Penumbra – The light and the shadow
Cognitive Computing stands as the single greatest disruption of our global society since the Industrial Revolution. Where the maturation of the Industrial Revolution created the magical gains in productivity that yielded the delivery of mass goods across the country and then across the world, the age of cognitive computing stands to amplify, augment, and multiply the productivity in ways that would have been unimaginable just a few years ago.
At the same time, that previous shift rapidly displaced global labor forces that were agrarian and yielded a different society- one where sunset no longer brought an end to the day but ushered in an age of electric night. It created rapid wealth and even more rapid destruction. It raised life expectancy immensely yet caused environment and health consequences that we are still grappling with. And it left a very different set of nations and individuals in positions of power throughout the world. Their values have dominated the last century and have prevailed – though not without great cost and challenge.
The values of freedom, self-determination, societal safety nets, and increasing equality of rights have made great gains, if still unevenly, over that Industrial Century. The power of Cognitive Computing and the great effects it will have could be every bit as influential on this century. Therefore, the question of our time is who will control this great power and what will they do with such a responsibility. Before tackling that question, I should probably begin by describing what we believe Cognitive Computing is and how it is used and could be used in the future.
What is Cognitive Computing? From Digital Data to Digital Reasoning
So what is Cognitive Computing?
It is the next wave of computing and software. Someday all software will learn and that day is nearly upon us.
Cognitive Computing is the platform that comes, necessarily, to deal with a society overwhelmed by the risks and possibilities of Big Data in an Internet-connected, mobile-device-oriented world where individual cognition and limited attention is no longer enough to cope. It has roots in artificial intelligence – in many ways it is the completion and extension of taking algorithms from complex automation to self-learning, and eventually generally intelligent agents with autonomy.
Today, with great computational effort and some human guidance, it automates the transformation of masses of data (literally millions or billions of human messages) into a coherent form (what we call the Knowledge Graph) that can be used for applications that make curiosity cheap and monitoring for known risks at global scale possible. It harnesses well-formed and digitally-organized data (called “structured data”) and human oriented data composed of text, images, and sounds (called “unstructured data”) into a knowledge representation (i.e. the Knowledge Graph) through a variety of integrated algorithms and storage mediums (very much like our brains). It reads, resolves, and reasons over data in a way that is increasingly like us. It learns. And it therefore gets better over time and with more data. It works at scales that transcend our human limitations and can be made more powerful by simply adding more hardware and storage. There are several groups working to make this Cognitive Computing future a reality. It is no longer a question of if this is possible, but how much of it is today and how will it extend further tomorrow.
The foundations of Cognitive Computing are algorithms that can find order in unordered information. Fundamentally, that is the ability to associate similar things and create segmentation of dissimilar things. This objective capability of distinction and doing it on a continuum allows it to create groups from raw data. The fundamental group in any representation is the alignment of real things that have different representations –i.e. real entities with multiple names – to be resolved into sets. These sets of multiple representations of real entities that have properties allow for another level of conceptual hierarchy to develop, namely categories. These categories can be as specific as people who have played the striker position in the Barclay’s Premier League in 2013 and extend as broadly to all people who live in the United States.
The creation of categories allows for operations to extend and generalize over abstractions of huge groups (i.e. properties and relationships) – making computation and thinking much more efficient than simply applying the same transformation or operation against each individually. Humans excel at creating general operations and rules of thumb (called heuristics) that extend over very broad groups and running though complex experiments (like “how will this audience respond to this product”). This has also been something that machines have been able to do for some time now, but only when the data was well structured and when the structure of groups was provided a priori. The high quality, high scale, and high throughput capability to read human communication and turn that into well ordered and resolved structured information that applications can use remained an open problem and a gaping whole in enterprise, much less consumer, capabilities.
Something as simple as reading a news story and then figuring out what people are talking about, what groups they belong to, who they are most similar to, and what properties they have (such as age, current position, etc.) and then linking it to all others that are similar for particular observations have been beyond the state of the art until rather recently. And, while in the past ten years some in academia have demonstrated the above capabilities; few have productized this capability into something usable and very few major companies or consumer applications have anything like this capability in use today. Using an analogy of the MP3 player market in the late 90s and early 2000s, we are in the pre-iPod age of Cognitive Computing.
The need for productized Cognitive Computing is now great and getting greater. The Government had to make a massive investment into the predecessors of this technology after 9/11 to find patterns in nearly infinite haystacks. Now industry is finding that the risks in their data are best dealt with proactively instead of settled reactively. And those on the cutting edge of finding opportunities see amazing opportunity in understanding Big Data through Cognitive Computing. This is leading the refining of a whole new hybrid field called Data Science and further exacerbates the need for something that can leverage Cognitive Computing to make human communication viable for Data Science.
There is a clear vacuum for a platform that can be trusted by all of these parties.
This is the vacuum that Digital Reasoning is stepping in to.
