Computer scientist Arthur Samuel is rumored to have said that machine learning is an aspect of his field that gives “computers the ability to learn without being explicitly programmed.” That’s why machine learning is also considered an element of artificial intelligence, or AI, which deals more generally with how computers can figure things out for themselves. Essentially, the idea is that, given a good set of starting rules and opportunities to interact with data and situations, computers can program themselves, or improve upon basic programs provided for them.
In the mid-1980s, computer scientists hoped to reshape computing and the ability of computers to understand and interact with the world. There was a huge infusion of interest, enthusiasm and cash at that time, but AI did not change the world as we knew it then. Over time, AI was found to be suitable for a relatively narrow set of computing tasks, such as creating viable configurations for complex computes. But AI neither set the world on fire nor redefined its boundaries and shape.
More than 30 years later, AI in general and machine learning are enjoying a spectacular renaissance. These technologies are being successfully applied to deal with all kinds of interesting problems in computing, and are enjoying a broad range of success. Notable accomplishments for machine learning include email filtering, intrusion detection, optical character recognition and computer vision. Machine learning and AI have proven quite effective in applying computation statistics to use data analytics to make predictions and spot trends.
Machine learning is hot, hot, hot
Because some companies build or use technologies that employ machine learning and AI, there has been considerable demand for skilled and knowledgeable researchers and developers. But if anything explains a sudden, sharp spike in demand for such people, it’s the increasingly pervasive use of predictive analytics across many fields of business. Most of the Fortune 500, and a great many other companies and organizations outside that fold, are now using predictive analytics to seek a competitive edge, or to improve their overall ability to deliver goods and services to customers, clients or citizens.
Individuals trained in machine learning are now in considerable demand across the entire employment spectrum. That explains the six-figure salaries that are increasingly the norm for those who land such jobs. Of course, for many who already work in IT or who are heading in that direction, this raises the question of “how can I get a job in AI or machine learning?” The answers are straightforward, if somewhat labor-intensive and time-consuming.
The traditional approach: Get a degree
The field is intriguing for many who may also have a bachelor’s degree in computer science, engineering or some similar discipline under their belts. In fact, it’s hard to find a reputable graduate computer science program that doesn’t include machine learning amidst its targeted subject matters. If you want to aim for the stars in taking the back-to-school route toward machine learning proficiency, I’d recommend consulting either Quora’s thread on the best graduate schools for machine learning or U.S. News & World Report’s list of the best artificial intelligence programs as good places to start looking for candidate schools.
Make the most of MOOC offerings
For those who can’t break away from life and work to pursue a full-time degree on campus, massively open online courses, aka MOOCs, offer a variety of alternatives. MOOCs can encompass actual degree programs at reputable universities, certificate programs that provide ample training but don’t confer a full-fledged degree, or mapped-out curricula in machine learning or AI that cover the ground in as much depth as one might wish to learn the subject matter.
A quick search on machine learning at the MOOC Search Engine produces millions of hits that include the following:
- Udacity offers hundreds of courses of varying length, complexity and depth in this area.
- edX’s machine learning offerings include a certificate program from Microsoft, as well as numerous graduate-level courses and curricula from well-known colleges and universities.
- MIT offers a plethora of online courses in this area, for paid-for college credit or free online audit.
- Stanford also offers a collection of machine learning courses for credit or audit.
Hands-on is where learning gets real
There’s no substitute for rolling up your sleeves and digging into development work if you want to really understand the principles of AI and machine learning. Expect to devote yourself to your mouse and keyboard, as you start small with toy data sets and basic applications, then work your way up to more serious, real-world problem-solving and solutions. The capstone project for the Microsoft Professional Program in Data Science (not a degree) runs for four weeks, for example, and challenges you to develop a solution to a data set using machine learning to test your skills.
Anyone who digs into this subject matter should anticipate spending upward of 15 hours a week on programming tasks, in addition to attending lectures, completing reading assignments, writing papers and all the other tasks that modern learning demands of students nowadays.
When you’re ready to rock, let the world know
Once you’ve finished that degree, obtained your certificate or knocked off a significant chunk of curricula, you can start positioning yourself to current or prospective employers as someone with skills and knowledge in machine learning and AI. Unless you also have picked up some hands-on, real-world experience in reaching this professional milestone, remain humble about your skills and abilities in this arena. Warnings aside, the prospects for those who can see themselves through the time, effort and expense of mastering machine learning and AI should be bright.
Read the source article at Business News Daily.