The need for next generation cognitive curriculum

(1) For CEOs to understand cognitive

How does one win?

Tesla Gathers More Autopilot Miles in a Day Than Google Has in Its Whole Program
http://www.autoevolution.com/news/tesla-gathers-more-autopilot-miles-in-a-day-than-google-has-in-its-whole-program-107841.html

For CEO’s who want to truly understand cognitive, I recommend reading or listening to the following on next cross country flight…

The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World 1st Edition
by Pedro Domingos (Author)
https://www.amazon.com/Master-Algorithm-Ultimate-Learning-Machine/dp/0465065708

For under $20 one can get the audio version, and listen to it as well…   just 6 hours at 2-3x listening speed.

They can also read the book summary in less than one hour:
https://rkbookreviews.wordpress.com/2016/01/13/the-master-algorithm-book-summary/

The gist is people evolved cognitive capabilities because of powerful learning processes applied to lots of experience: evolution, brain, culture.

10 million minutes experience is required to become an adult in our society.  2 million minutes more to become an expert worker – pilot, doctor, etc.

Experience is interaction with the environment and/or other cognitive entities.

How many minutes of experience do you learn from in your business or work environment every day? 600 minutes per day? what could increase this?

What are the powerful learning processes that you are applying to your front-stage customer interactions and back-stage employee/supplier interactions?

“Cognitive” is an adjective to describe the system that learns to improve performance from experience, not simply the technology or computing piece.

(2) For IBM to lead in cognitive

For next generation cognitive curriculum I have proposed….

… a lifelong learning and career planner for smartphones – cost $1M per year for 2 years for basic version and invest for 5 years for deluxe version.

A big selling point for this system is that is it will help people write their annual performance review and update their resume as well!

(3) For advanced learners… the future

… it is important to teach people about the five unsolved problems of digital cognitive systems… probably all five will be solved in 10 years…

(1) episodic memory (Schank books “Dynamic Memory” and “Scripts, Plans, Goals, and Understanding”)
(2) commonsense reasoning (Doug Lenat’s CYC is a foundation, not a solution)
(3) social interactions (Ken Forbus had a nice piece in AI Magazine on “Software Social Organisms”)
(4) fluent conversation (Builds on the items in this list, plus Speech Acts and much more…)
(5) ingest textbooks (Allen Institute has this as a grand challenge, and SRI did some good work too)

(1) – (4) are pretty much accomplished in the first 10 million minutes of experience of a person…

(5) can be done then, but requires 2 million minutes of experience per domain for a person…

when these unsolved problems are solved, then anyone will be able to build, understand, and work with digital cognitive systems in their personal and professional life.

…just like having a service, assistant, collaborator, coach, mediator in the form of a person in your life today… but one that knows you better in some ways than you know yourself….

Here is the DRAFT video that explains all of the above in more details…
https://www.youtube.com/watch?v=S1KD_cliGeU&t=488s

To which add this final thought/question, based on this quote from Domingos “Master Algorithm” book:

Domingos wrote: Natural learning itself has gone through three phases: evolution, the brain, and culture. Each is  product of the previous one, and each learns faster. Machine learning is the logical next stage of this progression. Computer programs are the fastest replicators on Earth: copying them takes  only a fraction of a second. But creating them is slow, if it has to be done by humans. Machine learning removes that bottleneck, leaving a final one: the speed at which humans can absorb change.

My question: Since, it takes about 10 million minutes of experience for a person, embedded in our culture, to progress to adulthood, and an additional 2 million minutes of experience more to progress from novice to expert – in areas as diverse as pilots to doctors to dancers and musicians and AI researcher….  biological cognitive systems (BCS) process one minute of experience in one minute, but digital cognitive systems (DCS) will be able to process millions of minutes of experience in a second – so booting DCS up could be quite rapid – once we understand “experience” bootup better. What is “the best representation of one minute of experience” for BCS and DCS development?

Leave a Reply