The present age is called “the age of big data”.Human resources who have acquired the basics of data science, which creates new value from huge amounts of data, are now in high demand from all over the world.But not all new technologies make people happy.In order for the technology to be truly useful in solving social issues, it is essential to consider it from all perspectives, including business and ethics.Hitotsubashi University started a new faculty in April 2023 to train generalists in social data science with knowledge of social sciences such as business administration, economics, law, and political science in addition to data science.
How to implement in society while using new technology
Fusion of social science and data science is the key
“We tend to think that we can do anything with new technology, but we must always assess the impact that technology will have,” warns Professor Mamoru Komachi, who conducts research in the fields of information and AI.
“Recently, several universities have issued statements regarding ChatGPT, but this is due to various concerns. It is not possible to tackle these problems with engineering alone, and we need the knowledge of the social sciences.”
Especially in recent years, the speed of technological development is remarkable.Around 2010, when deep learning (a method used in ChatGPT and Google Translate) began to attract attention, it was said that it would take 10 to 20 years from research to social implementation.However, the development period for Google Translate, which appeared in 2016, is about two years, and for ChatGPT, about one year.
“Until now, technology has taken the lead, and discussions on laws and regulations and social acceptance have followed suit. We need people with knowledge of both science and science.”
New development in natural language processing of artificial intelligence
Quality evaluation as the latest research theme
The progress of AI is astounding, but what kind of research can be considered at the Faculty of Social Data Science?
“My field of expertise is computational linguistics and natural language processing, and deep learning has had a major impact on research in natural language processing. Deep learning, on the other hand, is able to fluently connect data, because humans tend to appreciate things that look smooth, even if they look unnatural or wrong when you look at them in detail. Therefore, we are conducting research on the theme of "how humans evaluate sentences."This is the most advanced initiative in the world.”
Another big advantage of deep learning is that it can handle a variety of languages and applications with a single system.Until now, a dedicated translation system was required for each corresponding language, such as Japanese and English, or Japanese and Chinese, whereas generative AI represented by ChatGPT handles multiple languages in one system.In addition to translation, you can also perform various tasks such as taking minutes, brainstorming, and programming.
“In a project in our research group, we are researching methods for analyzing grammar and semantics. Conventionally, analysis could not be performed without providing dictionaries and complicated rules for each language, but deep learning can We can perform simple and fast analysis with one system.”
After studying philosophy and history at university, Professor Komachi himself has the experience of switching to science in graduate school to pursue the possibilities of language analysis using computers.At first, he struggled with mathematics, but he says that his experience of studying both the humanities and sciences is still useful in his current research.
“The field of artificial intelligence is rapidly advancing. Even so, I would like to discover fundamental principles that will remain unchanged for decades, and to undertake pioneering research that will lead to major developments in the future.”
Contribute to social issues and business innovation by fusing social science and data science
Acquiring the ability to pose meaningful questions to society,
Acquire proficiency in data science skills through PBL
In the learning curriculum, the graduation requirement is to systematically study the three areas of "business," "social issues," and "data science."
Courses taken are broadly divided into "social science" and "data science."In the social sciences, there are many common subjects with other faculties, such as business administration, economics, law, and political science, so there are plenty of choices.At Hitotsubashi University, students are encouraged to take lectures from other faculties, creating an environment in which they can obtain highly specialized learning beyond their departments.In data science, you will systematically study introductory-level mathematics and information courses, as well as “statistics,” “information/AI,” and “programming.”An understanding of high school mathematics is essential, but the curriculum is structured so that students can acquire knowledge from the basics.
In addition, there are PBL (Project Based Learning) exercises in the third year.In the exercises, students ask themselves questions using data from the real world, analyze the data, find some kind of agreement, consider problem-solving methods, and think about how to implement them in society.First of all, it is important to ask questions that are meaningful to society, and the key to this lies in the social science lectures that students learn by the second year.Mastering data collection and analysis learned in data science lectures is also a major goal.Practitioners involved in data science at companies and government offices also participated in the exercises.Students' presentations will be given feedback on whether they are actually useful to society and whether they are feasible from the perspective of economics and ethics.
“If you compare it to cooking, data science is like kitchen knives and other tools, while social science knowledge and data are like ingredients. Knives can be dangerous if not used correctly, but delicious dishes can only be made with the right tools and ingredients. The Faculty of Data Science handles both tools and ingredients.I hope that you will learn while interacting with teachers and seniors, and open up new fields with your peers and juniors."
Hitotsubashi University Faculty of Social Data Science
Professor Mamoru Komachi