The Last Iceberg – How Artificial Intelligence Is Unlocking Humanities Deep Frozen Secrets

Tip of the iceberg 90839Icebergs are common meme used throughout the Internet. You see them everywhere, from depicting social media to human behaviors. They are used to explain knowledge we know, above the surface, and those things we don’t know, below the surface. Icebergs are interesting. They’re secretive. The ten percent we see is the literal tip of what is possible. Below the waterline, just out of sight, are dark secrets. Secrets that are often out of the reach and unusable. That is, until now.

Artificial intelligence (AI) is changing our the way we live. AI is doing more than just helping us find patterns in data or helping us make better decisions, it’s truly unlocking unexpected insights and extending our knowledge in ways that only humans are capable of doing… capable of controlling. Prior to AI, human beings had to use their minds to harvest knowledge from everyday life events. It’s a hard process. A process that required countless hours of dedication just to discover one new meaningful insight that can lead to massive improvements in our lives. But this is now changing as we begin to rely on new cognitive technologies that generate knowledge for us…knowledge without us.

Iceberg Breaking Slowing Global WarmingAI is melting those data icebergs. In essence, it is becoming the global warming of the knowledge age. It’s unlocking their deep hidden secrets. AI is unleashing those hidden insights, producing more knowledge that is then used to melts even more icebergs. It’s an exothermic knowledge activity and that is exponentially generating more insights than used in discovery process. And herein lies a devastating, potentially life ending, problem.

As we rely, over rely, on new cognitive technologies, we lose our ability to discover new knowledge ourselves. The brain as an organ and capabilities are lost when not used. Take for example the Slide Ruler. Most people today do not know what a slide ruler is, let alone how to use one. This simple mechanical device, seen in the hands of most engineers hands in the 1970s, can perform amazing mathematics. With just two opposing rulers, one can do addition, subtraction, multiplication, division, logarithms, square roots, n-roots, and more. There’s nothing that can’t be mathematically done on the slider. It requires not batteries, no internet connection, and does not fail. It is brilliant in its complex simplicity. But today nobody knows how to use it. Why?

Figure1 Multiplication C DFor the slide ruler example, we have lost this cognitive ability as a society because it has been outsourced it to other systems, like the calculator and the spreadsheet. These are new productivity tools that were invented to help us more efficiently unlock knowledge. But the cost of using them is that we’re now no longer capable of exercising a part of the brain that used to physically discover insights through this mechanical manipulation. Artificial intelligence is now accelerating this kind of cognitive decay.

As humans rely more on AI to discover knowledge, we are slowly losing our own cognitive ability, our own mental capacity, to discover those insights ourselves. Our brain cognitively weakens. AI is in essence creating a mental defect in our executive functioning processes. Unchecked over time, we will become over dependent on AI to identify those new things that will lead to a better life. Eventually evolving to a point where we could literally die without this AI ability. Or even die because of it.

Main qimg 05a329fbabebe9e44945b8a336201176 cThis uncontrolled release of knowledge can be a destructive chaotic process. We see similar outcome with uranium, for example. With the right equipment one can control how neutrons are absorbed in uranium isotopes, producing a stable reaction which generates life-giving energy. left unchecked, however, the same neutron interacting with the same isotopes can produce devastating nuclear events. Controlled reactions lead to life, and control reactions lead to death.

Can humans survive the chaos of a world where AI is unlocking more knowledge that humans can handle? A future world where available knowledge is greater than the questions we can ask? Physics tells us we cannot. History shows us it is unlikely. AI unchecked, ungoverned, can be the nuclear weapon that we use on ourselves that will eventually melt not only every last iceberg, but society itself.

