In April 2006, Google launched the first iteration of its Translate service. Immensely popular right from the beginning, it was also the source of many memes and jokes — its literal translation of words couldn’t account for the metaphors, imagery, and abstract nature of human language and communication. Over years of usage, it collected a substantial amount of data and Google taught the software how to ‘learn’ using a neural network. Today, the service is capable of listening to speech and translating it instantly into text and voice in the target language. Like humans—in fact, much better than humans—neural networks are enabling software to engage in ‘deep learning’, a data-driven exercise that allows them to evolve and do their jobs better. This paradigm shift has opened up a world of possibilities for AI and its applications across several fields and industries.
The marvel of neural networks
Neural networks recreate and mimic the architecture of neurons in the human brain, and software leveraging neural network display a capacity to learn to recognize patterns with much greater speed and accuracy than ever before. Prior to Watson, IBM had built Deep Blue—the first chess-playing computer to ever defeat a reigning world champion. Capable of evaluating 200 million positions per second, the computer narrowly defeated Garry Kasparov. But other than playing chess, it couldn’t learn to deploy its massive processing power anywhere else. Watson, on the other hand, utilizes Deep Learning and neural networks. It not only beat world champions at Jeopardy—a TV gameshow that requires no little amount of abstraction and understanding of metaphors to win – but was eventually taught to recognize patterns in medical scans, and is more accurate than an average human radiologist at diagnosing ailments through scans.
The opportunity ahead
The ability of these computers to learn and adapt is an absolute game changer. Industries are already beginning to deploy neural networks in their conventional business processes; much of Wall Street has been overtaken by AI-driven high-frequency trading software, and a robot that leverages Watson’s AI has recently been hired as the first robot lawyer by a British law firm. With the proven width and scope of Deep Learning AI, no field of human endeavour is closed to technology—or to the humans who know how to build such Deep Learning software.
A recent FICCI and NASSCOM report stated that 37% of the Indian workforce would require radically different skill sets by 2022, and that nearly 65% of the current organized IT/BPM sector would have to be redeployed into profiles such as that of AI research scientists. The Indian economy and workforce would need to adapt to this drastic change, and there will be a scramble for AI and Deep Learning proficient candidates to enable these companies to deploy Deep Learning into conventional models and processes.
The ability of these computers to learn and adapt is an absolute game changer. Industries are already beginning to deploy neural networks in their conventional business processes.
Deep Learning and your career
For many of those people who have chosen careers far from technology and programming, the idea of picking up AI and Deep Learning proficiency might seem like too big a leap. However, we now know that AI will soon be entering into nearly every sphere of human life. There are applications for Deep Learning AI in healthcare, education, manufacturing, finance, retail, logistics, transportation, infrastructure, real estate, and even politics and governance—basically, every known and future industry or space of human enterprise. Learning about AI and programming now and riding the wave of change, instead of getting overwhelmed by it, would be a better choice for your career. In any case, there is no need for rigorous computer expertise to learn Deep Learning—your domain knowledge in your discipline and some basic logical/analytical skills will suffice. With the requisite skills in your current domain coupled with Deep Learning proficiency, you can build a neural network into your company’s processes with relative ease.
Imagine being able to create an AI program that can do nearly all the more mechanical aspects of your work, and slowly learn and grow to become more proficient over time. Soon, you will find yourself able to get a lot more work done, focusing on the more creative and engaging facets of your work. Regardless of your industry or job role, proficiency in Deep Learning will increase your value to current or prospective employers and make you a valued resource in any official capacity. Also, gaining such proficiency is far more accessible than you imagine—there are quality learning resources on this field available on the internet, especially at lifelong learning platforms that provide access to world-class faculty and instruction. So what are you waiting for? Get with the Deep Learning programme today.
Author is MD, India of Udacity