The Cognitive Systems Institute, which is a new set of IBM university programs in conjunction with IBM Research and the Watson Business Unit,  will focus faculty collaborators on building and evaluating cognitive assistants for every profession.   Artificial cognitive systems, or cognitive systems for short,  exhibit capabilities and/or perform tasks deemed intelligent by natural cognitive systems, such as people.  Professional cognitive assistants are cognitive systems designed to boost the productivity and creativity of professionals.   Cognitive systems researchers belong to a special profession, which improves building and evaluating cognitive systems, working on teams with other professionals, such as computer and information research scientists, human factors engineers and ergonomists,   sociologists,  operations research analysts,  mathematicians, statisticians, industrial engineers, and others.

The Cognitive Systems Institute will focus on professional cognitive assistants that exhibit the three L’s – language, learning, and levels.  Professional cognitive assistants should interact via natural language, learn by ingesting documents, and make recommendations with confidence levels.   For example, the IBM Watson Jeopardy! winning  cognitive system answered natural language questions, ingested Wikipedia and other sources to learn, and provided confidence levels with its answers.  The IBM Watson group is working on cognitive systems to help doctors, financial planners, researchers, and even chefs.   IBM Research is also working on cognitive system that will spar with debaters and politicians to help boost their performance.

The new Cognitive Systems Institute programs will be launched in late August, and will be designed to (1) help prepare faculty to set up Watson/Cognitive aligned courses, with the goal of enabling student teams to develop cognitive apps as part of the Watson Ecosystem,  (2) help prepare faculty and their top graduate students to submit aligned collaborative research proposals to funding agencies, with the goal of developing cognitive systems that boost the productivity and creativity of specific types of professions and professional teams (potentially a key part of what NSF calls “smart service systems”),  (3) help create linkages between faculty and IBM Researchers to define cognitive system grand challenges with clear measurable business and societal impact that might be achievable in the next 3-5 years, with the goal of further advancing the field of cognitive systems research and increasing aligned national research investments.  For example of a grand challenge, IBM’s Watson Jeopardy! system (2011) required close collaboration with seven universities to develop it, and it had a very clear measurable set of performance metrics that focused everyone across organizations and helped make decision-making easier.  For this last item, we are also (4) exploring academic interest in having an IBM-Researcher(s)-In-Residence at their universities, with the goal of accelerating collaborative research and achieving measurable grand challenge objectives.

Please let me know which of items 1 – 4 might be of most interest, or if your interests are in some other direction, and we can help guide you to the right IBMers to follow-up (for example, some universities already have data sets ready to be ingested, and investors interested in developing specific applications, so they are exploring the Watson Developer Cloud for Enterprise, as a path to get on site training and developer licenses more quickly).  Building cognitive systems to boost the performance of professionals, including research and teaching faculty at universities, is likely to be an important application area, and will have associated research grand challenges.

Jim Spohrer (spohrer@us.ibm.com)

It is time to re-read  Vannevar Bush’s “As We May Think” and Douglas Engelbart’s “Augmenting Human Intellect.”

Bush (1945) wrote: There is a growing mountain of research. But there is increased evidence that we are being bogged down today as specialization extends. The investigator is staggered by the findings and conclusions of thousands of other workers—conclusions which he cannot find time to grasp, much less to remember, as they appear.

Engelbart (1962) wrote: By “augmenting human intellect” we mean increasing the capability of a man to approach a complex problem situation, to gain comprehension to suit his particular needs, and to derive solutions to problems.  Increased capability in this respect is taken to mean a mixture of the following: more-rapid comprehension, better comprehension, the possibility of gaining a useful degree of comprehension in a situation that previously was too complex, speedier solutions, better solutions, and the possibility of finding solutions to problems that before seemed insoluble.

What advances are on the verge of reshaping how we may think and augmenting our intellect?  How might these advances contribute to a Moore’s Law for service science and smart service systems?  What are the related challenges and opportunities to this vision?

Smart phones are an everyday reality (not quite as envisioned in the 1930s, but close enough)…

