Forum Commemorating Launch of University of Tokyo Institute for Digital Observatory

The forum commenced with a welcoming address from the institute head, Masaru Kitsuregawa, a professor at The University of Tokyo who spoke about the university’s Data Integration and Analysis System for non-stop 24-hour live observation of the Earth. Professor Kitsuregawa made the following comments.

“This century has been termed the ‘age of observation,’ with observational capabilities that now dramatically outstrip what the 20th century was able to accomplish. Our ability to observe has been given a boost by live image acquisition. In the field of civil defense, for example, learning the many different behaviors of river systems enables detailed inundation predictions to be made. By utilizing multi-modal large language models (LLMs) and the careful observation of nature, conflicts, and other phenomena, the purpose of the institute is to create technologies for building resilient societies that do not put constraints on how people live their lives.”

In collaboration with various government agencies, the institute has set out to identify global trends through the observation of information in numerous forms, including an analysis of supply chain stresses that utilizes the visualization of an international input-output chart encompassing 190 countries and an analysis of import and export dependencies derived from trade data sourced from organizations such as the United Nations (UN).

The next speaker was Hitachi Senior Vice President and Executive Officer Jun Abe who spoke about the reasons why the two organizations were collaborating on the digital observatory, touching on Hitachi’s progress in the data storage field and its joint research with The University of Tokyo that has been ongoing for more than 20 years in data storage as well as work on resolving societal challenges.

“Given the changing circumstances in which social and economic activity is taking place, as epitomized by the COVID-19 pandemic, the question of how this will impact our lives has become a new issue for society to address. By leveraging Hitachi’s strength in data processing and the experience that we have acquired through the delivery of a wide range of services and solutions, we intend to utilize the digital observatory to create new forms of value to support social and economic activities that confront relentless change.”

Yasutoshi Nishimura, Minister of Economy, Trade and Industry, offered his congratulations on the launch of the institute with the following comments.

“What really brought home the importance of data to me was when I served as the Minister in charge of Measures for Novel Coronavirus Disease in 2020 when the COVID-19 pandemic was just starting. The use of posts on social networking services (SNSs) to analyze people’s behavior played a part in subsequent policy formulation and restrictions on activity. This reinforced for me the vital role of data in resolving societal challenges. Together with difficulties caused by the interruption of supply chains, the pandemic brought the concept of economic security into international focus. If we are to establish robust supply chains, then Japan too needs to put appropriate laws in place and build a self-sufficient economic structure. As we seek to further enhance economic security, I will be paying close attention to the work of the Institute for Digital Observatory.”

Toyoda:Having identified the building of a resilient society as a goal on the path to 2050, the Institute for Digital Observatory monitors and studies digital data on the many factors that impact supply chains. From your respective perspectives representing a university and a corporation, could you each talk about the society and economic activity you envision for 2050?

Fujii:If we are to obtain a broad view of the correlations between different phenomena, I believe it is essential that we work with a large number of people. This involves using the information contained in SNS posts on what the public are thinking to keep track of current issues and come up with ideas on what sort of arrangements we should be putting in place. I expect this will be just as important when thinking about the society and economy of 2050. Behind the success of the large IT platform businesses that have taken the world by storm in recent times, I believe there is a way of doing business that maintains close contact with users. I expect the same will be expected of public services in the future.

Another line of thinking in academia that has arisen recently is of “ethical, legal and social issues” (ELSI). When developing and implementing new technologies, this calls upon us to question those technologies’ ethical and social impacts. In practice, however, it is very difficult to predict how social and economic systems will be affected by a particular technology. This is because the individual problems are addressed in isolation from one another. Of particular significance in the lead-up to 2050, I believe, will be to interconnect across different fields, by which I mean having effective ways of anticipating what might happen and by what process of cause and effect.

Kojima:When it comes to creating an economy capable of sustainable growth by 2050, two major threats have manifested themselves already. The first is that economic activity is starting to exceed planetary boundaries. The second is that our well-being as human beings is in doubt. The question is being asked, “Are people really happy?” Achieving economic growth without well-being defeats the purpose, I would say. While the question of what to do to overcome these two threats is a difficult one, I believe that targets for the intermediate milestone of 2030 are important. We need to change the directions we are going in many different areas and establish a clear trajectory for turning things around.

