This document is our first public draft of research we have done in conjunction with the Cognitive Computing Consortium and Babson College.
Our objective in this document is two-fold. Cognitive computing as an approach to human-machine problem solving is new and hence somewhat unfamiliar. Consequently, like any other new technology, there is a certain amount of hype and confusion that contribute to clouding its adoption. Our first objective is to briefly distinguish how AI and cognitive computing differ not only from each other but also from traditional information systems. Our second objective is to present a set of tools that will guide buyers and vendors of cognitive applications through a set of decision points. Cognitive applications are new, and largely untested. We hope that as you select, deploy and test these new applications, you will discuss your experience with them so that we can test the decision tools described in the downloadable document. Please contact me at firstname.lastname@example.org with questions and comments. We hope that our work provides some clarity for this burgeoning field.
To Download, click here:Understanding-cognitive-computing
Like anything else in life except death and taxes (and even the particulars of these are uncertain), uncertainty is something that humans deal with every day. From relying on the weather report for umbrella advice to getting to work on time, every day actions are fraught with uncertainty and we all have learned how to navigate an unpredictable world. As AI becomes widely deployed, it simply adds a new dimension of unpredictability. Perhaps, however, instead of trying to stuff the genie back in the bottle, we can develop some realistic guidelines for its use.
Our expectations for AI, and for computers in general have always been unrealistic. The fact is that software is buggy; that algorithms are crafted by humans who have certain biases about how systems and how the world works—and they may not match your biases. Furthermore, no data set is unbiased, and we use data sets with built in biases or with holes in the data to train AI systems. These systems are by their very nature, then, biased, or lacking in information. If we depend on these systems to be perfect, we are letting ourselves in for errors, mistakes, and even disasters.
However, relying on biased systems is no different from asking a friend, who shares your world view, for information that may serve to bolster that view rather than balance it. And we do this all the time. Finding balanced, reliable, reputable information is hard and sometimes impossible. Any person trying to navigate an uncertain world tries to make decisions based on balanced information. The import of the decision governs (or should) the effort we make in hunting for reliable but differing sources. The speed with which a decision must be made often interferes with this effort. And we need to accept that our decisions will be imperfect, or even outright wrong, because no one can amass and interpret correctly everything there is to know.
Where might AI systems fit into the information picture? We know that neither humans nor systems are infallible in their decision making. Adding the input of a well-crafted, well-tested system that is based on a large volume of reputable data, to human decision making can speed up and improve the outcome. There are good reasons for this because human thinking balances AI systems. They can plug each others’ blind spots. Humans make judgments based on their world view. They are capable of understanding priorities, ethics, values, justice, and beauty. Machines can’t. But machines can crunch vast volumes of data. They don’t get embarrassed. They may find patterns we wouldn’t think to look for. But humans can decide whether to use that information. This makes a perfect partnership in which one of the partners won’t be insulted if their input is ignored.
Adding AI into the physical world in which snap decisions are required, raises additional design and an ethical issues which we are ill-fit to resolve today. Self driving cars are a good example of this. In the abstract, and at a high level, it’s been shown that most accidents and fatalities are due to human error. So, self-driving cars may help us save lives. Now we come down to the individual level. Suppose we have a sober, skilled, experienced driver who would recognize a danger she has never seen before. Suppose that we have a self driving car that isn’t trained on that particular hazard. Should the driver or the system be in charge? I would opt for an AI assisted system with override from a sober, experienced driver. On the other hand, devices with embedded cognition can be a boon that changes someone’s world. One project at IBM research is developing self driving buses to assist the elderly or the disabled live their lives independently. Like Alexa or Siri on a smaller scale, this could change lives. We come back to the matter of context, use, and value. There is no single answer to human questions of “should.”
This brings us to the question of trust. Should we trust AI systems and under what circumstances? That depends on:
- The impact of wrong or misleading information: Poor decisions? Physical harm? Momentary annoyance?
- The amount and reliability of the data that feeds the system
- The goals of the system designers: are they trying to convince you of something? Mislead you? Profit from your actions?
- The quality of the question/query
Is there some way to design systems so that they become an integral part of our thinking process, including helping us develop better questions, focus our problem statements, and reveal how reliable their recommendations are? Can we design systems that are transparent? Can we design systems that help people understand the vagaries of probabilistic output? Good design is the key—within the context of the use and the user.
Cognitive computing is emerging as a significant part of the next generation of computing. Because it is early days in this new generation of computing, there is still no widespread understanding of what it is and how it differs from some of its relatives: AI, internet of things, machine learning, conversational systems, bots, or NLP. We see that in both the US and in Europe that companies are very interested, but are mostly still at the experimentation and proof of concept stage. We will be tracking some of these projects as they develop their cognitive applications and roll them out more broadly. There is no question, though, that interest is high, and that the ability to augment and assist users, as well as to move from static to dynamic systems has great appeal.