Bringing Artificial Intelligence To Life

artificial-intelligence-recruitment.jpgArtificial Intelligence is on the verge of a premature extinction, unless we dramatically change the way we bring to market its abilities. The new goal of any organization should be to bring Artificial Intelligence (AI) to Life –  it is hard to do, but simple when done. Life is defined in terms of real problems being solved, express through actual uses cases, and not in the technology of their solutions. Uber has brought AI to life through self driving smart cars. Siri, Google Now, and Cortana have brought AI to life through digital personal assistants. Spotify, Pandora, and Netflix have brought AI to life through helping us enjoy music and film art in a highly personal way. All are living examples of AI that have real impacts on our daily lives.

BN-OY980_0718_c_J_20160718174929.jpgHowever, today artificial intelligence if often overly complicated by characterizing it in terms its underlying capabilities and technologies. Capability like machine learning, natural language processing, and robotic process automation are frequency points of discussion with consumers. When talking about AI, practitioners often invoke describe it in terms of genetic algorithms, neural networks, and evolutionary programming. While these capabilities and technologies accurately reflect the inner complexity of what makes artificial intelligence naturally hard, one still needs to bring AI to life in a way that simplifies our daily lives.

eabd7cf57b3a47e007d9c961dfb6152d.jpgWe are in the midst of a intelligence revolution that, by its definition, is destine to change our lives. Like the farmer being replace by the factor work being replace by the service worker, our lives will become more meaningful only when AI is as prolific as air. So, we need to bring AI to life by hiding the complexity that makes it hard, while transparently illuminating all the ways our lives become more simplified because it. It is only then when we will evolve to our next logical level of enlightenment.

Darkness, A Flashlight, and the Data Scientist

What you don t knowData sciences and data analytics not only use different techniques, that are often highly dependent on the distribution characteristics of the data, but also produce very different categorical types of insights. These insights range from a better understanding things you know you know (data analytics) to discoveries in area where you don’t know what you don’t know (data sciences). However, this knowledge metaphor can be a bit confusing, so I often use the “Darkness, A Flashlight, and the Data Scientist” parable.

Flash Light

In your mind, picture a darkened room, where you are standing, but do not know where in the room you are. In your hand is large flashlight. You raise it slowly, pointing it in a direction. You turn it on and white light radiates forward.

The light of the flashlight shines brightly on a distant wall, where you see several items. These are the things you know that you know. As you your eyes begin to scan outward, the wall turns to deep dark dark black where the light does not reach. In this darkness, there are things you don’t know you don’t know. You begin to look back into the cent of the light – that grey transitionary boundary between the light of what we know and the darkness of the we don’t know, are all the things we know we don’t know.

Singularity4

Data analytics is lot about understanding those things we know we know, that is quantifying the light. This is the world of descriptive and diagnostic analytics. On the other hand, data sciences help use understand the darkest parts of our world, where we look to predict temporal and spatial relationships  and prescribe means for achieving desired outcomes. Data analytics and sciences are different in their own ways, each very important in their own right.

However, in the case of the data scientist, the metaphorical role is to pull the flashlight back so that more areas of the wall are illuminated. So, as the flashlight is linearly pulled back, the data scientist enables an exponential increase in our knowledge. In essence, the data scientist works in the dark so that others can benefit from the light. Think about it!

Critical Capabilities for Enterprise Data Science

NewImageIn the article “46 Critical Capabilities of a Data Science Driven Intelligence Platform” an original set of critical enterprise capabilities was identified. In enterprise architecture language, capabilities are “the ability to perform or achieve certain actions or outcomes through a set of controllable and measurable faculties, features, functions, processes, or services.”(1) In essence, they describe the what of the activity, but not necessarily the how.While individually effective, the set was nevertheless incomplete. Below is an update where several new capabilities have been added and other relocated. Given my emphasis on deep learning, composed on cognitive and intelligence process, I have added genetic and evolutionary programming as a set of essential capabilities.

2015 03 04 10 52 16

The Implementation architecture has also be updated to reflect the application of Spark and SparkR.

