Pursuing the Endless Frontier in the Era of AlphaGo

Delivered at the THE Asian Universities Summit on the 7 February 2018 by Professor Alice P. Gast

 

This has been an impressive gathering and it is an honour to be here to speak to you today.

I am always inspired during my visits to China and throughout Asia. I find the vitality of the intellectual climate in the region very energising.

My Imperial College London colleagues and I greatly value our collaborations with peers in Asian universities. Over the years, we have seen these collaborations grow and thrive and we cherish our many friendships with academics at universities throughout the region.

It is wonderful to see so many friends here today.

As Times Higher Education points out, Asian universities are gaining strength.

They continue to expand their collaborative networks throughout Asia and throughout the world. They continue to produce large numbers of graduates in science and engineering; graduates who go on to do great things in academia, business and government.  And, their research is increasingly appearing in high-quality scientific research publications.   

There is strong government support of R&D in much of Asia.

And it is clear that this support helps universities produce first-rate research and provide a high quality of education to a growing number of students.

Asian government investments in R&D bode well for the future of Asian universities, and, I believe, for the rest of the world.

These are exciting times. The technological revolution that is sweeping the world is a watershed moment. It holds great promise for advancing knowledge and addressing many of the pressing societal issues faced by all nations. It is a time when investment in university research is more important than ever. 

As we have just heard, Universities, as providers of intellectual capital, will be leaders in the so-called fourth industrial revolution. 

This is yet another reason that Asian universities should look to the future with great anticipation and expectation.  We are already beginning to see the revolution in action.

Thanks to the growth in computer memory and computing power, artificial intelligence, or AI is beginning to demonstrate its promise to be a transformative tool for researchers, students, businesses and society.

There was perhaps no better public demonstration of this promise than when DeepMind’s AlphaGo defeated Lee Sedol, a professional player of 9 dan rank, in 2016.

The play drew the attention and admiration of the global GO community.  And it captured the imaginations of academics throughout the world.      

As you know, Go, or weiqi is a complex game originating in ancient China. It requires intuition and subtle thinking. Players spend years studying strategy, playing and learning from a master. It is impossible for a computer to play by brute force due to the 10^170 possible moves. 

Champions like Lee Sedol study with masters for years and build their personal approach to the game upon that foundation.  Great players develop the ability to combine proven, strategic moves with creative and unexpected moves that surprise their opponents and put them off balance. 

AlphaGo studied thousands of amateur games to learn the rules and then played itself millions of times, learning from each mistake and refining its moves. This reinforcement learning with neural network algorithms is a very effective way to optimise machine learning. 

The creative moves that AlphaGo made came from being free from biases passed down from generation to generation by Go masters.  It is an astonishing thought that computers can be “creative”. 

More recently, AlphaGo Zero has mastered Go without studying human games.  Instead, it starts with the basic rules of Go and competes against itself to learn.  As they put it, they have “removed the constraints of human knowledge”.  

This is both exciting and a bit frightening. This ability for machines to learn from their mistakes is especially interesting.  Rather than simply performing routine, non-dynamic tasks, machines have demonstrated their capability to deal with non-routine, complex matters. This opens a broad, new range of potential future applications.     

The ability to start with basic rules and not learn from human examples poses an interesting question about how humans should collaborate with such machines. 

At Imperial, we are collaborating with DeepMind on an important project to improve clinical data sharing, task management and quality care in the UK’s National Health Service. This is important given the explosion of the information and data in medical care. We have a critical need to help our healthcare providers optimise their time in caring for patients.   

Professor Nick Jennings, our Vice Provost for Research and Enterprise works with Rescue Global to pull together disparate streams of information during an emergency.  Working with AI/Machine learning algorithms, they can optimise emergency response and deploy limited resources where they are most needed. 

These are but two examples of the utilization of AI in the world.  We are seeing the important role that AI is playing in physics, healthcare, emergency response, financial analysis, materials science, and many other areas of importance in Asia and throughout the world. 

We are also seeing companies beginning to use AI to increase efficiency, reduce costs and improve experiences.

In 2016, Unilever began using a combination of AI and social media to review candidates for internships and entry-level jobs.   

Their approach replaces campus visits, eliminates the submission of resumes, reduces the number of interviews and increases the efficiency of the search process for both applicants and the company. 