I recently had the opportunity to attend a focus group, sponsored by SAS Institute, on cognitive computing adoption outside the US. Attendees came from Denmark, Japan, Finland, Serbia, Netherlands, Sweden, Switzerland, India and Ireland. They represented financial services, telecom, consumer product manufacturers, government agencies, and airline companies. Here are some gleanings from their wide-ranging discussion.
How are you using or how do you expect to use cognitive computing?
- Automatically revise and evolve rules to expedite adaptation
- Uncover and improve best business practices and processes
- Detect patterns of behavior. Detect abnormalities. Identify risks
Augment human agents who can’t handle the current workload by automating the more predictable aspects of the job.
Why move to cognitive computing?
- Handle large amounts of data with many more variables. Especially textual data.
- Reduce need for adding manpower. People just don’t scale.
- Stay ahead of competitors
- Uncover surprises. (This was a side benefit to demonstration project that was originally designed to augment the human workforce)
- Curiosity to see what benefits might derive from cognitive computing that we can’t get now
- Get rid of silos
- Automate predictable or repeatable work
- Augment human work by developing digital assistants
Examples of uses:
- Speech-to-speech product sales. Their innovation lab is experimenting with this. App will be personalized. Will use machine learning to replace hundreds of business rules and 20 predictive models. Machine learning will allow the models to evolve and will help to revise rules faster.
- Will discover and extract patterns of best business practices to establish KPI’s worldwide from hundreds of business managers. Need to understand what practices work and why.
- Expedite transactional processing. Moving from rules based process to teaching a system how to assess to minimize delays.
- They are tracking and analyzing invoices using rules and econometric models. Their goal is to teach a system to automate model development and modification for tracking and analyzing invoices. Extending beyond rules and econometric models they want to add sentiment from incoming, non-English communications.
- Recognize patterns of behavior to find anomalies and predict risk to more thoroughly assess people and goods.
- Automate responses to customers, but on a more individual level. Part of a project to analyze customer opinions—a big data project.
In all cases, augmenting existing applications and seeking net new benefit from the use of cognitive computing systems was consistent among all participants. Completely new product development or drastic changes to business processes from cognitive computing application wasn’t seen to provide the palatable business benefit needed to embrace adoption. In all cases, however, changes and improvements to existing business practices were expected.
This group of early adopters was proceeding with caution. They had the bruises from past new technology experiments and don’t believe the hype around AI today. In each case, it was apparent that they had support from high-level management, and that they were starting with a proof of concept, or several. We have heard this from other buyers. Several are working with more than one vendor, trying to compare dissimilar products with little in the way of best practices to guide them.
Some of the concerns that emerged were first that these systems are often a black box; that it was not clear why they were getting the recommendations that were delivered. Because business systems are traditionally databased, this ambiguity appears to be unacceptable to them for some uses today. The buyers felt that they needed the evidence behind the results. Probabilistic systems, including search engines, have long struggled with this problem. Although we know that information systems of all sorts deliver only what you ask for and not what you should have asked for, nevertheless, they are seen as precise and complete. Managing expectations is a challenge for vendors and for IT managers.
Other group members were concerned about the need for a lot of computing power. Several mentioned the challenge of developing non-English applications because most of the research has been in English-based systems. Perhaps most intriguing in terms of issues, though, were the predicted “What-if” questions: will we lose the institutional memory that originally trained the system? If so and if the system breaks down, will we be able to fix it? Centralized systems are always a problem, they said. They must be up and running 24/7. They must be reliable. That’s a challenge for any system.
Finally, they pointed to interaction design as a great unknown, especially for non-IT, non-analyst business users who need access to data stores but won’t understand the system design behind the interface. Right now there are experiments, but no accepted best practices.
It is apparent that SAS is seizing this trend toward cognitive computing. The announcement of SAS Viya ™ at this conference, along with a variety of tools for both their loyal developer and analyst base and a wider business user audience positions them nicely as both a partner with other cognitive and IOT platforms and as a potential competitor.
We will continue to track cognitive use cases and report on them. The field is evolving rapidly. Focus groups like this, and like the Cognitive Computing Consortium’s soon-to-arrive discussion forum will enable experimenters to teach each other, perhaps mitigating mistakes that might otherwise be widespread.
