, grade their form, and give Oprah Winfrey in media and enter- incorporate the data we create, and them feedback to improve. tainment, Tony Robins in behavioral process that into actionable insights What Peter and his team began to psychology, or Jillian Michaels in that will make us better and produce notice was those individuals who could personal fitness, industry leaders see the results we desire. compel themselves to come in every what is coming and act faster and more Like human intelligence, gathering Saturday would be much more moti- creatively than others. They observe the and analyzing data to understand the vated to have better outcomes. data and compute it into action. best path forward is a key element Through gauging intrinsic motiva- According to Marketdata Enter- of machine learning today. And like tion (self-motivation), extrinsic moti- prises, a market research firm that human intelligence, through data, vation (a prize or reward for comple- specializes in tracking niche indus- machine learning can find the stron- tion), and social responsibility present tries, Americans spend north of $60 gest indicators of causality. in fitness teams, iterative testing and billion annually on fitness and weight With a background in biomedical observation demonstrated the vast loss. With CNN Money putting the engineering, Peter Li, Founder and potential to implement intrinsic moti- median salary of a personal trainer at CEO of Atlas Wearables, was exposed vation in a digitized workout regimen. $56,000, with top pay of $128,000 to many different startup ventures After their initial experience with the (in 2012), it’s likely that a significant within fitness, healthcare, motivation Sound Body Challenge, Peter got on portion of that $60B goes to paying during his time at Johns Hopkins as the phone with Michael and through a professionals and experts to assist us in he earned his bachelors, then masters series of Google Docs exchanges, they

18 COGNITIVE TIMES Oct. 2016 launched Atlas Wearables. motivation factors built into Atlas THE TECHNOLOGY THAT SETS Peter, Michael, and their team technology, and how might that look started with an idea to help motivate to the regular user? ATLAS WEARABLES APART FROM people in the same way a personal Intrinsic motivation is the core of OTHER WEARABLE TECHNOLOGY trainer or fitness challenge program their technology. Atlas is focused on would, and they built an augmented augmenting the real-world experience IS THE MACHINE LEARNING personal trainer for your wrist. of a personal trainer and client–move- TRACKING CAPABILITY. ESSEN- Now in an advanced iterated version, ment + motivation. the Atlas Wristband2 is the first work- Like a trainer, Atlas will keep track TIALLY, YOU'RE ABLE TO PUT out tracker that learns to track your of different types of goals and metrics THIS DEVICE ON, DO JUMPING form for various exercises. It can auto- for you. It's sort of like a GPS; it will matically recognize most exercises and provide a roadmap to where you want JACKS OR ANY OF THEIR 100+ even help you target Training Zones to go, even if you want to drive some- EXERCISES, AND ATLAS WILL optimized to burn fat, tone, or bulk. It’s where you've never been. You just a waterproof device worn on your wrist, put your GPS on and you don't even KNOW THAT YOU'RE DOING THAT with a smartphone app and online think twice. SPECIFIC EXERCISE. portal to sync your personal data. By looking at the Atlas wearable The technology that sets Atlas model, it's clear the Atlas team have Wearables apart from other wearable dared to do things differently. Pedom- technology is the machine learning eter technology has made a strong ize their wearable to the nth degree. tracking capability. Essentially, you're name and case for itself in the fitness Peter tells Cognitive Times that the able to put this device on, do jumping market, especially alongside the rise of data collected by the Atlas wristband jacks or any of their 100+ exercises, and wearable step-counting fitness brands is “only the surface” of the potential for Atlas will know that you're doing that like Fitbit. future expansion. specific exercise. So pushups, trian- Right now, the market is beginning As is the case with machine learning gle pushups, military pushups, Atlas to see a trend towards consumers who technology, the development of Atlas’ can tell the difference between even have used a pedometer for a while and technology is iterative. They start with the most incrementally differenti- are now beginning to look for more, regular people (who represent and ated movements. as they are realizing that there may be provide data), and move up from there. Atlas uses motion sensors on the more to their activity than just the step Although the product might be great wristband: a three-axis accelerom- count. Atlas looks ahead to a fitness for professional athletes, Atlas would eter, and a three-axis gyroscope. If horizon where a wearable is just as rather build it for the masses. you imagine a paint dot on your wrist informative to the user as that user is Anyone concerned with fitness moving around in 3D space, Atlas to the wearable. would love to work with the world’s tracks that path, and each motion has In the constantly evolving intersec- top fitness experts. However, very few a definitive fingerprint that they're able tion of fitness and technology, devel- of us can afford to. Yet, that may be to identify. It's similar to what makes opers are often faced with expansive changing as we move toward machine Siri run, but instead of looking at voice, problems and no one clear solution. learning that can augment what has Atlas is looking at a completely differ- The success of Atlas is attributable historically been relegated solely to the ent dataset-motion. How does Atlas to its unique design, both inside and realm of human expertise. differentiate between movements? out. With its “Tetris-like” look, the At the end of the day, Peter and his Machine learning. interactive elements present in its user team want to be able to have the exper- The problem other wearables have experience, and the presence of high- tise an industry veteran like Jillian (those not utilizing machine learning) level machine learning elements to Michaels offers. They want to be able is in the similarity of movements. A produce a truly unique experience, the to capture her expertise and apply it pushup and power clean look noth- Atlas wristband truly is one-of-a-kind. to your unique goals, movements, and ing alike, so they have no issue. But As the needs and wants of consumers motivations. By leveraging advanced what about a normal pushup versus a change, Peter foresees many oppor- technologies and machine learning, triangle pushup? Visually, as a human, tunities to expand into solution-ori- they’re well on their way to building they look very similar with the slight- ented tweaks. it onto your wrist. est variation. Atlas Wearable picks up Atlas is already collecting a huge To pick up an Atlas wristband, check on those variations. amount of data to best optimize out Atlas Wearables on Amazon, at So Atlas has applied machine learn- consumer experience, and the wear- Dick’s Sporting Goods, Luke’s Locker ing to elevate precision beyond that able’s loyal community embraces in the Austin area, or at www.atlas- of other wearables. But what are the plenty of opportunities to custom- wearables.com.