Regarding how we may think and augmenting human intellect, my colleagues and I are working on a virtual community called the Cognitive Systems Institute (plan to launch next revision in September 2014) with one important goal being the creation of POVs to boost government and venture funding for cognitive systems research and startup development.   “Cogs” (Cognitive Systems/Cognitive Assistants) for boosting regional economic development will be appearing at an accelerating pace.  Gartner predicts 10% of all computers will be learning by 2017.   Many industries and professions will be disrupted including higher education.   The gist of the vision for the Cognitive Systems Institute is to have university researchers building “Cogs” for every profession, for every region and in every language, and with student teams launching startups using the Watson Developer Cloud.  IBM’s BlueMix on SoftLayer is the beginning of the API Economy leading to Cognition as a Service, which access to much cognitive componentry including Natural Language Processing on OpenPower and Pattern Recognition on SyNAPSE.  “Cogs” become our “Cognitive Bulldozers in the Era of Big Data/Internet of Things” because “Cogs” know individuals (you) and know professions (your job), boosting productivity and creativity (how to measure productivity and creativity for many professions is a challenge as well – I recommend this book). “Cogs” learn, use language to interact more naturally, and have levels of confidence in what they know and do not know – learning, language, and levels.   The Forbes Global 2000 companies generate nearly a 1/3 of the GDP of the world in just 2000 publicly traded companies’ revenue…. they have combined about 100M employees.  If there was a “Cog” to help each of their employees be twice as productive and creative – you move the needle on the GDP of the planet.  People enjoy being more productive and creative so quality-of-life might also be improved, if done right.  “Cogs” may well be a key part of a Moore’s Law for Smart Service Systems.

For those interested in helping, after re-reading Bush (1945) and Engelbart (1962), a practical next step would be to read this:
https://www.ibmdw.net/watson/docs/5-steps-powered-watson-application/

IBM has developed enabling technologies and proof-points, and is investing billions to realize this vision imagined by Bush and Engelbart. Aligning government funding and venture funding will enable university faculty researchers and student entrepreneurs to play a major role in making this vision a reality in the coming decade.

Jean Paul Jacob (IBM Research Emeritus) alerted me to this NYT article “Automation Alone Isn’t Killing Jobs” …

The author of the NYT article, Tyler Cowen, notes:

“Many other examples of automatable jobs are discussed in “The Second Machine Age,” a book by Erik Brynjolfsson and Andrew McAfee, and in my own book, “Average Is Over.” The upshot is that machines are often filling in for our smarts, not just for our brawn — and this trend is likely to grow.”

I especially agree with this point from the NYT article:

“There are unlimited human wants, so there is always more work to be done. The economic theory of comparative advantage suggests that even unskilled workers can gain from selling their services, thereby liberating the more skilled workers for more productive tasks.”

Ricardo’s Law of Comparative Advantage is very profound  – and greatly appreciated in the global service science community.  The advantages of interaction seems to be mathematically built into the fabric of universe, and certainly society.  Not just productivity (via Ricardo, etc.) but also creativity/innovation is enhanced by diverse perspectives and interaction (MIT’s Thomas Malone in his Science article on collective intelligence, and recent formulations by Pentland at MIT –  https://service-science.info/archives/3486).  This topic will surely be discussed at the National Academies Keck Center in this NSF-sponsored workshop in conjunction with the UIDP – research priorities for service science.

I especially like the perspective in this blog post by @JosieHolford  http://t.co/zdlS8g875m — don’t ask kids what they want to be when they grow up (that is 20th Century thinking – implicit a job and profession), ask them what problems they want to work with others to solve (21st Century thinking – implicit a team and purposeful effort) – their type of career may not exist yet.

If the transition to the industrial age is any guide, most of the jobs and types of careers in the age of smart machines do not exist yet.  The era of cognitive systems is just beginning to dawn.

How can a food truck become an example of a smart service system, known for creative new recipes?

Before explaining, consider recipes just a bit…

Recipes are new combinations and configurations of existing ingredients.  Some recipes are appealing, and stick around.

In his book “The Coming Prosperity: How Entrepreneurs Are Transforming The Global Economy” Phil Auerswald extends the notion of recipes in several interesting ways, including viewing entrepreneurs as people who bring new innovations to the world by seeking out new combinations and configurations of resources.

In “Technology: What It Is And How It Evolves,” Brian Arthur describes the evolution of technologies as new combinations and configurations of existing technologies, with occasional scientific discoveries expanding the list of ingredients.

Resource integrators are fundamental, and described in the Service-Dominat Logic literature of Vargo and Lusch.

In the emerging service science literature, built on the foundations of the S-D Logic worldview, service systems are defined as dynamic configurations of resources (people, technology, organizations, and information) interconnected by value propositions, for the purpose of value co-creation and capability co-elevation.  Service science is the study of the evolving ecology of nested, networked service systems entities, complex systems with capabilities, constraints, rights, and responsibilities.

Smart service systems are an emerging focus area for the National Science Foundation, and a first set of proposals are now under review.

As the era of cognitive computing and smart machines dawns, the people inside service systems will benefit from Cogs.  By boosting the productivity and creativity of people in roles in service systems, Cogs will lead to smarter service systems.

So how can a food truck become an example of a smart service system?

Nick Lavars explains it this way:

IBM has put its cognitive computing system in control of the menu at a food truck feeding attendees at this week’s SXSW festival and the appointment has resulted in some particularly imaginative dishes.