Facebook’s Dave Feldman writes a blog on medium.com that discusses interaction design. The most recent post examines how useful chatbots are vs. GUI’s. (https://medium.com/@dfeldman/bots-conversation-is-more-than-text-1c76d153e13d) Cutting through the hype, Dave gives some examples of how adhering to one design camp or another can create a frustrating experience for a customer. When you move the online dialog into the real world of ordering a meal in a restaurant, it’s quickly apparent that neither type of design approach makes sense. In one case, the experience is cold and inhuman, even though it’s efficient. In the other, the amount of information conveyed makes it impossible to consider all the choices and come to a decision. The trick is to use each approach judiciously, and probably together, depending on the amount of information to convey, the type of dialog needed in order to make a decision, and the human element that creates a warm, satisfying customer experience. Check it out. It’s both entertaining and instructive.
By Sue Feldman
In the cognitive computing era, there are plenty of tough technical challenges. Their difficulty pales, however, when compared to the social and legal issues these new technologies raise. Increasingly, we rely on algorithms to help us sort through the complex factors that lead to making a decision. For most of us, there is no way of knowing whether the algorithm is well suited to handle our current situation. In fact, by their very nature, algorithms cannot be crafted to react dependably to the unforeseeable. Articles by Julia Angwin in the New York Times and ProPublica on Aug. 1st celebrate a decision by the Wisconsin Supreme Court to limit the influence of algorithmic recommendations for sentencing offenders. The algorithms predict the risk that an offender might commit a crime in the future. Based on these recommendations, an offender might face jail time or probation. See http://www.nytimes.com/2016/08/01/opinion/make-algorithms-accountable.html?ref=opinion&_r=0 or https://www.propublica.org/article/making-algorithms-accountable.
There is no stuffing the algorithm genie back in the virtual bottle. The fact is that we need help in making sense of the welter of data that showers us whenever we make a decision. From choosing a carpenter to treating a cancer patient, the human mind can’t take in every available data point quickly enough to make the most optimum decision in a reasonable amount of time. For the most part, that is not a problem. We don’t need to know everything to make an acceptable decision. There are plenty of good carpenters, restaurants, or books. Rarely are day-to-day decisions a matter of life or death. But sometimes they are. From self-driving cars to medical treatment, when lives are at stake, should we rely on algorithms alone?
Our society tends to rate the accuracy of computer results much more highly than that of human decisions. For some reason, we leave our skepticism behind when recommendations are digital. What has created this aura of infallibility? As a young researcher, I found that I could hand a client the same information in the same words as a digital record and as a photocopy and have the digital version more readily accepted. The believability bias hasn’t changed much since then. It’s time to develop a more mature approach to melding digital evaluations with human common sense. We need to ensure that the path to digital recommendations is transparent and that the underlying data is reliable so that we can judge the conclusions for ourselves. We also need to teach skepticism.
Computers and humans complement each other. Neither is perfect. Combined, human sense making and algorithmic pattern detection make for more complete (but still imperfect) understanding. Angwin says we must require “the right to examine and challenge the data used to make algorithmic decisions about us.” That’s a good first step.
When IBM’s Watson burst upon the scene in 2011, little did we know that it would kick off a new category of computing. Since that time, IBM has drawn most of its major divisions into the cognitive fold. That’s no surprise: cognitive computing is the ultimate Venn diagram, drawing on hundreds of technologies, from AI to Zookeeper, in order to create systems that “interact, understand, reason, and learn.” It was apparent at the Watson Analyst Day on May 23rd that IBM’s message has been refined, and that it has begun to gel. Just as we in the Cognitive Computing Consortium have moved from a vague understanding that we had something fundamentally new, so too has IBM’s understanding of what cognitive computing is, and what it’s good for become much more solid.
Realizing that the complexity of cognitive solutions can be a barrier to entry, IBM Watson has begun to offer “App Starter Kits” around clusters of technologies that are pre-integrated, like conversation agents, business intelligence, or audio analysis. But markets require more than a single vendor, and we have already seen the rise of new vendors that are not part of the Watson Partner constellation. Being able to mix and match platforms, apps, and technologies will require new standards for not just formats but also storage and terminology if all types of data are to be exchanged easily. Making Watson’s cloud-based cognitive services like sentiment extraction, NLP, predictive analytics, or speech-to-text on both Bluemix and Twilio is a good step in this direction. So are the emerging sets of tools to guide adopters through data selection and modeling, analytics selection, visualization choices, and interaction design.