COGNITIVE TIMES Oct. 2016 19

MACHINES ARE PREDICTING What You’ll Buy by Thu Le

our favorite retailer already knows what you’re going to buy next. Customer spending data are goldmines of intelligent insight. Typically, a company will look at how you spend to determine how you are likely to spend in the future, and thus, how to target their Ymarketing to your likely purchases. All of this is made possible through predictive analytics. COGNITIVE TIMES Oct. 2016 21 First of all, it’s important to note that Dallas published a study that shows most predictions are based on the how companies can target consumers assumption that consumers spend based based on the amount of money they on their interests. Financial behaviors have already spent elsewhere. This model that companies consider are the types requires a new kind of methodology--in- of products customers are purchasing, stead of looking at how a consumer buys when they purchase, and where they are at their own stores, companies can now spending their money. Having access to examine share-of-wallet, or the amount this kind of information allows compa- of a consumer’s spending in a defined nies to improve their value proposition product category that a business captures. by personalizing product offerings and For instance, if a customer spends a total delivering them to you when you’re most of $100 on toiletries every month, with likely to need or want them. $60 of that amount at Walmart and the The story of Target predicting a teen- other $40 at Target, the share-of-wallet age pregnancy before her father knew for each company would then be 60% has become a classic case study among and 40% respectively. retailers. Target was sending the teen- This method makes sense as it also ager coupons for baby clothes and cribs looks at a consumer’s total size of wallet, a few months before she was due, while which means how much they can afford her father was completely unaware of for a specific product category alone. If the pregnancy. Based on the daughter’s a consumer owns a wallet size of $100 updated spending and browsing behav- for toiletries per month compared to iors, Target knew she was algorithmically those who can only afford $50 for the more likely to be a pregnant shopper. same products, then Target and Walmart However, there are only so many would want to put more efforts into the things retailers can do until a consumer former customer. The same goes for moves on to his or her next major life banks. A customer can either decide to Target was sending the event - getting married, buying a new put 100% of his or her savings into a house, or getting pregnant. That’s when bank, or split them at 50/50, or whatever teenager coupons for baby the consumer’s usual routine breaks, proportion that they’d like. If a bank can and retailers need to adjust. To account entice the customer with a certain size of clothes and cribs a few for this, retailers have statisticians like wallet into putting more of their savings Andrew Pole, who created what he into that bank’s system, then it subse- months before she was called a ‘pregnancy-prediction model’ quently gains a higher level of profit. for Target. This model was built from However, getting competitive data can due, while her father was an existing database associated with be challenging as it isn’t always available. their customer ID numbers which Without accessible data, companies can completely unaware of the included their historical transactions turn to predictive models to gain an through credit cards, coupons used, estimation of their customers’ spend- opt-in survey responses, emails, demo- ing. Most of the data feeding into these pregnancy. Based on the graphics, and various other types of data. models originates from past surveys Using these data, the system can predict or information aggregators, as well as daughter’s updated spend- which females are most likely to be preg- historical transactions and approxi- nant, and Target can start placing ads mated customer average income within ing and browsing behaviors, and offerings for an array of maternity a demographic or geographic segment. products to those customers, while not In addition, interrelationships between Target knew she was algo- making it intrusive. product spending behaviors can also The act of following a consumer’s be used to facilitate cross-promotional rithmically more likely to be spending pattern does not occur solely efforts. If customers tend to spend more on the retail floor. on soaps after buying shampoos, but not a pregnant shopper. Recently, The University of Texas at quite the reverse, then companies might

22 COGNITIVE TIMES Oct. 2016 want to promote shampoos more heavily. through simple messaging. The bot can match your interest, but also know when Using consumer data for retail is one also show you bargains that you might you need them most. thing, but aggregating life patterns is be interested in based on your spend- For such a long time, banks, retail- quite another. Banks in this modern era ing history and geography (for exam- ers, and companies have retained a are in a tremendous position of power ple, deals from the exact store you’re fluid amount of data streams that have to understand, predict, and exploit standing in). gone untapped. With Artificial Intelli- customer trends. Furthermore, even the banks that gence, machine learning, and predictive There is a lot of value that a bank can you are connecting with can tap into analytics advancements, companies can derive from your monthly bank state- your data and offer additional banking now put these data into use–not only ments. For example, Kasisto developed services when it makes sense. By comb- to increase profits, but also to improve KAI, an Artificial Intelligence based ing through your statements, credit processes, and in turn, benefit their conversational platform with an ‘exper- scores, reviews, and many other streams customers. tise’ in finance, to help consumers with of resources, banks can start offering managing money and tracking expenses deals and saving options that not only COGNITIVE TIMES Oct. 2016 23

