Significance of Data Science Education at a university specializing in the social sciences
March 5, 2021
Hitotsubashi University Faculty and Graduate School of Social Data Science (tentative) will welcome its first class of students in the 2023 academic year. Even so, in 2020 before the official opening, the university established a course entitled “Introduction to AI” open to students in all grades. This course has proven very popular. We sat down with course instructor FUKUDA Haruaki, Associate Professor at the Graduate School of Business Administration, and asked him about the course objectives as well as the significance of data science education at a university specializing in the social sciences.
Just the beginning of data utilized to create new value
Associate Professor FUKUDA Haruaki, Graduate School of Business Administration
――Could you tell us what the course objectives are and give a general outline of the course?
FUKUDA: This course provides an overview of the principles and functions of AI and data science methods. The objective of the course is for students to learn methods and approaches for using computers and information technology to resolve problems, which will be useful for them in pursuing more specialized study in the future. The main focus of the course is on practical exercises where students write actual programs. Mathematical concepts are also necessary, but prerequisite knowledge is not required.
――What is the significance for Hitotsubashi University students to learn about data science?
FUKUDA: Data science is the fourth science paradigm after science grounded in Aristotelian experience, science based on Newtonian principles, and science based on computer simulations. Data science has also been called the “fourth industrial revolution.” A large quantity of diverse data has been accumulated as the internet has evolved and AI and IoT utilized far and wide around the world. Nevertheless, there is a sense that the use of data to create actual new value has only just begun. This seems to be because such operations are currently conducted mostly within a sphere visible only to IT specialists. As people possessing a greater knowledge of our society and what it entails acquire data science techniques, this evolution will expand the potential applications of data science.
We hope that teaching data science to Hitotsubashi University students and educating the people who will sustain this potential will produce significant value for our world.
To achieve that aim, it is important for students to learn how to use currently-available technology tools to analyze present challenges. Yet, what I would like to do in the “Introduction to AI” course is to take it one step further so that students develop the ability to continually incorporate technologies that come available in the future and apply these as they discover and solve problems encountered in the future. In other words, I would like to present evidence of a type that is different, both scientifically and socially, from what has been offered so far in order to develop in students the capability to come up with solutions to social challenges. Even though we can learn using case studies, such as those company data scientists are currently working on, I do not believe such studies would be very helpful in encouraging students to think about the future.
Social demand for ideas conceptualizing the “it would be awesome if we could do this” type of thinking
――Will we be able to solve social challenges and create new value by having social science professionals master the use of data?
FUKUDA: Yes, we can do that. I believe it is social science professionals, not information science experts, who are able to offer solutions to a variety of social issues. AI and IoT are being deployed everywhere. Yet, data science is currently perceived as something difficult and unwieldy. But, make no mistake, in the future, the benefits that these technologies reap will help our society.
――Can you give us an illustration of what a specific solution might entail?
FUKUDA: For example, the use of wearable sensors to help prevent suicide is possible solution that we can investigate. Research on this is already being conducted in Canada and the United States. The wearable sensor measures an individual’s heart rate, breathing, body temperature, and other vital signs, which are then analyzed to infer the person’s mental state. The sensor transmits an alert if the application determines a likelihood these factors may lead to suicide. Information science and other science-focused technicians have a tendency to consider only whether such solutions are tangibly feasible. Yet, even if they are, whether it would even be ethically permissible to use such a technique to peek into someone’s mental state is controversial. So, by validating such methods further from a social science perspective, I believe we can come up with even more effective ways of using such sensors. In other words, what today’s data scientists lack is not the technical perspective, but the social perspective to guide the practical application of ideas, about which people say “wouldn’t it be great if we could do this.”
I want students to learn that AI is hardly all-powerful
――As this course is mainly programming exercises, to what degree do you expect your students to achieve that goal?
FUKUDA: When I conducted a survey beforehand of students asking what they thought of AI, I found that many who had no experience with AI at all and imagined it to be something akin to magic. That is why I want students to actually experience AI and visualize its potential, feeling themselves: “I can’t even do this with AI.” There is a lot of media coverage about AI, but no news about what you can’t do with AI. Therefore, it is easy to make the mistake, but AI is by no means all-powerful. My hope is that students in this course create their own programs and actually run them so that they are able to achieve, in their own way, an understanding that “this logic is difficult” or “these lines are important.” It is also fine if they conclude that “AI is not interesting at all” or “it’s not for me.” Once they actually try it, they will have reaped benefit even if they only start to think in a way different from before the course when they had no experience with AI at all.
――I think that not a few students will actually being dealing with data once they find a job with a company or other organization. How exactly will what students learn in this course be useful to them?
FUKUDA: When they actually develop AI, they will probably need to rely on the skills that engineers possess. In such interactions, we can imagine engineers telling them “that can’t be done.” This class will help students to understand why something can’t be done. This course is not enough to provide students with detailed instructions about what sort of AI they actually want developed, so they will need to study more.
At Hitotsubashi University, many students want to gain a proper understanding right from the start
――Having taught courses before, what do you feel is the attitude of Hitotsubashi University students toward learning?
FUKUDA: This course is also conducted fully online, which makes it difficult to grasp students’ demeanor, so there is some concern there. I have also sensed disparities in students’ computer proficiency. Nevertheless, even from students who are not very computer literate, I have found they are passionate about learning this material. Overall, many Hitotsubashi University students are enthusiastic about learning. It might be the online material or something else, but it has been evident that quite a few students have also endeavored to study on their own. I have felt the students want to properly understand the lessons right from that start. In many cases, programming is skill that people with an interest in it acquire when they try it, desiring to create something, while also researching and investigating how to do that. Of course, it is also worthwhile learning how to properly program right from the start, but it is perfectly all right for students to take it easy and proceed to learn in their own way. In class, students have to write a lot of source code. Even so, we are not trying to train computer science experts or engineers, so the purpose is not hone their programming ability. Through programming, I would like the students to learn ways or approaches to use computers and information technology to solve problems.
――Specifically, what sort of AI will students be programming?
FUKUDA: There is a lot of data circulating around the world that we can make use of. For example, there is a type of AI for NBA games that predicts whether rookies with specific attributes will be able to play professionally five years from now based on data such as which rookies are playing and how they are playing, how many rebounds the players are getting, how many shots they are taking, and how many points scored. What I hope to foster through these sorts of practical exercises is the ability to intuitively sense causal relationships from data and conclusions deduced from that data.
Moreover, I also want to develop in students, who are walking around with smartphones, a recognition that they are walking around with a terminal full of their own personal information. When we use the internet, we are providing it with personal information without even realizing it. I think that we need to know how and where that data is being used, and I hope students will learn that in this course.