Two years ago, IBM launched its Watson Division. It now has 550 partners in 45 countries, thousands of developers, and programs in conjunction with 240 universities. It continues to add new languages and services. This is the beginning of a market, but we believe that this phenomenon is bigger than a single technology market. Rather, IT will evolve from the current deterministic computing era to one that is more nuanced. We already see elements of cognitive computing creeping into new versions of older applications—more intelligent interactions, better, more contextual recommendations, In this new world, we will add probabilistic approaches, AI, predictive analytics, learning systems, etc., but we will also retain what works from the old. That calls for a much deeper understanding of which technologies solve what problems the most effectively. What kinds of problems demand a cognitive computing approach? The processes that IBM delineated as possible elements of a cognitive solution are:
They also emphasized the importance of data—curated, annotated data that is normalized in some way using ontologies for both categorization and reasoning. This should come as no surprise to those of us from the online industry, who know that there is no substitute for the blood, sweat and tears that go into building a credible, usable collection of information. The question today is how to do this at scale, and at least semi-automatically, using NLP, categorizers, clustering engines, and learning systems, training sets, and whatever other tools we can throw at this barrier to sense making.
By far, the biggest advances in cognitive applications have been made in healthcare. With good reason. Medicine has a long history of information curation. Advances in ontology building, controlled vocabularies (normalization) and categorization date back to the 1950’s. PubMed and its predecessors had already built multilingual online collections of medical publications, clinical data, toxicology, and treatment guidelines as early as the 1980’s. These resources predate IBM Watson health and have enabled it to address health information problems with an existing well-curated knowledge base. Healthcare requires extreme accuracy, big data analytics, advanced patient-doctor-machine natural interaction, and a probabilistic approach to solving a medical problem. Because the amount of possibly relevant information is staggering, and the outcome is a matter of life and death, the reasons for investment in cognitive systems are obvious for healthcare insurers and providers alike. There are also, of course, billions of healthcare dollars at stake. Customer engagement, retail sales, mergers and acquisitions, investment banking, security and intelligence are not far behind in their promise, but they lack that initial bootstrapping of existing knowledge bases.
In summary, cognitive computing is moving from dream to reality. New tools and more packaged applications have reduced the complexity and the time to deploy. Early adopters are still at the experimentation stage, but from IBM and other vendors and services firms, we see gradual adoption with associated ROI, a virtuous loop that attracts yet more buying interest.
IDC’s Third Platform –the next computing generation—rests on cloud computing, big data and analytics, social business, and mobility. Together, these form a foundation to provide scalable anywhere-anytime-any device computing. As these trends become ubiquitous, they enable and accelerate the Internet of Things (IOT), cognitive systems, robotics, 3D printing, virtual reality, self-driving cars, or better security.
At the same time, this brave new world wreaks havoc on the old one of PC’s, client-server software and legacy apps. I would also add another disruptive ingredient to the mix—open source software, which is no longer for hobbyists and is now embedded in most new applications. IDC predicts that 2016 is the year in which spending on third platform IT will exceed that for the second platform, with a CAGR of 12.7% for 2015-2020. At the same time, they predict that second platform investment will be down 5%.
Their recent surveys show that, in terms of maturity, today most companies are in the digital exploration or early platform development phase, with 14% having no interest in digital transformation, and only 8% already using digital transformation to disrupt competitors or markets. That will change by 2020 as 50% of businesses will be using this platform to disrupt and transform.
- Business owners, not IT will control more of the IT budget
- Health services and financial services are two of the top industries to invest, reaping the rewards of faster, cheaper, and more comprehensive uses of their data.
- Other top applications now in the works include marketing and sales, retail, security, education, media and entertainment.
- Technology will be embedded in most applications and devices.
- Start-ups are rife, and the shakeup has not yet begun
- Cognitive computing and AI is a requirement for developer teams—by 2018, more than 50% of developer teams will be using AI for continuous sensing and collective learning (cognitive applications and IOT).
Where does existing IT infrastructure play in this game? In our scramble as analysts to pin down trends, we often neglect the fact that existing systems and applications a still valuable. They may well be good enough for a given task or process, or they may continue to churn on, feeding into newer layers of technology stacks when appropriate. Unlike newer versions, the kinks have been worked out. The challenge for business and IT managers will be to distinguish between the promise of the new and the security of the old: when to invest, when to explore, and when to stand back and watch. Good question!
Click here or more information on IDC’s take on the 3rd Platform
[The following article is a transcript of a video of Sue Feldman’s keynote session at KMWorld 2015 in Washington, DC. View the full session video at http://www.kmworld.com/Articles/Editorial/ViewPoints/Cognitive-Computing-and-Knowledge-Management-Sparking-Innovation-108929.aspx?PageNum=2]
Download the slides: Sparking Innovation
Innovation is perhaps the biggest test of a knowledge management system. We’re used to capturing information. We’re used to locking it down. We’re used to accumulating it. We’re used to creating some kinds of access to it. Innovation makes us go far beyond that. What I’ll talk to you about today is what innovation is. What the process for coming up with a new idea actually is. Then I would like to talk to you about cognitive computing because I think that it solves some of the problems that our older, traditional technologies cannot really address adequately. I’ll end by talking a little bit about where I think knowledge management has to go.