For its “Cognitive Cooking” project, IBM has enlisted the services of four prominent chefs to work with Watson, who is consulting a database of tens of thousands of recipes and ingredient combinations to conceive dishes that the regular foodie has probably never thought of.

“If you were to look inside the system that is running in the IBM cloud you would see a system that is trained on 35,000 different recipes, as if it was digesting a giant cookbook,” said Steve Abrams, Director of Watson Life at IBM. “From reading that cookbook it has learnt an awful lot about different ingredients that are often used in different cuisines, and the ingredients that are often paired together.”

So where is the world of Cogs and Smart Service Systems headed?  To get a glimpse of the answer, Ramez Naam has begun to articulate clearly some of the key issues.   The notion of “learning” that Naam begins to illuminate is a step in the right direction, especially when he remarks:

And, indeed, should Intel, or Google, or some other organization succeed in building a smarter-than-human AI, it won’t immediately be smarter than the entire set of humans and computers that built it, particularly when you consider all the contributors to the hardware it runs on, the advances in photolighography techniques and metallurgy required to get there, and so on. Those efforts have taken tens of thousands of minds, if not hundreds of thousands. The first smarter-than-human AI won’t come close to equaling them. And so, the first smarter-than-human mind won’t take over the world. But it may find itself with good job offers to join one of those organizations.

A grand challenge research problem in cognitive systems research and service science is to minimize the amount of energy, material , and time required to rebuild more and more capable systems (cognitive systems and service systems).   This could mean finding the optimal set of training data to boot-up a Cog that can assist a human in some particular job role (“how many recipes does a Cog need to read and understand before it can help a chef innovate?”).  Or it could mean finding the optimal population of Cogs to help a new Cog train up rapidly.   Or it could mean, how rapidly can one rebuild the entire societal infrastructure, and ecology of service system entities.  These are new fundamental problems in accelerated learning rates, for individuals in isolation (just data stream inputs), for individuals in a society of other entities, and for entire societal boot-ups.

In a few weeks a group of us will convene in Washington DC to develop a proposed research agenda for service science and innovation, with an eye towards how technology advances will enable smart service systems of the future.  Empowering the people in service systems to be more productive and creative with Cogs will surely be one of many topics for discussion.

In thinking about the jobs of the future and jobs in the age of smart machines, it is important to think about jobs where there is no end in sight of work to be done.   Scientists, artists, ethicists, entrepreneurs, explorers, coaches, and chefs are seven jobs that immediately spring to my mind.  Surely, there are many more.

“Cogs” can be thought of as a new species of intelligent agents (“smart machines“) that can learn and communicate with us, and know a person and a person’s history very well.  For example, already doctors are starting to recommend healthcare apps to patience, and perhaps soon they will prescribe apps, and insurers will pay.  However, “Cogs” will have capabilities far beyond simple apps.

Perhaps like our pets, Cogs know us; but unlike pets, Cogs process natural language and do pattern recognition as smart machines – with capabilities that are on a tremendous improvement trajectory, driven in part by Moore’s Law.  Cogs will be one of the key innovations in the era of cognitive computing, or what some have called The Second Machine Age.

The IBM Watson Group is building APIs that will be available in the IBM Cloud (SoftLayer) — and people will be able to build personal Cogs as well as professional Cogs for various applications (each job type or professional will have a GenericCog, that can be customized to know a specific person in a job role).

For example, imagine a ChefCog, it already exists, helping chefs create and prepare amazing new recipes.  If a chef has some unique style, then a ChefCog would adapt and help that person be more creative and productive.  Or like free-style chess, imagine teams of chefs and their ChefCogs all working together to advance the culinary arts.

One can also imagine DoctorCogs, MedicalEducatorCogs, CustomerServiceCogs, and SalesCogs, all helping people be more creative and productive – and like the ChefCog, early versions of all these types of Cogs already exist.

So you may have a Cog that knows you holistically, and then a separate Cog for each of your roles in life (e.g., various service systems – at home, at work, at school, at the hospital, etc.).

Technically, the set of Cognitive Systems includes entities such as people, pets, and Cogs…. but also larger entities, such as companies that know you and communicate with you and can learn about you like Facebook, Google, etc. as well as nations, states, cities that know you as a citizen and provide service offerings customized to your needs.  Any entity that stores information about you and builds up a profile that helps the entity interact with you naturally to co-create value and co-elevate capabilities can be viewed as a type of Cog.

So Cogs are somewhat like a new species, without rights or responsibilities (so not a formal service system entity yet), but definitely with capabilities and constraints.

The relationship: All service system entities are cognitive system entities, but not all cognitive system entities are service system entities.   The set of Cogs is a subset of the set of Cognitive Systems, but until Cogs have rights and responsibilities – they are disjoint from the set of Service Systems.