People who click on a Facebook ad are directed to a company site where they can provide a few basic pieces of information, including their LinkedIn account. An algorithm sorts through the applications to find candidates who meet job requirements.     

The candidates selected then play a set of online games based on cognitive neuroscience. The games are designed to assess memory, the ability to focus, aversion to risk and various other characteristics important to the workplace.

The candidates that make it to the next step use their computer, smartphone or tablet to respond to questions from an avatar about various situations or issues that may come up in the workplace. AI is used to evaluate how quickly they answered the questions, their choice of words, and their facial expressions.

Finally, those selected at this step are invited to an in-person interview. 

Applications doubled in the first year. Yet the average time for a candidate to be hired dropped from four months to four weeks. The company reduced its costs of recruiting.  And it lowered the window of uncertainty for candidates. Applicants learned sooner rather than later of the company’s potential interest in them.

AI also has applications that can help address societal issues and enhance lives on a personal level.

At Imperial, researchers are working with partners in Europe using AI and audio, visual and facial recognition technologies to help children with autism improve their learning and emotional understanding.   

Those affected by autism spectrum disorder have difficulty in expressing their emotions and understanding the emotions of other people. They often find human interaction confusing and stressful. 

The technological platform developed by the Imperial team can decipher sounds and visual clues.  It is connected to a robot that interacts with the children. Those with autism spectrum disorder find the robot easier to interact with than another human. And because of that the robot can conduct simple lessons that a human cannot.

The importance and utilisation of AI is growing. In Asia, it is used in a number of ways:  

In China, hospitals use virtual doctors to read CT scans.

In Thailand, AI is being used to improve the care of cancer patients. 

In Singapore, room service robots powered by AI assist in customer service.

In Vietnam, the potential of AI to help farming is being explored.

And in Indonesia, AI is used to improve customer telecommunications services.

The Fourth Industrial Revolution is underway.  Artificial intelligence is here.

At the same time, it presents challenges because of the pace of change and the disruptions that it will create.

What will its impacts be on the future of the workplace? 

How will it affect teaching?

Will universities be able to respond effectively to the the technological revolution?

University leaders have important roles to play in helping to manage the changes that define these times. These changes will affect our students.  They will affect our staff.  They will affect our neighbours. And they will affect society as a whole.   

What is the role of universities in the age of AlphaGo?

I think it is useful to look back in time…. to look at another watershed moment… for some of the answers.

Over 70 years ago, Vannevar Bush--an American engineer, academic, civil servant, and MIT entrepreneur—submitted a report to the President of the United States, a report entitled, “Science, The Endless Frontier”.  

Science, The Endless Frontier is a wonderful phrase that captures the essence of the pursuit of knowledge. Starting points are known.  End points are not. 

Bush was asked by President Roosevelt to recommend what America should do to improve the national health, create new enterprises bringing new jobs, and better the national standard of living. The idea was to capture the advances in scientific knowledge made during the war and ensure that these advances would continue in peacetime.

Science, The Endless Frontier led to the creation of the National Science Foundation with its mission to support fundamental research in science and engineering in the US.

The values and visions contained in Bush’s report have lasting importance not just in the US, but across the global higher education community.  His insights are as relevant today as they were in 1945.      

Bush said that science is a proper concern of government

He advocated for the importance of fundamental research.

He urged the removal of financial barriers to education.

He underscored the value of international collaboration.

And he emphasized the importance of competent and inspired teaching of science.  

Chuck Vest took on the mantle of Vannevar Bush.  During his 14 years as president of MIT, he was a forceful and untiring advocate for the importance of research and the need for government support of fundamental research.

Each year of his presidency, Chuck wrote a letter to the MIT community in which he shared his thoughts on the issues of the day.  These were later published as a book titled Pursuing the Endless Frontier: Essays on MIT and the Role of Research Universities.

His letter in 2000 was titled: Disturbing the Educational Universe: Universities in the Digital Age-Dinosaurs or Prometheans? It dealt with the role of information technology in university education. He wrote:

“Does the future of education, learning, and training belong to a new machine-based digital environment, or will the best learning remain a deeply human endeavor conducted person-to-person in a residential campus setting? I believe the answer is “Yes”—to both.”  