Let me start by telling you a story. Once upon a time, there was a biologist and physicist and they went for a walk. The first thing they started to do is to fall into a conversation about a fairly arcane subject: DNA. The physicist was interested in the electrical properties of DNA. The biologist knew a fair amount about that because he was also a chemist and something of an inventor. They talked and they walked and then the physicist went back home, continue to ask questions, do research, etc. The biologist kept on sending information but really went back to what he liked to do best which was tinkering with ideas and things because he was something of an inventor. After several years of research, the biologist, Esther Conwell, won the National Science medal in 2010 for her work on the conductive properties of DNA and how to enhance those because she was interested in semi-conductors. The inventor was my Dad, and that’s the kind of thing that he enjoyed, which I think is a story of what happens when you have two innovative, open-minded people.
Ingredients of Innovation
Let’s take a look at the ingredients. First of all, you need a problem or research direction. In this case it was semi-conductors. You also need opportunity. You need cross-fertilization, in this case biology and physics, which are adjacent but certainly not congruent. You need colleagues who, like you, interested in discussing. In the research that I’ve done over the years on the process of innovation, I’ve found that innovative inventions of various kinds, and discoveries tend to be sparked by good food and a bottle of wine. It’s almost a requirement. You need curiosity. You need that serendipitous encounter to create the “aha moment,” a happy accident. You also need information and you need support both in the sense of an organization willing to let you model around and support in the sense of the information that is provided to you.
What is innovation? Well, it’s a lot of things. When President Obama presented the medal to Esther Conwell, he said that innovation is fueled by a combination of caffeine and passion. Obsession actually. Certainly, it requires a new idea, but it’s rarely entirely novel. It builds on what came before and that should be of importance to knowledge managers. Game-changing innovations occur at the boundaries between subjects and organizations. It’s a group effort rather than an individual one. Developers, users, partners, and colleagues all have a part in it because they provide not just the ideas but also the need that spurs the innovator to solve a problem. It tends to occur at the lower levels of organizations. Those of us who are at the top of the organization, beware. It may disrupt industries or companies for good or ill and it is both risky and rewarding. That’s innovation.
What’s the business case for supporting innovation? Because very often it doesn’t pay off. Those of you who have R&D departments know that that’s the case. First, revenue. If you’re successful, it drives revenue because you are first to market. That means you are able to dominate that market and in fact that’s what’s happening with cognitive computing right now. You can attract and keep customers, build customer loyalty and market buzz, shape that market the way you want. It helps you avoid disruption and stay competitive. It helps you expand into new markets.
By creating a fertile environment for R&D, you also have a pipeline of new ideas to avoid stagnation and being bypassed by competitors. You attract outstanding researchers who soon burn out and leave if you don’t provide them with that kind of organizational support and latitude because innovation is a fragile flower. It gets trampled very easily.
On my second job, a very long time ago, I got hired by someone who called me in after two weeks and said, “I heard that you are innovative, Susan. You haven’t had any ideas yet.” She was right. I never had another one for her.
The Innovation Process
What is the innovation process? It’s quite different from what goes on in knowledge management normally. First, you have to have that idea or interest. There’s no question about that. You have open discussions, wide readings, you bump into people, you talk to them in the hallways, you go out to dinner with friends who are not in the organization, and gradually you discover that there is a need, which you find intriguing.
This is a very individual process even though it requires other people. You define the problem. You eliminate some of the common ideas. You discover that other people have been there before you and you give up and do something else. Then something interesting happens: you’ve taken in all this information, you’ve stuffed it into your head, and you let it simmer.
We had a graphic designer who is tremendously innovative. He used to go home and take a bath. I’d go for walks. Other people do other things. They knit. They cook. They garden. Whatever it is, they have to distract the front of their brain so that the back of the brain can allow that ferment to happen and that’s great fun. But if you have too tight a deadline, you’re not going to follow that elusive idea which is half-formed because you don’t have time for it. You have to meet the deadline and the idea gets squashed. That’s a very important thing for organizations to understand. These people who are innovators need some direction, but they also need a great deal of latitude and freedom as well.
You have to explore broadly. (This is where cognitive computing and knowledge management coincide, as I’ll discuss later). You have to filter and winnow and focus and rethink and iterate and go back to the beginning and start all over again. Finally, you have something concrete enough to develop and off you go, maybe. You find the problem, do research on it, then go off and develop. You commercialize it, you throw it into the market place and you see what the consequences are–big revenue, losses, whatever it turns out to be.
You identify the problem by talking to customers, talking to colleagues, talking to sales people, and talking to other researchers. Coming back is very iterative, as most of you know. You do research and you redefine the problem.