2015 03 04 10 53 13

Deep Learning Intelligence Platform – Addressing the KYC AML Counter Terrorism Financing Challenge

NewImageTerrorism impacts our lives each and every day; whether directly through acts of violence by terrorists, reduced liberties from new anti-terrorism laws, or increased taxes to support counter terrorism activities. A vital component of terrorism is the means through which these activities are financed, through legal and illicit financial activities. Recognizing the necessity to limit these financial activities in order to reduce terrorism, many nation states have agreed to a framework of global regulations, some of which have been realized through regulatory programs such as the Bank Secrecy Act (BSA).

As part of the BSA (an other similar regulations), governed financial services institutions are required to determine if the financial transactions of a person or entity is related to financing terrorism. This is a specific report requirement found in Response 30, of Section 2, in the FinCEN Suspicious Activity Report (SAR). For every financial transaction moving through a given banking system, the institution need to determine if it is suspicious and, if so, is it part of a larger terrorist activity. In the event that it is, the financial services institution is required to immediately file a SAR and call FinCEN.

The process of determining if a financial transaction is terrorism related is not merely a compliance issue, but a national security imperative. No solution exist today that adequately addresses this requirement. As such, I was asked to speak on the issue as a data scientist practicing in the private intelligence community. These are some of the relevant points from that discussion.

2014 12 16 21 38 08

Determining if a transaction is terrorism related, requires more that analyzing the anomalous nature of the activity, but the correlation of seemingly unrelated signals (profiles, transactions, interactions, etc.) through behavioral analyses.  Data (enterprise, IT, open source) is the historical debris of human activity. While any single data record is associated with one person, two physical independent events can be found through the causal behavioral analysis of data chains.  

2014 12 16 20 26 13Know Your Customer (KYC) is a common means through which one can learn about structures and behaviors of each individual in a community (e.g., commercial banking, insurance, etc.). It is the governing program through which customer due diligence is performed as part of compliance activities associated with on boarding and on going monitoring activities. 

2014 12 16 20 12 24

Over the years, through ongoing regulatory additions and changes, KYC has grown in complexity and, as a result, has become a significant multifaceted challenge to institutional employees. In additional to knowing about customer,  there is now a need to know more about the customer’s customers (KYCC). There are significant deficiencies  associated with determining propensity (probably), intelligence, and monitoring activities; even though most organizations are adequately dealing with a few of the ingestion, processing, and reporting activities.

2014 12 16 20 13 04

There are six major components to an effective know your customer program. Terrorism Financing Monitoring is one of the least mature and the hardest technically to solve. Traditional approaches encode simple transactional behaviors found through manual investigations into rules engines and event monitoring systems, an approach that does not scale as fast as the terrorism financing activities they are designed to defeat. 

2014 12 16 20 13 43Money laundering (ML), as defined by the United Nations, is the process through which the proceeds of criminal activities are disguised to conceal their origins. Fundamentally, money laundering is about financial structure (where) and behavior (how). The Financial Action Task Force (FATF) has established international standard for ML monitoring and reporting.

2014 12 16 20 14 31

While the mean through which money is laundered is beyond the scope of this presentation, there are several concrete examples that have been discovered as part of an ongoing money laundering ontology. The High Invoicing Scheme is often used to launder licit funds through commercial business enterprises by exchanging low value goods for high value illicit funds.

2014 12 16 20 15 07

Terrorist Financing (TF) involves the solicitation, collection or provision of funds with the intention that they may be used to support terrorist acts or organizations. In addition to understanding the structure and behavior of financial sources, understanding their intended use is also necessarily. This “intent” is one of the characteristics that make identifying terrorism financing so difficult.

2014 12 16 20 15 58

Terrorism financing and money laundering are interrelated. In money laundering, funds are always illicit in their origin, where funds for terrorism financing can come from both legal and illicit sources. Because of the dual funding source and the intended use of the funds, it is extremely difficult to identify whether financial activities are related to terrorism financing.

2014 12 16 20 16 33Below is a set of real account, transactional, and international profiles. Are they normal? Are they an example of money laundering? What about terrorism financing? In additional to answering these questions, would traditional ML and TF monitoring systems identify each activity or tie them together? The answers are at the bottom of this article.

