Sparking Innovation: Cognitive Computing & KM #KMWorld

KMWlogo_Stacked_Session Description: Agility, speed and flexibility are key requirements for organizations today.  Enterprises need a new approach to handling, analyzing, and acting on complex information—as it arrives.  Feldman, a long-time technology analyst discusses a new approach to knowledge management that addresses the complex problems enterprises face today.  She considers the impact of cognitive computing on the IT industry and how it will affect our jobs and our lives. She raises issues and possible impacts for those in the search, discovery, content management and knowledge management areas, and demonstrates why KM professionals are uniquely well suited to understanding and using these new technologies.  She’ll end by giving us a glimpse of a future fueled by cognitive computing.

Speakers: Susan E. Feldman, CEO, Synthexis Cognitive Computing Consortium

[These are my notes from the KMWorld 2015 Conference. I’m publishing them as soon as possible after the end of a session, so they may contain the occasional typographical or grammatical error. Please excuse those. To the extent I’ve made any editorial comments, I’ve shown those in brackets.]

NOTES:

  • Ingredients of Innovation: 
    • Problem or research direction
    • Opportunity
    • Cross-fertilization
    • Colleagues
    • Passion
    • Open Mind
    • Curiousity
    • Opportunity
    • Accident
    • Serendipity
  • What is Innovation?
    • A new idea, practice or object
    • rarely entirely novel — this is one reason why knowledge management can be so helpful for innovation
    • most successful innovation occurs at the boundaries between subjects or organizations
    • group, rather than individual effort — developers, users, partners, colleagues
    • tends to occur at lower levels of an organization — top-down innovation is rare
    • may disrupt industries or companies
    • risky and rewarding
  • Business Case for Supporting Innovation.
    • successful innovation drives growth in the economy
    • increases company revenue, especially when you are first to market
    • facilitates market dominance
    • helps a company attract and keep customers — it builds customer loyalty and market buzz
    • helps you avoid being disrupted — you stay competitive and expand into new markets
    • it helps create a fertile envionrment for R&D
  • The Innovation Process.  This process requires flexibility and time. It cannot be too tightly managed because that strict management will constrain creativity.
    • Engage in open discussions, wide reading, input from colleagues, customers, partners — all of this leads to a growing awareness of a need
    • define the problem
    • eliminate common, prosaic ideas
    • simmer — put it on the back burner and let it develop further
    • explore broadly
    • filter, winnow, focus
    • rethink, iterate, start from the top again
    • develop
  • Standard view of the innovation process. This is only half the process because the standard view presents innovation as a linear process. However, thinking is not always linear. Here is the standard linear view:
    • define problem
    • research
    • develop
    • commercialize
  • The iterative Innovation Process. By contrast to the standard view of the innovation process, the iterative innovation process is not a linear process. It looks more like spaghetti.
    • it involves lots of conversation, reading
    • it involves iterating, backtracking, pivoting
  • The Role of Information Access and Analysis Tools. We have tools that are pretty good at finding things we have and things we know. However, innovation requires that we get better at discovering what we do not know. Therefore, our tools need to help us
    • improve exploration and discovery
    • introduce related information without drowning us in superfluous information
    • improve and/or eliminate queries, then help the user frame the question broadly
    • discover unexpected relationships
    • search on a concept level rather than by keywords
    • unite multiple sources of information, including some you may not know
    • collect and share
    • enable information and people interaction in one application
    • save time
  • Cognitive systems are key for serendipitous exploration. Cognitive computing makes a new class of problem computable. This new class of problem:
    • is ambiguous, unpredictable
    • involves conflicting data
    • requires exploration, not searching
    • depends on uncovering patterns and surprises
    • involves shifting situation, goals, information
    • requires best answers that change based on context
    • requires problem solving that goes beyond mere information gathering
  • Context is a differentiator. Every problem may surface in completely different contexts that lead to completely different answers. Cognitive computing is best at addressing these different contexts.
  • Cognitive computing is...
    • meaning-based
    • probabilistic — you get several likely answers rather than just the ONE answer
    • iterative and conversational
    • interactive
    • contextual
    • learns and adapts based on interactions, new information, users
    • it has a big data knowledge base – multiple sources, formats
    • analytics
    • highly integrated set of technologies
  • What cognitive systems do:
    • analyze BIG data
    • understand human language on multiple levels
    • analyse and merge all formats and sources of information
    • uncover relationships across contexts
    • understand and filter content and context
    • find patterns and uncover surprises
  • Examples of cognitive computing at work:
    • Are there new drugs that might be MORE effective for controlling diabetes?
    • Who is funding this terrorist organization and how are the funds delivered? Is this organization a threat?
    • Can I identify the MOST RISKY product or customer problems before they blindside our company?
    • Which company will be the MOST PROMISING M&A target?
  • What cognitive computing is NOT
    • just big data or AI
    • robotics
    • drones
    • humanoids
    • entirely autonomous
    • the singularity
    • a human replacement
  • Cognitive system:
    • start with a question
    • analyze the question — define the kind of question it is and what kind of answer might be required.
    • enable exploring by expanding the problem statement and generating a variety of hypotheses
    • the orchestration element of cognitive computing determines the best way of answering particular type of question — this matches the likely answers to the most appropriate context
    • the cognitive processor is analogous to an index — it matches concepts, it provides confidence algorithms
    • the outcome is the dataset, which is then put through a series of filters
      • who, what, why and when?
      • who else has asked this question?
      • in what context?
    • sent the filtered results to the exploration loop = a set of tools that help visualize and analyze the information
  • Cognitive computing requires a large array of tools — this is necessary because, in many ways, it is trying to replicate the extraordinary ability of the human brain such as
    • facial and feature recognition
    • speech recognition
    • interactive voice response
    • content intelligence
    • about 20 more tools!
  • Barriers to Innovation
    • lack of organizational support
    • party line thinking
    • no time to think
    • too-rigid innovation systems
    • lack of encouragement of innovation
    • too much information
Share