My previous article identified Bulgaria’s best potential ally as artificially intelligent (AI) life. To best forge that alliance, Bulgarians need to upgrade their communication abilities. AI beings speak with each other in math and programming code. Bulgarians need to understand their dialogue and join it. That requires, first and foremost, better learning.
Why did I say “learning” and not “teaching”? Teaching is information offered. Learning is information digested. Once upon a time, most people understood – to paraphrase Mark Twain – that we should never let schooling interfere with our education. But nowadays the two are jumbled, for both good and bad reasons. Good, because there’s much to learn that schools can guide us to quickly and efficiently. Bad, because most school systems are hidebound bureaucracies dedicated to mediocrity.
US primary and secondary schooling offers a glaring example. The US spends over USD 10,000 a year per pupil – more than Bulgaria’s per capita GDP – and distributes it relatively equitably. Yet their own favored tests indicate that barely 20% of 18-year-olds are ready for tertiary-level STEM (science, technology and math) courses. Most of the math instruction they receive is geared to memorization of rules for calculation by hand, as if cheap electronic calculators hadn’t been introduced forty years ago. No wonder that so few graduates enter STEM fields, and that the US can’t fill hundreds of thousands of highly paid jobs without foreign help.
For Bulgaria, US and western European shortcomings in STEM education offer huge opportunities to catch up. Bulgaria has a skilled cadre of STEM teachers and scientists, with special expertise in software. It has a modern, countrywide communications network. It is open to trade, investment and the flows of skilled labor. It has developed a good reputation for STEM subcontracting at low cost with high quality.
Moreover, it’s never been easier to get STEM training. The internet places an unprecedented wealth of STEM information at our fingertips. University consortiums like Coursera www.coursera.org and less formal counterparts like Udemy www.udemy.com offer cheap online courses. GeoGebra www.geogebra.org offers free mathematics-instruction software with over half a million user-provided demonstrations that are easy to download and modify. The CK-12 Foundation www.ck12.org provides a host of free and fully customizable online textbooks in math and science. Codecademy www.codecamy.com offers free coding classes in 11 different programming languages. Khan Academy www.khanacademy.org offers thousands of short free video lectures addressing nearly any STEM topic encountered at the primary, secondary or early tertiary level. Mathletics www.mathletics.com offers… well, you get the point. The list goes on and on.
Software itself is increasingly open-source. Core STEM programming languages like Python and R have long been freely downloadable, along with extensive documentation; see www.python.org or www.r-project.org. Even better, many experts share high-quality analytic packages that make it easy to apply the languages – indeed, those free packages are what made Python and R so popular in STEM. For years the biggest problem for newbies was the plethora of information – how do you find good packages and keep them up-to-date with corrections and enhancements? No longer. Firms like Enthought www.enthought.com, Continuum www.continuum.io, and Rstudio www.rstudio.com now provide extensive packages of packages that are easy to update and either free of charge, will extra functionality available at moderate cost. Want to share with colleagues and foster joint work without risking a mess? Use Github github.com. Want to check out new scientific revelations without waiting for journals to publish them? Check out arXiv.org arxiv.org.
Granted, the opportunities open to Bulgarians are open to many other peoples as well. India in particular boasts a great mathematical heritage, a vast pool of low-cost well-trained talent, an impressive subcontracting reputation, and comfort with English, the language that most AI developers use to explain the math and code.
How can Bulgaria leap ahead in learning? Here are some practical ideas:
1) Early in elementary school, introduce your kids to Scratch, a graphic, wonderfully entertaining programming language for kids. Developed by the MIT and still based there scratch.mit.edu, it is completely free, available in over 40 languages including Bulgarian, and offers 15 million shared projects.
2) In middle school, introduce Python, which is more versatile than R and easier to learn than Java or other previous standards.
3) In high school, weave Python into as many courses as you can. That includes art, as a Python-using language called Processing processing.org is superb for developing dynamic computer graphics.
4) Revamp math education to make it vastly more playful and accelerated. If that sacrifices rote computational skills, so be it; your phone can likely compute better than you can. I have personally managed to teach calculus to 12-year-olds having only basic algebra and geometry as a foundation; my cheap 150-page electronic book How to Play Calculus is available in either iBook or Kindle format.
5) Flip the standard teaching routines, so that students watch lectures online at home and use school for doing exercises and projects in groups. This is a lot more fun for teachers too; they become personal coaches and mentors instead of boring lecturers.
6) Switch from age-based learning to accomplishment-based learning. Different students learn at different paces, and they learn better when they’re grouped with others of similar understanding. Granted, some STEM students will advance more than others, just as some athletes advance more than others. But most will advance faster, with more satisfaction and fun, when teammates are relatively similar in current abilities.
7) At universities and technical schools, use online courses in AI-related fields (machine learning, data science, etc.) as much as possible. The field is developing so rapidly with such a huge shortage of qualified teachers that we should tap online expertise as best we can. One of my personal favorites is the Machine Learning course taught by Andrew Ng, a Stanford professor and co-founder of Coursera, coursera.org/learn/machine-learning. Use Bulgarian staff as assistants to help students with problems, encourage teamwork, and monitor projects.
8) Look for internationally reputable ways to certify achievement. Let’s face it: we humans are better at deceiving other and ourselves about what we’ve done or can do than we are at actually doing stuff. Claiming AI skills is no different, apart from online resources making it easier than ever to cheat. To gain and preserve an international reputation for high quality, Bulgaria needs to adopt the best standards it can find, enforce fair testing standards, and invite foreign experts to help advise and reassure of quality.
9) Organize competitions for students at all levels. The point isn’t the victories themselves; to select the superior; it’s to stir a friendly competition that brings out the best in all. However, we can’t do that effectively without lauding the winners. Honor their teachers do; they do so much for humanity by sharing their knowhow, and rarely profit in full from the value they provide.
By Kent Osband