2014 12 16 20 17 14

A wide variety of Anti-Money Laundering products are available today. At a baseline level, AML systems automate mandatory legal and regulatory compliance requirements and support the necessary enhanced due diligence and Know Your Customer policies.

2014 12 16 23 03 48 

Use cases in Risk are centered around connecting all business and financial information systems to enable enterprise regulatory, monitoring, and reporting requirements in order to further better risk decision making. Identify fraudulent behavior before it happens, with proactive intelligence and investigation tools, that are all capable of operating across multiple channels and nations.

2014 12 16 23 05 19

Data and intelligence analysts, as well as KYC AML & TF specialists, face an exponentially increasing challenge to thoroughly identify new customers and monitor all customer behaviors on a ongoing basis.

2014 12 17 08 38 01What is the new TF intelligence paradigm given the global regulatory requirements, the maturation of terrorist, the complexity of financial services information technology systems, and the national security imperative to find, fix, finish (exploit, analyze, and disseminate) terrorism actions pre-boom? It starts with the recognition that tradition enterprise (ERP, CRM, etc.) and IT (transactional logs, click through, etc.) data sources are insufficient. Additional data deep web and open source data needs to integrated into the analyses as a means identify networked behaviors.

2014 12 16 20 20 11 

In addition to new data sources, man and machine need to be integrated into a deep learning enabled ecosystem. Modeling the behaviors of bad guys is often counter productive, given their speed of adaptation. A more viable approach leverages modeling good guys and removing them from the target population under investigation. Machines automate this process of removing good behaviors from the system through black list aggregation and human guided machine learning algorithms. Intelligence experts perform enhanced investigations through Human, Physical, and Cyber Intel programs. All of these activities are wrapped in deep learning machines that learn from those highly utilized behaviors, driving the search from new data source and intelligence procedures.

2014 12 16 20 20 37

The new enterprise solution delivers (outside the box) the identity of bad people and organizations, behavioral activities, FinCEN SAR filings, and xml integration into the banking enterprise. In order to achieve these outcomes, banking enterprise and IT data, 3rd party black lists, and deep web and open source data is consumed. Bank AML and TF experts work in conjunction with Data Science, Behavioral, and Intelligence teams. As part of an enterprise learning system, the intelligence results are feedback into the platform as a means through which knowledge is grown.

 2014 12 17 09 36 26

Ienterprise architecture language, capabilities are “the ability to perform or achieve certain actions or outcomes through a set of controllable and measurable faculties, features, functions, processes, or services.”(1) In essence, they describe the what of the activity, but not necessarily the how. For a data science-driven approach to deriving insights, these are the collective sets of abilities that find and manage data, transform data into features capable of be exploited through modeling, modeling the structural and dynamic characteristics of phenomena, visualizing the results, and learning from the complete round trip process. The end-to-end process can be sectioned into Data, Information, Knowledge, and Intelligence.

2014 12 16 20 21 57

Data science is much more than just a singular computational process. Today, it’s a noun that collectively encompasses the ability to derive actionable insights from disparate data through mathematical and statistical processes, scientifically orchestrated by data scientists and functional behavioral analysts, all being supported by technology capable of linearly scaling to meet the exponential growth of data. One such set of technologies can be found in the Enterprise Intelligence Hub (EIH), a composite of disparate information sources, harvesters, hadoop (HDFS and MapReduce), enterprise R statistical processing, metadata management (business and technical), enterprise integration, and insights visualization – all wrapped in a deep learning framework. However, while this technical stuff is cool, Enterprise Intelligence Capabilities (EIC) are an even more important characteristic that drives the successful realization of the enterprise solutions needed to address the emerging KYC ML and TF threats.

2014 12 16 20 22 31

Terrorism financing came into the limelight after the terrorist attacks in the United States on the 11 September 2001. Global anti-terrorism programs, now manifested themselves through nation state regulations such as the Bank Secrecy Act, can be more effective through the use of deep learning ecosystems that integrate both machine and man. This is one such platform capable of achieving this goal. 