Maintaining Century-Old Trust with Cutting-Edge Cognition

by John King

According to a recent Accenture to regain and maintain the centu- tional technologies in this new era. report, global investment in finan- ries-old trust in financial institutions. As the financial sector continu- cial technology (fintech) ventures The 2008-2009 financial crisis ally evolves, trust is becoming more in the first quarter of 2016 eachedr started an unprecedented regula- important than ever. Technologies $5.3 billion, a 67 percent increase tory response across the globe with like blockchain are providing trust over the same period last year. At a focus on risk management. It through a transparent and feder- the same time, consumer faith in had wide-ranging implications for ated digital ledger system. Artificial financial institutions has been low financial services including technol- intelligence is providing next-gener- for years. This loss of confidence has ogy and the response to regulation. ation cybersecurity and surveillance the potential to eventually undermine Regulatory, consumer, and cost pres- solutions to reduce risk and curtail financial growth. However, the rise of sures have meant that capital market illicit activity in financial markets. fintech provides new opportunities firms needed to look beyond tradi- In a world where trust is being asso- ciated with uncrackable crypto algo- rithms and artificial intelligence, the role of trusted banks and institutions is evolving, and for the third-larg- est stock exchange in the world, The London Stock Exchange Group (LSEG), trust is being maintained by applying cutting-edge technology. Chris Corrado was working at IBM as a programmer when he was approached by a recruiting firm specializing in placing mathemati- cally minded personnel in financial services firms. They were looking for a quant. If you’ve read the book, “Flash Boys”, you know that large financial organizations, in addition to spending big on the best hardware, have been scooping up the very best computer science and data analytics talent for some time. But it would be almost 30 years until Michael Lewis would write “Flash Boys” –this was in the mid-80’s. Morgan Stanley had already seen the writing on the wall. Corrado joined Morgan Stanley as a programmer when it was exclu- sively a capital markets firm. But he rose through the ranks to eventually become Morgan Stanley’s CIO of International IT. After a decade of covering appli- cations development and infrastruc- union of finance and technology is market surveillance and post trade ture in Europe and Asia for Morgan blended into a very effective gradient, systems for over 40 organizations Stanley, Corrado served in CTO propelling his industry to new heights and exchanges globally, including or CIO roles with Deutsche Bank, through streamlined technology and the Group’s own markets–London Merrill Lynch, AT&T, eBay, Asurion, a client-oriented experience. And Stock Exchange, Borsa Italiana and UBS, and MSCI. Corrado is putting his experience to Turquoise,” Corrado explains. “We Banks know what they need. And work, leveraging the latest innovations are building trust through this resil- what Corrado offers is at the top of to ensure that LSEG, which main- ient exchange technology and an their list. No surprise, then, that once tains a $6 trillion short scale market operational focus on clearing and the he had amply proved his mettle, the capitalization, remains one of the quality of indexes we provide as well as global financial icon, London Stock most trusted financial organizations platforms for efficient and high-quality Exchange, brought him on as their in the world. trade reporting.” COO and CIO. At LSEG, Corra- So, just how, specifically, is Chris An important facet of the drive for do’s mission includes providing access Corrado redefining trust in an industry change is understood by paying atten- to capital formation, intellectual enveloped in transformation? tion to the underlying drivers. The rate property, and risk and balance sheet of change in fintech is nothing short management. “LSEG is a leading developer and operator of high-performance tech- of breakneck. While industry trans- For a leader like Corrado who is nology solutions, including trading, formations in finance are often driven always focused on innovation, the by endogenous factors, they are more COGNITIVE TIMES Oct. 2016 27 often brought on by specific exoge- companies are a vital part of the finan- oping new products will significantly nous catalysts. And in considering the cial services sphere–they complement reduce risk and margin requirements global transition to a risk-based focus and support the ecosystem in achiev- while delivering the opportunity for in finance, post 2008, this trend proves ing efficiencies and cost reductions,” he deeper regulatory oversight.” true: Corrado isolates the 2008-2009 says, referring to leverage of technol- One of LSEG’s unspoken goals also financial crisis as a chief driver under- ogy to mitigate losses while supporting includes providing and protecting the lying a rethink of the notion of risk as the link between clients and partners relationship of trust between all play- it existed before the crisis. and engendering trust. ers within its federated network. As In a paper prepared for the 2010 With great hopes being tied to a leading developer of high-perfor- Financial Crisis Inquiry Commission new technologies such as the secure, mance technology solutions for over by the Booth School of Business at the distributed ledger system, blockchain, 40 organizations and exchanges glob- University of Chicago, the problem of Corrado is focused on leveraging what ally, Corrado tells Cognitive Times poor risk management was apparent in is working, albeit with caution. that LSEG is committed to a collab- actions leading to and evident through “Blockchain is an early stage tech- orative, open source industry effort to the crisis. Research and insights nology that has the potential to drive advance distributed ledger technology, acquired by subpoena revealed that change across the industry. We are “as a trusted institution with expertise in many financial firms, the “preferred excited about it but believe that the in post trade and risk management we explanation for why bank balance technology needs to be developed in feel we have a lot to offer the exciting sheets contained problematic assets, a considered and rigorous manner, in debate that is currently taking place.” ranging from exotic mortgage-backed partnership with clients, to provide With security being front and center securities to covenant-light loans, the right service and benefit to them,” in LSEG’s operations, Corrado refer- is that there was a breakdown of Corrado says. “Amongst other initia- ences the importance of building incentives and risk control systems tives, we are contributing to a market- secure and dependable relationships, within banks.” wide steering group comprised of and applying the most cutting-edge The opportunity to learn from the clients and other partners to look at available technology. crisis and re-orient finance came the key challenges and opportunities “This is the future,” he points out. about in the aftermath of 2008. Lead- that blockchain and distributed ledger “Essentially, we operate in a sector ers looked for opportunities to grow technology more generally offers.” that can attract bad behavior and it is without committing the mistakes that London Stock Exchange Group getting worse. To live safely, we need had led to the previous disaster. An has significant technical expertise to to proactively predict what undesir- eye for innovation was instrumental bring to the discussion and Corrado able behavior could happen next, and in ensuring a positive growth curve sees real opportunity for the tech- prevent it from happening and/or to on the way forward–and innovation is nology in trading, particularly in risk ensure that this behavior cannot impact what Corrado lives for. management. As has become clear us, or our clients. That is our responsi- “Innovation is essential to service with the meteoric rise of the fintech bility, and to do so requires advanced clients’ needs, so startups and fintech sector, “driving innovation and devel- analytics applied through a variety of WITH A CLEAR UNDERSTANDING OF THE CAPABILITIES AND SCALABILITY THAT A.I. BRINGS TO FILTERING OUT SCAMS, SPOOFING, AND OTHER UNWANTED ACTIVITY, CORRADO BELIEVES, “THIS IS THE FUTURE.”