Again, you iterate. Test it on the market like at social media. Do competitive intelligence. Then you might commercialize it and see what happens after that.
Discovering What We Don’t Know
There is set of information tasks that we try to support with knowledge management, research, and text analytics. Any sort of information access and management tool is aiming to support all of these tasks, but the tools rarely do. The problem is that we have separate tools. The creation tools may not be well integrated into the process. If they are, the fact is that in innovation we’re on the phone, we’re sending emails, we’re discussing in the hallways. We’re not capturing that.
The reasons why we make decisions and change directions are poorly known and can’t be modeled for the process to happen again. We’re losing information that’s falling off the table. We’re pretty good at finding in some ways. We’re not so good at discovering what we don’t know and uncovering patterns we don’t know enough to look at. That discovery and uncovering are key to innovation, because what we want is to find out what we don’t know so that we can invent it. We’re pretty good at analyzing information and getting better. The discussion is very often not integrated into this whole picture and the decision-making is fairly diffuse. These are information tasks that we need to be able to support.
The Role of Information and Analysis Tools
The role of information access and analysis tools in this case is to improve exploration and discovery, to introduce related information. Although we want related information, we don’t want all the information in the world.
How do we manage to promote those happy accidents without burying the searcher? We have to help with the information-finding process to eliminate queries perhaps in favor of exploration of some sort. We have to help. This is where our traditional systems also fall down, in helping the user to frame the question broadly, helping the user understand how to ask for the information they need if they don’t know they need it. We used to have knowledgeable intermediaries who did a lot of this, but that’s not what’s happening today.
The tools have to help us understand and discover unexpected relationships across all sources of information. They need to search on a concept level rather than on keywords because those are also limitations. They need to unite multiple sources of information no matter what format they’re in or where they reside. They need to collect and share and discuss. They need to enable information and people to interact in one place. Then of course they need to save us time so that we can look at enough information in order to have those ideas. That tools that have started to emerge over the last couple of years are key to supporting these expanded roles for knowledge management. Cognitive systems are the next logical step.
As an analyst, I’ve been watching the markets develop all kinds of tools: business intelligence, search, text analytics, graphics of various kinds, reporting tools, creation tools, and drawing tools. They all solve a piece of a problem.
We used to call that “Sneakernet.” The Sneakernet that goes on in the creative and innovation process is overwhelming. It’s a tremendous waste of time because it means you’re constantly rummaging back through stuff that you did 10 years ago because you know you did it already. In fact, when I was preparing this talk about innovation, I had to go back to research I did 10 years ago because I knew I’d done something about this, but I really didn’t remember where it was. It was really hard to find it; desktop search is terrible. Yet, there it was in the back of my head.
Enter Cognitive Computing
See separate article in KM World, “What is Cognitive Computing”,
Cognitive computing is going to bring us another step closer to solving some of these problems. What is cognitive computing? Last year I brought together a team of 14 or 15 people to try to define it before marketplace hype completely screwed up any idea of what it was. I don’t know if we’re succeeding or not.
What are the problems that cognitive computing attacks? They’re the ones that we have left on the table because we can’t put them into neat rows and columns. They’re ambiguous. They’re unpredictable. They’re very human. There’s a lot of conflicting data. There’s no right and wrong, just best, better, and not such a good idea but maybe. This data requires exploration not searching. You just have to keep poking at it and shifting things around.
When I’m at the beginning of a project, I find myself jotting down ideas and then arranging them on a large table because sometimes they fit together one way and sometimes they fit together another. You need to uncover patterns and surprises, and computers are very good at this because they don’t get embarrassed by wrong ideas. Although they all have the biases that they get from their programmers, their biases are different from yours.
The situation is shifting as well. As we learn more, we change our focus and our goals. We go back and ask the same question that we often do, but we do so hoping for different results because we’ve already learned that stuff is not so easy in today’s systems. If you go to Google and ask the same question, you’re not always going to get the same answers, but you’ll get similar answers. But if you were looking for pictures of Java because you’re planning your vacation today and two months from now you want flights, the system won’t know your progress and your decisions.
The Value of Context
How do we make a cognitive system into a partner so that it keeps track of who we are and what we want to know at this time? It gives best answers based on who you are, where you are, what you know, what you want to know, and when you want to know it. It is very individually focused. Its aim is problem solving beyond information gathering. It gives recommendations based on who you are. I want to give you a couple of examples because context, we have found, is one of the key differentiators of a cognitive system.
In 2011, IBM designed a computer named “Watson” that won Jeopardy against two human champions. That was the beginning of cognitive computing. (You can see it on YouTube: https://www.youtube.com/watch?v=Puhs2LuO3Zc).