Post – The financially related transactions above where those associated with the 9/11 terrorists in 2001.

46 Critical Capabilities of a Data Science Driven Intelligence Platform

NewImageData science is much more than just a singular computational process. Today, it’s a noun that collectively encompasses the ability to derive actionable insights from disparate data through mathematical and statistical processes, scientifically orchestrated by data scientists and functional behavioral analysts, all being supported by technology capable of linearly scaling to meet the exponential growth of data. One such set of technologies can be found in the Enterprise Intelligence Hub (EIH), a composite of disparate information sources, harvesters, hadoop (HDFS and MapReduce), enterprise R statistical processing, metadata management (business and technical), enterprise integration, and insights visualization – all wrapped in a deep learning framework. However, while this technical stuff is cool, Enterprise Intelligence Capabilities (EIC) are an even more important characteristic that drives the successful realization of the enterprise solution.

2015 02 04 08 50 01

In enterprise architecture language, capabilities are “the ability to perform or achieve certain actions or outcomes through a set of controllable and measurable faculties, features, functions, processes, or services.”(1) In essence, they describe the what of the activity, but not necessarily the how. For a data science-driven approach to deriving insights, these are the collective sets of abilities that find and manage data, transform data into features capable of be exploited through modeling, modeling the structural and dynamic characteristics of phenomena, visualizing the results, and learning from the complete round trip process. The end-to-end process can be sectioned into Data, Information, Knowledge, and Intelligence.

2014 11 08 14 10 45

Each of these atomic capabilities can be used by four different key resources to produce concrete intermediate and final intelligence products. The Platform Engineer (PE) is responsible for harvesting and maintenance of raw data, ensuring well formed metadata. For example, they would write Python scripts used by Flume to ingest Reddit dialogue into the Hadoop ecosystem. The MapReduce Engineer (MR) produces features based on imported data sets. One common function is extracting topics through MapReduced programmed natural language processing on document sets. The Data Science (DS) performs statistical analyses and develops machine learning algorithms.  Time series analysis, for example, is often used by the data scientist as a basis of identifying anomalies in data sets. Taken all together, Enterprise Intelligence Capabilities can transform generic text sources (observations) into actionable intelligence through the intermediate production of metadata tagged signals and contextualized events.

2014 11 08 14 21 11

Regardless of how data science is being used to derive insights, at the desktop or throughout the enterprise, capabilities become the building block for effective solution development. Independent of actual implementation (e.g., there are many different ways to perform anomaly detection), they are the scalable building blocks that transform raw data into the intelligence needed to realize true actionable insights.

Deep Learning – Being Almost Human

NewImage

Deep Learning is the application of unsupervised learning paradigms to machine learning capabilities in order to identify, extract, and leverage features leading to insights that enable actionable outcomes. While computer science is a relatively young discipline, Deep Learning is even a young discipline within this field.

Deep Learning is one of the more remarkable resources available to research and practitioners in this emergent field. The goal of the website is to host a variety of resources and pointers to information about Deep Learning. In these pages you will find:

  • reading lists of relevant article and books,
  • links to software and tools,
  • datasets and databases,
  • a list of deep learning research groups and labs,
  • a list of announcements for deep learning related jobs (job listings),
  • as well as tutorials and cool demos.

NewImage

For example, under the “Recent Post” section, there is an interesting article “An Article about History of Deep Learning.” The author, Daniela Hernandez, describes how deep learning has been an disciplinary outlier in the academic field for the last 3 decades. Emerging from this intelligence slumber, most practitioner are merely “fashioning artificial neural networks that mimic at least certain aspects of the brain.” While not much of an advancement in unsupervised learning, it nevertheless has extended our ability to creates intelligence systems that scale with minimal human interaction.

Like deep learning itself, this is a relatively young site that should mature over time, becoming a key resource in the data science field.

SaveSave