machine learning algorithms.” using machine learning can detect ing with the underlying power of Security is becoming more import- unique activities or behaviors (“anoma- computing platforms (hardware, large- ant than ever. Risk management falls lies”) and flag them for security teams. scale data analytics infrastructure), on the shoulders of automated systems, The challenge for these systems is to Corrado indicates that an investment with higher-level validation and inves- avoid false positives–situations where in providing the necessary infrastruc- tigations becoming the primary duty “risks” are flagged that were never risks ture to enable “that level of scalability” of analysts. System monitoring is a in the first place.” is certainly worthwhile. tedious responsibility where poten- To stay ahead of technological “There are a number of areas where tial for human error and exhaustion is trends, LSEG has been working AI can offer significant benefits,” he vast. When technology takes on the with SparkCognition, a leading arti- tells us. “Cybersecurity comes to mind: job of cogency and mitigating risk– ficial intelligence startup, to develop predictive analytics around maintain- whether in fraud, error, or otherwise– next-generation surveillance solu- ing resilient operations, improving the burden of security becomes a little tions. With a clear understanding of the quality or our financial data. Ulti- easier to manage. the capabilities and scalability that AI mately, this affords us the opportunity In handling partner-client rela- brings to filtering out scams, spoofing, to make our people smarter and to tionships as well as monitoring the and other unwanted activity, Corrado make decisions faster.” transition of capital, Corrado says it believes, “this is the future.” Humans will perhaps always be is imperative for modern-day finan- Always honoring the trust between needed to provide strategic guidance, cial firms to maintain security as a top LSEG, their clients, and the market and bear the ultimate responsibil- priority in mitigating the potential at large, security is top of mind for ity, but the efficiency enabled by AI threats brought on by high levels of Corrado. Understanding, avoiding, has proven instrumental in allowing risk. In order to maintain the highest and preventing cybersecurity and human analysts to provide improved levels of security, Corrado, as he has fraud “is our responsibility and to do services and support ever-more-com- throughout his career, is staying ahead so requires advanced analytics applied plex risk management needs. Algo- of the technology before the technol- through a variety of machine learning rithms, with their extensive exposure to ogy gets ahead of his organization. algorithms,” states Corrado. Machine the inner workings of such a complex Predictive analytics in machine learning technologies address the system make the perfect information learning and artificial intelligence question of speed and efficiency by “sponges” that learn, adapt and imple- have been instrumental in scaling the isolating system anomalies and learn- ment security and trust at machine security component of financial risk ing continually from existing financial scale for large organizations such as management. The speed and effective- activity–whether internal or external LSEG. With much more to come in ness of monitoring risk is extremely to a group’s digital infrastructure–this the AI revolution, and AI-accelerants important to large digital infrastruc- technology ensures a baseline of secu- like quantum computing on the hori- tures such as those within finance. rity previously unattainable by human zon, Chris Corrado is making abso- As the publication, TechEmergence, standards alone. lutely certain that the only direction reported earlier this year, “systems By coupling AI and machine learn- risk management and financial security has to go is up. COGNITIVE TIMES Oct. 2016 29 THE ONLY TRACKER THAT LEARNS EVERY EXERCISE