For me, as a person who has been in the chaotic world of search and text analytics all my life, it was a validation that the kinds of things that we do–the search index as opposed to the database–were actually really useful for very complex problem-solving. That was the beginning.
For another example, think about patient care. The emphasis on who, what, where, and when you are is one of the differentiators for cognitive computing. We all need slightly different slants on the same question. Let’s say we have a patient who has a disease. We know his genetic makeup, his age, his history of smoking, that he has certain allergies, etc. We also know where he is, what kind of access he has to medical care. We also have access to enormous amounts of information especially in health care and possible treatments and confidence scores. How does this change health care, because this is life and death?
Today, in standard health care, if you have a disease, or a particular kind of tumor, there are treatment guidelines. It doesn’t matter if you’re black, white, female, male, young, or old. That’s how you treat them. That’s not the way it needs to work. Instead, imagine you have all that information–more information than any doctor can amass in his head–and you’ve ingested it and you can start to match that person as a query, against that information and all of the applicable drugs side effects and what’s known from clinical trials. You come up with 2 or 3 treatments. Maybe the system says, “Have you considered that if you did this test we would have more confidence in recommending?” It’s a dialogue now. It’s a dialogue that supports the doctor and the patient in their decision on a treatment and that’s the kind of medical care I want. That’s another kind of context.
Suppose you’re an investor. In that case the context is for the portfolio, the personality. Are you conservative? Are you a risk-taker? How old are you? Do you want a lot of data or do you just want to be told what to invest in? Are you an influencer? How old are you? What’s your previous investment history? What are the market trends? What is your investment strategy?
All of those things need to be taken into account. That’s what human investment advisors do, but they’re not all-knowing. Starting with the evidence, the information, and then the ability to make a better judgment instead of a gut, intuitive decision is a very good idea–especially if it’s your money.
Consider the company, CustomerMatrix. They have a sales application. It sits on top of sales force. They do a lot of this. They look at who the sales person is. They look at who the manager is, who the strategist is for the company. They give different answers, but the thing that I’m fascinated by is that they also have ingested your business goals and your business strategy. They will make recommendations according to the usefulness of approaching one prospect or another for acquisition, or sales, or another department based on how a positive outcome will influence the business of the entire company as opposed to just that sales person’s commissions. It’s not a bad idea.
Another example is ExpertSystem. I’m just showing you that this can be very familiar. These systems do the usual text analytics things. They extract sentiment. One example is about a cat that was a popular resident one of the train stations. But they extract things like sadness and give people who are writing news, for instance, a very good idea by comparing the number of hits, number of readers, and number of Tweets, etc., against the kinds of extractions they’ve done already in order to understand better what readers are looking for. This is kind of a building block and the context is this particular event.
Elements of Innovation
What are the elements of innovation? People, collaboration tools, access tools, information of different types, and a work environment that is designed for cross-fertilization.
What kills innovation? Lack of organizational support, party-line thinking, no time to think, too-rigid innovation systems, lack of encouragement of innovation, poor or limited information and information access, and of course, information overload. We want lots of information but we don’t want too much. That’s a tall order for knowledge management.
Can we re-imagine knowledge management? What can we do to give us a sort of informed serendipity? How do we do this? Cognitive computing can help with this, but there are some changes we need to make–not just in our systems and our tools, but also in our thinking. To bring us back to that DNA metaphor, we need to get away from structures in some cases. They’re useful, but not only do we need to capture information and conserve it, we also need to cut it loose. We need to loosen our grip on the information bits that are attached to those taxonomies, that are contained on those cells of the databases, and that are in the document and the text analytics systems categorized to within an inch of their lives.
Let them loose. Let them float around, bump into each other, and give innovators the opportunity to create their own information soup, if you will, to explore without forcing them into the structures that we have created. Because what they want to do is to find the unexpected by creating schemas and taxonomies we are giving them what we expect in terms of how information works. This is a tall order for knowledge management, and I leave it with you as a challenge and a question.
This week, a research group asked me, “For cognitive computing to take off, what advances do you think we need to make the in next five years?” I answered the question, first listing the major components of a cognitive system, and then discussing which ones were still fairly primitive. But the question continues to haunt me. The fact is that we’ve had most of the components for cognitive computing for a very long time. Language understanding, machine learning, categorization, voting algorithms, search, databases, reporting and visualization tools, genetic algorithms, inferencing, analytics, modeling, statistics, speech recognition, voice recognition, haptic interfaces, etc., etc. I was writing about all of these in the 1990’s. As hardware capacity and architectures have advanced, and our understanding of how to use these tools has evolved, we have finally been able to put all these pieces together. But the fact is, that we have had them for decades.
