A remarkable number of innovations in technology have been made by humans who were tinkering in their garages. A remarkable number of scientific breakthroughs have been made by humans who got their starts as youngsters mixing chemicals in the basement, or as amateurs who took up their pursuits (stargazing, for example) as a hobby at a later age.

Before you classify yourself outside the group of mad scientists poised on the edge of 21st century discoveries, consider the following ways you can participate in what Ray Kurzweil calls "accelerating intelligence." This column has previously mentioned some crowd sourcing avenues and online games for do-it-yourself artificial intelligence, but more opportunities are becoming available.

Because AI is a young field, it's possible to learn the basics and begin to explore further with the investment of a few hours. You can be an AI scientist without straining your brain, without learning to program, or, if you prefer, you can stretch your limits. Here's how.

Help bring Siri out of beta

Siri is the new iPhone intelligent personal assistant, currently in beta. Siri's voice recognition artificial intelligence is dynamic and improves as you and others give Siri exposure to more variations of a language, including dialects and accents.

Enter challenges and contests

The Defense Advanced Research Projects Agency (DARPA) and other government, academic and commercial organizations have increased the number of their AI challenges and contests that are inexpensive and open to the general public. For many of these contests, no prior AI or science background whatever is required. Creative and simple ideas often prove to be better solutions than traditional approaches. Winners report being surprised that they won. Here's a sampling of recent contests:

A small San Francisco-based team won DARPA's challenge to reconstruct shredded documents, competing against almost 9,000 other teams, by using custom-coded computer vision algorithms to suggest fragment pairings to human assemblers for verification. DARPA extended its challenge not only to computer scientists but to "puzzle enthusiasts and anyone else who likes solving complex problems." Some experts did not think a solution would be possible.

In 2009, an MIT team won DARPA's Red Balloon challenge, which required participants to locate 10 large, red balloons at unknown U.S. locations. The MIT team entered the challenge at the last minute and from start to finish completed the challenge in less than four days, competing against teams that had been planning for months. The team used a website and social media to enlist over 4,000 worldwide collaborators in less than 48 hours, providing small financial incentives to collaborators who were part of the chain that located a balloon.

AI Challenge's "Ants" contest ended recently. Previous challenges were "Tron" and "Planet Wars." Look for future challenges. The Ants challenge began with a five minute starter kit that showed participants, including beginning programmers, immediate results. Programming tutorials helped participants complete their entries.

The U.S. Patent and Trademark Office, together with the NASA Tournament Lab, launched a contest to develop new algorithms to aid in patent examination. Contestants are currently working in teams of two to develop solutions and deliver an algorithm, requiring text recognition and image analysis, that can automatically identify and locate specific elements within patent documents. The grand prize is $10,000, but all active competitors win special T-shirts.

Learn in online classes - free

This writer recently completed Stanford University's non-credit, free online Introduction to Artificial Intelligence class. If I can do it, you can too!

Innovative and successful beyond anyone's expectations, this class reportedly originally attracted 160,000 students worldwide, and rumor suggests that 20,000 or so completed the course in the advanced track (doing weekly homework and taking mid-term and final exams, all online). Compare this to the average size of a university class.

The instructors, Stanford professor Sebastian Thrun and Google Director of Research Peter Norvig, taught in a series of short video segments that included many interactive quizzes to promote understanding. Online discussion groups operated on reddit and aiqus, a Q&A user-moderated forum set up especially for the class. Less formal groups ebbed and flowed on Twitter (#aiclass), Facebook and LinkedIn. In "office hours" videos, the professors answered questions proposed and voted upon by class members. This "distance learning" experience felt so intimate that many students reported sadness that it was ending with the final exam.

As a result of this class, this column will no longer contain passages such as: "This AI requires hidden Markov models (Do I sound like I know what that means?)" This is because I now know exactly what hidden Markov models are and what their significance is for AI. After refreshing my memory via Wikipedia about how to add and multiply fractions, I was surprised to find myself spending hours doing complex math calculations for hidden Markov models, Bayes networks and other seemingly opaque AI techniques so that I could complete the homework and answer the exam questions.

Considering my long-ago and limited math education, especially in probability, I finished the class with a respectable score, experiencing difficulty primarily with those pesky Bayes networks, especially in the final exam. Also figuring in the final exam was The Towers of Hanoi, a classic puzzle, mentioned in the October 2007 AI and Humans column. In that column I joked that to work on the puzzle one should start with 4 disks - and have some snacks on hand. I confess to snacking on chocolate chip cookies while working on the final exam, but I also found myself calculating the number of state spaces (valid configurations) in the Towers of Hanoi 4-disk version, and answered that one correctly!

Many class video segments remain available to the public on YouTube. Due to the success of this and two other Stanford AI classes offered free online last semester, Stanford is offering several new free classes in AI and other fields starting this month. Not to be outdone by Stanford, the Massachusetts Institute of Technology (MIT), an innovator several years ago with its OpenCourseWare initiative, will introduce similar online classes later in the spring. There's also a wealth of educational information free at Apple's iTunes U.

Choose a problem important to you and solve it

One of the best ways to learn computer programming or artificial intelligence or another new field, is to find a problem that interests you personally. Start to explore it, find others working on the same conundrum, and actively look for ways to solve it.

This approach is recommended by Introduction to Artificial Intelligence professors Thrun and Norvig, and is used successfully by MIT's Media Lab, which requires participants to jump into innovation by building prototypes. Listen-in and participate in online discussion groups

Listen-in and participate in online discussion groups

You can listen-in and participate in online discussion groups about AI on Facebook, LinkedIn and other social media sites. Search on the hashtag #artificialintelligence in Twitter and you'll find recent news items and developments to follow up on. Communication with others interested in your chosen problem could allow you to leap forward quickly yourself, or result in a collaborative solution.

Predict the future

The beginning of a calendar year is traditionally a time to predict the future. One of the traditions in AI, at any time of the year, is to make predictions about how soon AIs will reach the level of general human intelligence, out-perform humans at specific tasks (from cooking to complex project scheduling), or take over the world. No credential or scientific background is required to participate in this venerable AI tradition. What do you think artificial intelligence will be able to do in five years, ten years, twenty-five years?

Sources and additional information:

How to use Siri: www.apple.com/iphone/features/siri-faq.html

DARPA shredder challenge: www.shredderchallenge.com/

Frank Moss, The Sorcerers and Their Apprentices: How the Digital Magicians of the MIT Media Lab Are Creating the Innovative Technologies That Will Transform Our Lives, Random House, NY, 2011.

AI Challenges: aichallenge.org/

USPTO algorithm challenge currently underway: community.topcoder.com/ntl/?page_id=536

Videos of Introduction to Artificial Intelligence class segments (search on "AI class unit"): www.youtube.com

Free online Stanford University AI classes starting shortly: http://www.class-central.com/

iTunes U: www.apple.com/education/itunes-u/