www.AtlasWearables.com AUSTIN STARTUP SPOTLIGHT

Greener Futures: Honest Dollar, Fintech, and the Fish That’s Eating the Whale

by John King and Caroline Lee

COGNITIVE TIMES Oct. 2016 31 e met Whurley on a Friday advocating for open source, or rubbing successfully acquired company, Honest afternoon at WeWork, one elbows with Presidential candidate Dollar, Whurley continues to have an Wof the national co-working Hillary Clinton. outsized impact on the industry he’s franchises that have exploded in Whurley started his career in Austin chosen to displace. popularity over the past few years. at Apple, then went on to IBM as a When it comes to the charge Housing individual consultants, Master Inventor. It was then that he for starting a company like Honest entrepreneurs, and startups, the office got the startup bug. He has seen success Dollar, Whurley makes it clear that buzzed with activity. and failure as an entrepreneur, yet has the company was a problem-oriented Yet the Honest Dollar founder– never stopped creating. He is the author response to a clear need. “There’s a despite his own packed schedule, from of two books and various articles, savings crisis in America,” he explains. conference calls to addressing a deluge holds over ten patents, fundraises for “If you’re an average American citizen of emails armed only with a mobile political causes, has founded multiple and a $400 unexpected expense arises, keyboard–was cool, completely at peace. non-profits for social good, and has it could potentially bankrupt you.” Whurley, a successful serial been a producer for two-award winning This number isn’t arbitrary–in January entrepreneur and inventor, along with films. All of this while launching of this year, Princeton Survey Research his most recent startup, Honest Dollar, companies and building game- Associates International conducted fit in well with the hopeful, bright- changing technology. a survey that reported only 37% of eyed, and in many cases inexperienced, His first really successful venture Americans have enough savings to entrepreneurs at WeWork. The was a mobile software design and cover a $500 or $1000 expense, such as difference is, Whurley is leading a development company. In 2010, along an emergency pet surgery or a washing fintech startup which was acquired with Ben Lamm and Mike Erwin, machine repair. The other 67% resort within one year of launch by global Whurley founded Chaotic Moon to spending cutbacks (23%), charging financial behemoth, Goldman Sachs. Studios. Chaotic Moon developed expenses to a credit card (15%), or Whurley, who goes by what was the first iPad-only digital newspaper, borrowing funds from friends or family originally his Unix username, an built applications for Microsoft, CBS (15%). In addition to this, the survey abbreviation of his full name, William Sports, Sanrio, Pizza Hut and others, found that 4 out of 10 Americans Hurley, is famous. He started his career developed products such as a Microsoft surveyed ran into an unexpected major as a touring musician–“as everyone Kinect-controlled skateboard (which expense last year, which is a testimony now knows,” he points out with a smile, the company named the “Board of to Whurley’s point: the savings crisis “because someone put it on Wikipedia.” Awesomeness”), a shopping cart that is a problem that has existed without He never went to college, and he’s follows a shopper around the store, being properly addressed for some time accomplished more by the age of 45 and a bicycle helmet fitted with seven now. “And that should worry everyone, than most do in a lifetime. If you follow cameras that begin recording on impact, right?” him on social media, you’ll see him functioning like an airplane’s black box Interestingly, the problem became hanging out at the White House for in case of a hit-and-run (the “Helmet of more complex as it was considered state dinners, visiting with President Justice”). And in July of 2015, Chaotic further–a fact that doesn’t come as a Obama, leading The Global Partnership Moon was acquired by Accenture. surprise, especially given the historical for Gender Equality in the Digital Age, Sitting at the helm of another focus of current personal finance

