Why India needs Artificial Intelligence talent
By Ashok K Nag, consultant and retired adviser Reserve Bank of India*
In 2017, Ke Jie, the world Go champion, lost all three matches to AlphaGo, an artificial intelligence software. It was “…one of the greatest matches in history,” 19-year old Jie said after his loss. This contest between China’s Jie and a machine was held in Wuzhen, China.
Go is a 4,000-year-old board game for two, popular in Japan, China and Korea. Contestants win by placing 181 black or 180 white pieces on a board to corner the most space. (See photo.) The world’s top Go players make far more complex calculations than do grand masters in chess.
Go is played on a board checkered by 19 horizontal and 19 vertical lines, while a Chess board has 8 x 8 lines. A typical chess game ends after about 80 moves, while winning a Go match takes twice that number. The number of potential moves in a chess game is 10 x 123 zeros, known as a Shannon number. In a Go match it is triple that amount.
Rapid rise of artificial intelligence
In 2016, Lee Sedol, a top Go player from South Korea, also lost to AlphaGo. The contest was watched live by over 280 million people around the World. While Jie and Sedol’s losses can be viewed as a defeat for humans, the contests also displayed the recent major advances in computing being made by engineers. The matches brought wide attention to the growing, vital role of artificial intelligence (AI) software, which powered AlphaGo’s victories.
AI tools use neural networks, which are self-learning software systems, first proposed in 1944 by two scientists at the University of Chicago. It is only over the past decade that new computers were devised which are powerful enough to enable neural networks to solve massively complex problems, faster and more efficiently than a human brain.
AlphaGo is the creation of DeepMind, a London based company. It is “attempting to distil intelligence into an algorithmic construct (which) may prove to be the best path to understanding some of the enduring mysteries of our minds” says founder Demis Hassabis. In 2014, he sold the company to Google for a reported $525 million, but continues as chief executive. Google organized the two man versus machine contests as a display of its AI capabilities.
AI is automating jobs
AI will transform society in the 21st century in ways similar to the impact of the Industrial Revolution during the 19th century, says Carl Benedikt Frey, an economic historian at Oxford University. The use of AI tools will spread across the entire economy as well as have a major impact in the defense area, Frey and other experts say. Self-driving vehicles, for instance, will radically change the automobile industry as well as ground warfare.
Over the short term, Frey notes in his book The Technology Trap, the automation AI brings will lead to a loss of jobs, including for middle class professionals. This is already happening in a major way In the U.S. and China, through the use of a variety of hardware and software robots, in manufacturing, warehousing, retail, computing and office work and other businesses. So far, the “…despairing middle class has not resorted to physical force, but their frustration has led to rising populism and the increasing fragmentation of society,” writes Frey.
AI helping discover COVID-19 drugs
Over the long term, Frey says, the widespread use of AI will lead to major economic benefits. This is already evident in the medical field. A DeepMind software, for instance, diagnoses eye diseases as effectively as the top doctors. Another of Its software predicts the complex 3-D shape of proteins, which is transforming the invention of new drugs, including for COVID-19.
In February, researchers at MIT announced they used AI to screen more than a hundred million chemical compounds in a matter of days to formulate a new antibiotic drug. In laboratory tests on animals, the drug killed many of the world’s most problematic disease-causing e-coli bacteria, like that of tuberculosis, and including some strains that are resistant to all known antibiotics.
Use of AI-based tools revealed a molecule which could be one of the most powerful antibiotics ever discovered. The lead author of the study, which was published in the journal Cell, is Jonathan Stokes, a post-doctoral fellow at the Broad Institute, run jointly by MIT and Harvard university.
America leads China in AI
There is an intense, competitive global race to develop AI tools for commercial as well as defense purposes. Google, Microsoft, Amazon, Facebook, IBM and other major American technology companies are investing billions of dollars to find ways to use AI to grow revenues and profits. Google, for instance, is using it to improve its search results, pricing of digital advertisements and personal assistant software and for speech recognition, automotive self-driving and other products.
The giant American companies lure the best AI talent from around the world. They offer a starting annual salary of $500,000, plus a large signing bonus, for a Ph.D. in AI from a top university such as MIT and Stanford in the U.S., Oxford and Cambridge in the U.K. and the University of Toronto and McGill in Canada.
Alibaba, Tencent, Huawei and other giant Chinese technology companies are similarly spending billions of dollars on AI. They attract the best talent within China by paying attractive salaries. In addition, both in China and the U.S., there are several large companies focused mainly on AI.
Currently the U.S. leads the world in AI talent and hence in AI. An indication of the talent is that, in 2018, nearly half of the 22,400 authors, published in 21 leading AI journals and presenting papers at major AI conferences around the world, earned a Ph.D. in the U.S. Authors from China were 11%, followed by the U.K. with 6% and Germany with 5%. About a fifth of the authors have a major impact in their field, based on citations of their papers. Of those, a quarter were from the U.S., while 6% were from China.