The Applications of Cognitive Computing
Great products are known for their utility to the user and the way they inspire great loyalty and drive the imagination of the customer. They are trusted and they are loved. They rarely are created quickly but often appear on the scene as if from nowhere and before anyone knows why… they just have to have them. They are often not pure inventions but the assembly of many capabilities into an elegant, integrated and focused design that extends what the user does today while at the same time enabling types of use they couldn’t imagine before the product existed.
We believe the first and foremost application of Cognitive Computing is the amplification of awareness – the scaling of detecting and triaging patterns in data that are known to be risky. We call this the Proactive Use Case. By combining the innate ability of our platform, Synthesys, to build up from customer data a rich Knowledge Graph through algorithms with trained patterns that are known by experts to represent Key Risk Indicators, we can find risks at a scale that was previously impossible. Whether this risk is in masses of intelligence messages or in the billions of emails flowing through investment banks every year, the Proactive Use Case is making people safer and protecting customers and shareholders in new and increasingly necessary ways. Imagine being able to teach a computer how to spot when two people are talking about a deal and how that talk impacts their activities? Imagine no more. That is what customers are doing today with Synthesys and Proactive applications for Compliance.
Now imagine you wanted to take one of these risk patterns, such as a known bomb maker in another country, and you wanted to see who he’s talked to and when and where he has been. Imagine you had a giant conceptual Map of the World based on the data you had collected trying to prevent an attack on key public spaces in other countries. This is where the Reactive Use Case steps in making investigations over unimaginable amounts of data available in radically less time than previous tools. Thanks to technologies like Impala from Cloudera, this is not just an analyst dream anymore. This is the reality of what can be done today.
Billions of relationships that started in news and human communication, learned and connected by machine, kept secure and private and owned by our customers to support protecting their citizens, their shareholders, their employees, and their customers. Most importantly, since the customer owns their data, they get the value from it to support their mission. We are just providing the technology to refine these masses of data into something they can use. Some are now using it to help them make better investment decisions or better customer connections. Others want to use it to make sense of doctor and nurse’s notes to radically improve the quality of care and drive another generation of innovation in medicine. This may prove to be the most exciting application of our technology to date. And beyond the walls of the Pentagon or the enterprise or the hospital, still others want Cognitive Computing to tell them what’s important today in an objective and personalized way.
Protecting the public.
Discovering cures and helping treatment.
Knowing what’s important to you now.
These applications are using the power of Cognitive Computing through Synthesys to automate the understanding of their data and scale their own expertise across all of their data.
Customers trust it to create products of their own. This is what makes it a trusted platform. Trusted to work better and be more complete than other competing technologies, trusted to work with the most sensitive data in the world, and trusted to provide them freedom and value on their data – not try to seize that value for ourselves. As these current applications become cemented and ubiquitous and future applications both change the way businesses work and how we live our lives, the power of Cognitive Computing technology will expand in our society and begin to transform it. It is important that we think about this before these things play out – both in terms of how we safeguard from excesses and be careful in trusting any one group with too much power and control over our lives.
Choosing the Society We Want While We Still Have The Freedom To Choose
Cognitive Computing can transform our society for the better. Or it can be used for control, exploitation, and manipulation.
We have the freedom to choose between technology companies that deliver products to us or technology companies that turn us into the product.
If you aren’t paying for technology, then you are the product.
The government and Big Business understand this and are increasingly moving to a model like the Digital Reasoning model of Cognitive Computing where we supply the refining technology and they create and market the products while protecting the customer and employee data. That is what we believe makes us a trusted Cognitive Computing company. Trusted to deliver an open platform that makes cognitive computing real and that can be owned by our customers to create value for them- in many languages and across a limitless number of domains.
It is a great accomplishment of history that technology that many thought would not be personal or be owned by the common man ended up being so. Henry Ford believed that every person would want their own car – it required innovation to democratize that and make it affordable for all as well as beating the entrenched interest groups that controlled car manufacturing at the time.
He was right.
Steve Jobs believed that every person would want their own computer at a time when the leadership of the world’s largest technology company (IBM) did not.
He was right.
Bill Gates believed (like his childhood hero Henry Ford) that while the hardware platform might change, the value that could reach every home and every office through software was something everyone should own.
He was right.
We believe that every person will want to own their cognitive computing system as an extension of themselves. We believe we are on the right side of history. We believe that the Hadoop clusters of the enterprise and the shared compute possibilities of the Utility Cloud like AWS are just the beginning and a bridge to a ubiquitous world of Cognitive Computing that provides the environment for the software people will own and use at work and in the home and across every device.
With this enduring mission, it seems appropriate to bring this all together into something that speaks to the world about what we are doing and will continue to do that is most relevant today.
Deliver trusted cognitive computing for a better world.
It’s a challenge that must be met. And it will be. Begun in the great and long shadows of the aggregators, it will peer out from behind them and overtake them, bringing light and good fruits for its users. No man or company should own the Knowledge Graph, because every company or man or woman should own its Knowledge Graph.
We are on the right side of history.