32 COGNITIVE TIMES Oct. 2016 research. Unemployment statistics are Individual Retirement Accounts (IRAs) Honest Dollar in May holds potential often too insular to provide insight into designed for the way people actually live for innovation beyond any current the how and why of personal finance, and work. Retirement accounts are set offerings for the financial services even though the two may be linked. up for an industrial era where workers industry. “Think of it as a financial According to David Johnson from stayed with the same company for most company that has been transforming the University of Michigan, “people of their lives, contributing equally to itself into a tech company, and is now have studied savings and debt, but a long-term savings account. Today, becoming a software company,” he this concept that people aren’t making two years seems to be a long tenure concludes decisively. ends meet or the idea that if there was for younger generations. Saddled with Given the track record of Honest a shock, they wouldn’t have the money college debt and stagnant wages, these Dollar’s development, it’s likely that to pay, that’s definitely a new area of younger generations are forced to put Whurley plans to make the most of research”–especially since the Great off savings for longer periods of time. Honest Dollar and Goldman Sachs’ Recession. In early September, The Pew partnership, but he’s not yet saying how. Of all the potential savings-related Charitable Trusts released an analysis Historically, SXSW has spelled major spheres of the financial industry, the of access and participation in retirement announcements for Honest Dollar. “The area of focus that Whurley landed on savings plans that found Americans idea is to continually top everything, was retirement. “I look at retirement on average haven't actively saved for right? This year at SXSW will be the as kind of the Achilles' heel of the retirement. The reasoning for this answer to questions like “how have financial system,” he says. “There are so observation envelops a variety of they fared?”, “did the big company many areas of our lives that are touched catalysts: from enrollment variations crush them?”, or “did they make the and dominated by finance. And what based on industry and region, to access big companies super innovative” – but do people do? They choose not to differentials based on age, gender, between now and then, it will be a think about it.” This became a root education, and ethnicity. whole series of new people that we cause analysis that would eventually Honest Dollar, Whurley tells us, aims brought on, the humans we brought on, lead to Honest Dollar’s inception as to simplify the process and generate and creative, innovative projects making a web-based way to make retirement wider access to retirement under the their way into the public sphere.” investing easy and accessible to customer’s own terms. The company’s From mere observation of the everyone, allowing individuals to create recent acquisition by Goldman Sachs, company’s past track record paired with and manage their own IRAs. announced at SXSW 2016, will ideally the enthusiasm in Whurley’s words, it Honest Dollar’s initial charge was achieve this goal even faster than before. seems like Honest Dollar has its sight to find insight that would better When asked about working with set on rocking fintech with no sign of answer these questions. Several of the Goldman Sachs, Whurley is quick to stopping. Though he wouldn’t dare give essential inciting topics included a basic answer with two questions of his own. away any teasers as to what to expect understanding of financial behavior: if “Are we now a part of Goldman?”, from Honest Dollar next, Whurley people aren’t thinking about their future leaning back with a smile. “Or is did, with the same cool he carried financially, what are they thinking Goldman now a part of us?” His point walking into the room, leave interested about? holds true – the knowledge exchange spectators one important piece of Essentially, Honest Dollar offers incited by the company’s acquisition of advice: “Just wait and watch.” Where No Derivatives Company Has Gone Before: Rumi Morales and the Future of Fintech by John King and Caroline Lee

“When you think about a bank, or Ventures, Rumi Morales is charged with tems are giving way to new technolo- when you Google one, what’s the image the execution of venture investment in gies and digital renovations: “If you go that comes up?” emerging technology worldwide, so she here today,” Morales begins, showing “Usually a Greco-Roman temple. But knows a thing or two about the history an image of a massive floor at Stamford when’s the last time you went to a bank and trajectory of fintech. Morales grad- UBS teeming with trades, “most of this that looks like that?” uated magna cum laude from Wellesley is shuttered for a number of reasons - Ripples of laughter move through the College and has an MBA from NYU. whether it’s the increase in regulations, audience before she moves to her true She’s worked with personnel and port- the pretty banal trading environment question: “When’s the last time you went folio companies around the world, we’ve seen so far, but also because of to a bank?” including the Global Markets Institute this.” She switches to a slide featuring In a talk delivered to the MarketsWiki at Goldman Sachs, and international a mosaic of different financial services Education (MWE) World of Opportu- corporate development strategy with the apps displaying trending projections nity Education Series, Rumi Morales CME Group. Morales is such a mover rising and falling on a digital dash. “You demonstrated a charge to rethink ways in and shaker in her industry that she was don’t need all the hardware since we’re which the Chicago Mercantile Exchange named to Crain’s Chicago Business 2014 becoming increasingly mobile.” (CME) Group has engaged with the Tech 50, Institutional Investor's 2015 In thinking about the future, Morales financial ecosystem of the future, and Fintech Finance 35, and Crain's Chicago isolates the importance of recognizing to go where no derivatives company has Business 2015 40 Under 40 lists. the origins and importance of a basic gone before. Rumi Morales knows banking. need when considering the future of a While speaking to the audience of She knows fintech. And according to solution. The trajectory of the futures young professionals in Chicago, Morales Morales, the growth of online banking market through history is no stranger to established the trajectory of financial and mobile transactions are just two of this concept: the world’s first recorded technology, or fintech, through its early the factors relegating brick-and-mortar futures exchange was organized in and most well-known iconography. banks to relics of the past. Japan in 1697. The process behind such As Executive Director of CME Even the most iconic trading ecosys- a market has undoubtedly grown in

34 COGNITIVE TIMES Oct. 2016 complexity throughout the last three digital age, this change happens with the ability to process what’s happening hundred years, but the concept itself - more haste and expanse than ever before. around you. much like the best ideas, Morales asserts Morales pauses at a point in her timeline Much like a seasoned trader, a - harkened to a need, and that need was to examine the transformation of CME computer is able to achieve each of to protect and manage capital. Group’s famed Merc Club into an Inno- these prerequisites to effectively learn Digitization, advanced security vation Lab, designed to foster collabo- the markets and offer predictive anal- systems, and the next generation of big ration and the exchange of ideas. “You ysis to rival human experts, while still data are the three areas that Morales think about the evolution of finance over mitigating human error and completing