These are findings from an annual study done by Element AI. It is a leading AI firm based in Montreal, Canada, with Microsoft as one of its early investors, since its founding in 2016.
China’s ambitious AI plans
Chinese president Xi Jinping’s wants his country to lead the world in AI by 2025. His government is taking several steps, including spending big sums, to attain the goal. In 2018, China spent an estimated $8.6 billion on AI research and development (R&D,) in both defense and commercial areas. In contrast, in 2020, the US government will spend about $5 billion. But these numbers, especially in the case of China, likely understate the actual amounts spent for defense purposes.
China has a big potential pool of AI talent, given its large population and emphasis on math and science education, starting from elementary school. Last year, 180 colleges in China were teaching AI and related majors for undergraduate students. More than 60 universities in China offer Masters’ and Ph.D. programs in AI. Each year they graduate over 1,200 students with advanced degrees.
Leading Chinese universities are paying two to three times what Stanford and MIT pay their faculty in an effort to attract the best researchers from around the world in AI, as well as in other advanced technical fields.
Only IIT Hyderabad offers AI degrees
Since the 1950’s, 23 Indian Institutes of Technology (IIT) have been set up and funded by the government of India. The computer science programs at the IITs in Hyderabad, Bombay, Madras, Kanpur and Kharagpur and at the Institute of Science (IISC) Bangalore are ranked 78 among the top 100 in the world by CS Rankings, a metrics-based system.
Each year, the IITs admit about 10,000 students, based upon one of the most competitive entrance exams in the world. They admit an additional 8,000 for their masters’ programs, though the criteria are much easier, including with students sponsored by the government and companies. Both these programs can be a source of AI talent, with some additional training.
It was only last year that IIT, Hyderabad, became the first reputed college in India to offer under graduate and masters’ degrees in AI. At the other IITs, AI courses are offered only as part of their computer science programs.
Some of the IITs offer a specialization in the AI related field of data science. Since 2017, IISc offers a two-year Masters’ program in computational and data science. The Indian Statistical Institute (ISI), Kolkatta, another top college, offers a similar program. Both IISc and ISI are also government-run institutions. While they provide students a monthly stipend to cover their costs, each year they admit a total of only about 100 students.
The most important part of the AI talent chain is top quality Ph.D.’s. Most of the IITs as well as IISc and ISI have reputed Ph.D. programs in mathematical algorithms, computer science and statistics, which are all related to AI skills. But currently the number of students in such programs are less than a hundred. IIT, Bombay, which has the best computer science department in India, has only seven faculty members working in AI field.
So, it is not surprising that India only has about 500 researchers with AI skills, according to the Element AI study. Given the lack of AI programs at the top engineering and science colleges, there is also very little talent in the pipeline.
Need a virtuous cycle of good faculty and students
Unlike in the U.S., India does not have top colleges, funded by donors, which are keen to find and fund the best talent in the world to study for their AI degrees. And unlike in the case of China, the government of India is not funding the development of AI talent.
In fact, in 2019, the government doubled the fees for the masters in technology degree at the IITs to Rs. 202,500 ($2,700) per year. It also stopped providing students with a stipend to cover lodging and food expenses. As a result, it is likely that many talented students who cannot afford to pay the fees and costs, may decide against enrolling in the program.
India can start building its AI capabilities by funding programs at the IITs and other top technology institutes. The faculty should be allowed to set up, run, advise and consult for businesses as well as recruit students. This will attract good talent, especially as part-time faculty. A reputed faculty will bring in good students. The major banks, which are mostly government-run, should provide low interest loans to students who are unable to pay the fees and other costs. Students, even in high school, can begin learning AI skills on their own through online courses and certification offered by IBM and other companies.
This process will create a sufficient mass of talent and trigger a virtuous cycle: reputed faculty, quality courses, top students, a culture of business start-ups and successful entrepreneurs funding more faculty and courses at the institutes.
For this to happen, India’s ruling politicians must realize that spending on AI may not translate into votes in the next elections. But it will vastly improve the country’s defense systems and health care services, and, over the long run, help modernize the economy.
To receive Global Indian TImes stories each week:
send your name and email to: gitimescontact@gmail.com
Or connect through Twitter or Facebok:
https://twitter.com/GlobalIndianTi2
fb.me/GlobalIndianTimes
*Ashok Kumar Nag is a Mumbai based consultant in information management and data analysis. He spent over 25 years in the Statistics and Information Management department of the Reserve Bank of India, retiring as an adviser. Ashok is an alumnus of the Indian Statistical Institute, Kolkata, India. Link