Here’s what we don’t have: an understanding of how people and systems can interact with each other comfortably. We need to understand and predict the process by which people interact to question, remove ambiguity, discuss and decide. Then we need to translate that process into human-computer terms. Even more, we need a change in attitude among developers and users. Today, we tend to think about the applications we develop in a vacuum. The human initiates a process and then stands back. The machine takes the query, the problem statement, and processes it, spitting out the answer at the end. Users, because of their expectations that machines will not be information partners, helping the information problem to evolve and then finally be resolved.
That’s not the way a human information interaction happens. If two people exchange information, they first negotiate what it is they are going to discuss. They remove ambiguity and define scope. They refine, expand or digress. This process certainly answers questions, but it does more: it builds trust and relationships, and it explores an information space rather than confining itself to the original question. That’s what we need to improve human-computer interactions: first, help in understanding the question. Then, we need better design to enable that question to evolve over time as we add more information, resolve some pieces and confront more puzzles.
IBM Watson’s Personality Insights claims it can deduce who you are based on 100 words of your writing. Unable to resist, I entered the preface from my book, The Answer Machine, into IBM Watson Personality Insights service.
You are shrewd.
You are philosophical: you are open to and intrigued by new ideas and love to explore them. You are self-controlled: you have control over your desires, which are not particularly intense. And you are compromising: you are comfortable using every trick in the book to get what you want.
Your choices are driven by a desire for prestige.
You consider helping others to guide a large part of what you do: you think it is important to take care of the people around you. You are relatively unconcerned with tradition: you care more about making your own path than following what others have done.
In 2011, a very clever machine from IBM named Watson defeated two human champions in the quiz game, Jeopardy. Watson is an answer machine, and its Jeopardy win was proof that it could be done. The press was immediately abuzz: would machines replace humans? Would we need teachers, programmers, or writers in the future? Could we automate doctors?
The short answer to these questions is no, we still need people. But the better question to ask is how to join man and machine so that we can address more complex problems than either can manage alone. Machines excel at performing repetitive tasks. They don’t get tired and they don’t get bored. They are good at crunching vast amounts of information to find patterns, whether they make sense or not. They have no emotional investment in theories, and are consistent to a fault. They don’t get embarrassed if they return the wrong answer. Machines, however, are very bad at making the intuitive leaps of understanding that are necessary for breakthrough thinking and innovation. Humans excel at this kind of thinking. People can balance the imponderables that are almost impossible to program: diplomacy, subtlety, irony, humor, politics, or priorities. People are good at making sense of data and synthesizing ideas. Above all, people can understand and make exceptions to rules. Machines can’t. We need people to make decisions, but we need machines to help us filter through more information in order to make better-informed decisions. People need this assistance because they are swimming in an overwhelming sea of information, and need time to think if they are to innovate and act wisely.
For this kind of help, new types of more “intelligent,” language-capable machines, like IBM’s Watson, are a necessity. Marrying intelligent machines with humans holds great promise: machines to do the repetitive work and forage through massive amounts of information looking for patterns and evidence to support or reject hypotheses, while humans supply the necessary judgment, intuition, and system override to determine which patterns make sense. This collaboration divides up the work into what each party—machine or human—does well. Anything that follows a predictable pattern is a good candidate for automation. Health care tasks are a good example of this duality: let the machine enter diagnostic codes based on existing rules. Let advanced information systems find the latest research on treating illnesses. Gathering information, organizing it, weighing the probability of its pertinence to a particular patient—this what a Watson does well. This frees clinicians to work with patients, assess the evidence and use it to improve patient care.
Watson is an answer machine. Part search engine, part artificial intelligence, part natural language technology, and stuffed with the specialized information to answer questions on a specific subject, be it medicine, finance, or Jeopardy. Answer machines sift through mountains of information to find patterns. Although they don’t make complex decisions that try to balance costs, emotional aspects, or ethics, they free up humans to do what machines can’t do: consider the factual and thenon-factual and then make well-informed choices. They provide better answers, faster, than current search engines do. Watson is one visible, well-publicized example of an answer machine, but there are many others that are arriving on the scene, albeit with less fanfare. Search engine technologies– the focus of this book – have undergone a metamorphosis of their own. The simple search engine of the 90’s, which matched keywords and phrases, has been transformed into a multifaceted access point to all kinds of information—in multiple formats and from a multitude of sources. Indeed, today the term “search engine” is a misnomer. Like Watson, search today comprises more technologies than keyword search: categorization, clustering, natural language processing, database technologies, analytical tools, machine learning, and more.
This book examines the metamorphosis of search from its awkward command line youth to the emergence of something much more complex, something I call an answer machine. The following chapters look at the role that information plays in the work and personal lives of people today, the tools we have developed to interact with digital information, and the future for these technologies as they move from engineering marvels to everyday tools.]