believes have the power to transform time and how things continued growing, onerous tasks in a fragment of the time. the financial services industry landscape. and you can become humbled because “We’re at a stage today where comput- “We feel that these are the three things things can change.” ers have access to far more historical data that not only keep us up at night because But the changes don’t stop there. One and analysis than we can hold in our we’re scared,” she says with a smile, “but of the most pervasive technologies of heads,” Morales explains. “Computers they keep us up at night because we’re the day is artificial intelligence, and it’s are getting a feel of the market now.” excited.” Fintech, like many major indus- quickly making its way into the finan- And much like the neural network of a tries today, is an amalgamation of many cial industry. human being, the “feeling” computer is elements combined to create a well-run- Morales highlights artificial intel- able to problem-solve in ways previously ning machine. ligence as a key player in the future of unimaginable. This level of critical think- The interest that CME Ventures keeps the derivatives market, since a change to ing is “not about machines becoming in mind while seeking out new areas automation could usher in new potential better machines,” the executive director of innovation parallels the mission of for growth, resources, and development. concludes, “but instead about machines emerging technologies themselves: how She gives an example from SparkCog- becoming better brains.” might one improve an existing system to nition’s CEO, Amir Husain, in under- An ongoing dialogue between the better address a need and allow for more standing the intersection of a trader’s feel needs of corporations and the solutions accessible future innovation? for the market and a computer’s ability of the startup world has occurred on a In an industry built on change, to recognize patterns. Posing a question resoundingly silent level. Rumi Morales Morales sees each of CME Ventures’ to the audience, Morales challenges the believes that establishing a dialogue will partner groups as an opportunity to group to consider what comprises the allow more solutions like artificial intel- propel fintech into a more secure future; expertise within the overlap: “If you’re a ligence and machine learning to affect but most of all, one that channels more really good trader and have a feel for the the derivatives industry in focused, perti- effective communication between people, market, where does that feel come from?” nent areas that will drive the financial finance, and technology. The answer that Husain offers, services sector into the digital age, full New technologies have historically Morales says, is multifaceted: experience, speed ahead. ushered in change. In an increasingly a knack for recognizing patterns, and Political Forecasting and Predicting the 2016 Presidential Election

by Ashvin Govil e are in the midst of an election season that a result, we have to turn to sources of data from the present, has rewritten the way politicians view the rather than from the past, to try predict the future. Whistory books, shifted how Americans look The most well-known political forecaster actively predicting at their fellow citizens, and created seemingly unsurpassable this year’s election is Nate Silver, thanks to his past successes rifts between them. However, regardless of whether a voter in predicting elections. After being named one of the World’s is riding the Trump Train, feeling (felt) the Bern, or is with 100 Most Influential People in 2009 for correctly predicting Her, one basic curiosity looms at the back of their mind: who’s the results of 49 out of 50 states in the 2008 presidential elec- going to win? tion, he went on to predict the 2012 election without missing It’s a question that has long drive people in political science a single state. and statistics to find models, polling methodologies, and Nate Silver’s models mostly use poll data to calculate the patterns in history to create models that can accurately and probability of each candidate winning, as laid out in these steps: consistently predict the results of presidential elections. More Step 1: Collect, weight and average polls. recently, people in computer and data science have started to enter the field, wielding new tools, such as neural networks, Step 2: Adjust polls. machine learning, or advanced statistical analysis, to harvest Step 3: Combine polls with demographic and (in the case of polls- new kinds of information from the same old data. plus) economic data. Predicting an election is not an easy task. It requires fore- Step 4: Account for uncertainty and simulate the election thou- telling the beliefs and the intentions of well over 100 million sands of times. people, months in advance, well before many of them have even made up their minds on who they will vote for. The most important factor here is the averaging of the polls. The first step in creating an accurate prediction model Media outlets love to report outlier polls showing one candi- involves choosing the appropriate data source. In a standard date or another up in the polls by a surprising amount, imply- election, political forecasters like to use historical economic ing that something significant has changed about the election data to predict the outcome, often with good results. For based on just one poll. example, in 1988, Michael Dukakis led in the polls against However, due to the uncertain nature of polling, it is George Bush Sr. for much of the campaign season. But models expected for individual polls to sometimes be wildly apart considering the strong economy at the time and the fact that from one another, in the same way you would expect that in a Bush was the Vice President of a popular incumbent president, random group of people, some of them are much taller than Ronald Reagan, predicted that Bush would eventually come others. Averaging polls helps eliminate these statistical errors up on top and win the election (which he did). and allows for predictions that have a smaller margin of error But of course, there is hardly anything that could be consid- than the polls on which they are based. ered “standard” about this election. However, even Nate Silver’s models, and therefore the polls, Thanks to a full year that has thoroughly broken all tradi- have had trouble making sense of this election. tional theories on political science and is beckoning an entirely This entire election cycle has been wrought with endless new era in American political history, predicting based on the scandals, controversies, and leaks draping negativity over both past has suddenly become much less reliable. candidates. Public opinion polls have shifted rapidly as news Due to the historical unpopularity of both major candidates cycles shift from one candidate to the other, one scandal to the for president, these traditional avenues of prediction have been next, giving a very noisy picture of the race over time. deemed inapplicable to this cycle, even by their creators. As

Nate Silver’s Polls Plus predictions from June 8th through October 12th COGNITIVE TIMES Oct. 2016 37 Nate Silver’s Polls Only model, which doesn’t account for convention bounces or economic factors, has had an even bouncier cycle.