In the same letter Vest also wrote, “Machines cannot replace the magic that occurs when bright, creative young people live and learn together in the company of highly dedicated faculty.”

He added, “…. as universities find their way in the digital age. …. our emphasis should be on one thing-the enhancement of learning.”

Of the many wise things Vest said, perhaps the most relevant point today is:

“Universities are our primary vehicle for educating talented men and women and for producing new knowledge, insight, and techniques. In order to serve well, universities must balance continuity and change—continuity of their deeper values and guiding principles, and commitment to intellectual excellence and the life of the mind, are essential.”

These are what I take from Bush and Vest as the four core values for universities in the age of AI, the age of AlphaGo. 

1. Commitment to Fundamental Research

Continued support of fundamental research is essential. To pursue The Endless Frontier requires a commitment by governments to provide funding for research in areas that expand knowledge even when near-term applications cannot be discerned. Basic research fosters innovation and long-term competitiveness. The support of fundamental research represents a belief in talent and potential. And the ability to demonstrate patience.   

To use one more quote from Chuck Vest, “Ignoring fundamental research is just another way of living for the short term at the expense of future generations.”

Asia’s investment in R&D has been essential for its rapid economic advancement. But a lasting commitment to basic research is essential to its future. Some countries in the region are making impressive strides and committing to increase funding of basic research. All countries should make the funding of basic research a priority in the coming years.     

2. Commitment to Teaching 

The most important function of a university is education.  Students are our most important and lasting output.

Our responsibility is to impart knowledge, and in doing that to provide students with the intellectual ability and confidence to adapt to changing times. We need to produce talented graduates that companies, universities and other entities need in the changing technological environment.

Our students, in all disciplines, need to be comfortable and conversant with AI.  They need to understand both its potential and its limitations.  We need to be alert to new ways to incorporate AI into the university curriculum.   

Research and teaching are intertwined. Advances in research have and should continue to be shared in the lecture halls and labs. AI has the potential to bring professors and students closer by paving a two-way street between what is shared and what is questioned.  Students need feedback, and AI can provide analysis to augment the judgement that professors provide.

3. Commitment to Collaboration

Collaboration is more important than ever.  The explosion in data in areas such as healthcare, our microbiome, financial markets and cosmology is both exciting and daunting.  AI is a tool to extract knowledge from these superhuman scales of information. 

It will also provide a growth in opportunities for universities to collaborate more broadly than ever before. 

The increased ability to work in new ways on complex issues will require high-level skills that are unlikely to be found within one university, one sector, one country or one region. The smart use of AI will require the expertise, insights and perspectives from a disparate range of people.

International, interdisciplinary collaborations among universities, businesses and governments will become increasingly important. 

It is also important for universities to collaborate more actively with their local communities.

4. Commitment to a Global Perspective   

While AI has the potential to be a major, global force for positive social change, it presents challenges as well as opportunites.  

In an address to the participants in the recent AI for Good Global Summit, UN Secretary General António Guterres offered the following caution:

“Artificial Intelligence can help analyze enormous volumes of data, which in turn can improve predictions, prevent crimes and help governments better serve people. But there are also serious challenges, and ethical issues at stake. There are real concerns about cybersecurity, human rights and privacy, not to mention the obvious and significant impact on the labor markets. The implications for development are enormous. Developing countries can gain from the benefits of AI, but they also face the highest risk of being left behind.”

In order for AI to help create a better future for all countries, and for all individuals, we will all need to support and guide its development.

University leaders must be leaders in addressing the disruptions created by AI and in helping to broaden its positive impacts. Universities must be places developing beneficial approaches to the societal stresses that the technological revolution brings.

We must continue to be sources of knowledge for all parts of the world and for all parts of society. 

All of us must think about how we are going to educate the next generation of students to succeed in this new world. All of us need to find new ways to work with governments and companies. All of us will try to make the very best use of technology in our research.

We need to be leaders, not just participants, in these times of change.

Our students and our research must be more than simply relevant to the times.  We should steer the way forward. 

The investment of Asian nations in basic research is essential for them to capture the opportunities of the technological revolution. 

Beyond the investment, it is the commitment of universities to the core values of higher education that will ultimately determine the societal benefits of this revolution. 

It is these values that should guide all of us as we pursue the endless frontier.