This isn’t a fault of Nate Silver or his models, but rather the nature Through the power of the financial incentives created by these of the poll data they rely on. Rapid-speed news cycles combined websites, prediction markets offer a powerful way to calculate with the abnormally high number of undecided and third-party the true odds of an election. They operate as data aggregators in voters that will eventually support one of the mainstream candi- similar fashion to Nate Silver’s models in order to create predic- dates creates a huge amount of ripples in public opinion and tions. However, instead of crunching raw numbers to create the the polls. probabilities of each candidate winning, they are powered by thou- But what if we want a crystal ball that looks past all that volatility sands of people crunching a combination of data, news events, and and makes a concrete prediction of the future that we can trust personal perceptions. now, and two weeks from now as well? These markets tend to be more stable than polls or poll-based markets. The graph below taken from PredictWise, a website PREDICTION MARKETS that aggregates prediction markets into actual predictions, shows just how stable these markets are even in the face of multiple Prediction markets are these peculiar online websites where thou- conventions, scandals, and blunders taking up the news cycles sands of traders place bets on elections or other political events for weeks on end. in the same way a gambler might bet on a certain horse winning a horse race. Betfair and PredictIt are two of the more popular prediction markets.

Compared to the data from Nate Silver’s Polls Only model over the time frame, these predictions are practically flat for much of the race, until both started converging towards Clinton after a series of leaks and accusations of sexual assault dramatically hurt Trump’s standing with women voters in the last few weeks. So now we can predict the election in a relatively consistent manner (prediction markets), and directly see the beliefs and opinions of the public (polls). But still, both polls and prediction markets have their flaws. Polls are slow to show changes in opinion, prone to bad methodology, and are not very useful individually. Prediction markets are able to update their odds in real time and are able to stabilize themselves despite large changes in the polls. But at the end of the day, they don’t offer any actual new data, merely a new option to consolidate what we already have into an actual prediction. What if we wanted to find a new data source that could tell us something that history, polls, and prediction markets could not? The answer is simple–social media. The internet has exploded in Social media platforms also provide more simple ways to view the last few decades, with the last ten years bringing the rise of data about certain candidates that can be predictive. The graph social media. Even since the first election of President Obama in below shows how the number of Facebook likes Bernie Sanders 2008, the percentage of adults who use at least on social media site had in a particular state corresponded to his vote share in that has skyrocketed from 25% to 65%, and social media grew from state’s primary or caucus, with a surprisingly strong correlation. being just another distraction in our lives to an extension of our personalities. A useful side effect of all this growth has been an exponential increase in the amount of data people create, share, and enjoy with each other using social media websites such as Google, Facebook, and Twitter. Political discourse wasn’t left out of this growth, making social media has a potentially ripe source for real time data on how people think and feel about certain politicians at any given time. Although social media’s demographical bias towards younger people makes it a challenge to extrapolate trends to the general population, it can provide a window into how people’s thoughts, beliefs, and actions are changing in real time in a way polls never could. The tricky part is teasing out the data from the unorganized mush of words that makes up what people say on the internet--people https://twitter.com/tylerpedigoky describe their feelings using words and links to articles mirroring their opinions, rather than the discrete and easy-to-understand data a poll or a prediction market provides. Despite the advantages of using social media as a data source, This is where the technology and data analysis experts come in. hardly anyone has been using it to make meaningful predictions. With the use of cutting-edge natural language processing (NLP) With the continued emergence of new and robust data sources techniques such as sentiment analysis, a sequence of words in a such as social media, breakthroughs in cognitive algorithms such tweet, Facebook post, or Reddit comment that would otherwise as Natural Language Processing and Machine Learning, and be meaningless to a computer can be converted into a discrete increased computing power, it is now possible to create predictive numerical representation of how a certain person feels about a models that approach problems with more complete perspective, certain candidate based on that post. This representation could and ultimately, human intelligence at machine scale. then be aggregated and fed into a model that could show changes For now, we’ll continue to follow Mr. Silver’s work while pushing in how candidates are perceived over time by tracking changes in the limits of cognitive security and predictive analytics in energy, the average tone used about that candidate on social media. The security, and finance. data could even be trained to predict the results of new polls that are still being conducted, by training a model through supervised machine learning to predict them based on how older social media data correlated to old poll data. Google Trends, a service provided by Google that aggregates and displays the relative popularity of any given search queries over time, provides a different lens on the same concept. Instead of seeing what people are saying, Google Trends shows us what people are trying to learn. When a large news event about one candidate happens, search queries about that candidate tend to increase proportionately to the magnitude of the event. The graph below comparing search trends of Donald Trump (red) and Hillary Clinton (blue) shows this in effect--queries of Hillary Clinton surged as the FBI investigation surrounding her private servers wrapped up, and interest in Donald Trump similarly spiked during the Republican Convention.

TM Congratulations to the UT Austin Computer Science Department on five decades of leadership in academia, innovation, and research. The top-ten nationally ranked UT Computer Science Department is a global leader in the development of scientific knowledge and applications that are defining our technological future.