That all of our attention is currently occupied by immediate challenges amid a deluge of problems that include pandemics and conflict is a result, one might say, not so much of newly manifesting risks, but of having failed to pay attention to problems that have been around for a long time. Meanwhile, those risks are now coming to pass also.

A key question to address if we are to create an ideal society in 2050 is how to bring to light those things that are happening right now but that we have failed to recognize, and how to start addressing them. In this regard, I would like to put this concept of a digital observatory to work in our business.

Issues with Data Utilization and Expectations for New Technology

Toyoda:The Institute for Digital Observatory has set out to address a wide variety of issues by monitoring data and putting it to use. In relation to its activities, are there any of these issues you would like to raise in particular? Also, please tell us about your expectations for new technologies such as generative artificial intelligence (AI).

Kojima:I see generative AI as really a major breakthrough that will impact all sorts of knowledge work done by humans. For ourselves, generative AI raises two questions that will become the basis of competition. The first is how to collect the text and insights that are the inputs of generative AI and large language models (LLMs). The second is in which applications will LLMs deliver genuine value. Given this context, how to obtain the data is of extreme importance. In a world that is awash with data, the process of collecting and analyzing the many different forms of this data and using it in conjunction with meaningful insights as an input to generative AI will be crucial. This is where I believe the digital observatory can deliver the greatest benefits.

Fujii:A key point, I believe, is how to incorporate important data that is difficult to acquire in digital form, what you might call the “hard-to-hear voices.” One example is the challenge of acquiring information from the most upstream reaches of supply chains in emerging nations and linking it together. The University of Tokyo is strengthening student exchanges between us and universities in Africa and Southeast Asia based around social entrepreneurship, because it is important to actually go to these places to see what is happening on the ground. We need to realize that, through its cumulative effects, data in itself is hugely powerful.

Moreover, there is scope for adopting a quantum way of thinking whereby the collection of large amounts of data from a wide range of fields enables many different time series to be observed simultaneously. In doing so, the correlations between them can be used to find out what is going on. I expect there is scope for crossover with the field of quantum computing.

Conclusions

Toyoda:Finally, could you please say a few words on the future of the digital observatory?

Kojima:In this era when digital technology is flourishing, it is important to distinguish between those areas that are subject to competition and those that are not. At Hitachi, we see the monitoring and analysis of digital data as belonging to the non-competitive realm. In areas like this, the public sector, academia, or companies should collaborate on creating the digital observatory. My hope is that we can get more people involved and expand our scope to make the digital observatory a reality under the leadership of its director, Masaru Kitsuregawa.

Fujii:At The University of Tokyo, we are guided in our daily activities by our philosophy of wanting to be an institution that brings together stakeholders from throughout the world. In the sense of expanding the scope of data, some of which will be open and some closed, we intend to do our best to forge these links as widely as possible across the world’s many different places, organizations, and sectors. My hope is that we can combine the expertise that The University of Tokyo has built up in sociology and many other academic fields to listen to the “hard-to-hear voices” and transform this data into a strength for Japan. Thank you for your time today.

The discussion was followed by a message from Hideya Yamada, who heads the Statistics Department of the Minister's Secretariat at the Ministry of Agriculture, Forestry and Fisheries. Mr. Yamada touched on several issues, including Japan’s low level of food self-sufficiency and how the aging population and declining birthrate are shrinking the workforce in agriculture, forestry, and fisheries. He also discussed the ministry aims to revise the Basic Law on Food, Agriculture and Rural Areas with a focus on strengthening food security, and had the following to say about the hope of close working with the Institute for Digital Observatory and the timeliness of its launch.

“We have been visiting the Institute of Industrial Science at the Komaba Campus to exchange views with the University of Tokyo academics since the Institute for Digital Observatory’s inception. The work they have done for us analyzing trade in products such as vegetables or fertilizer has given us an appreciation for the importance of this data platform. We look forward to even higher levels of collaboration between industry, government, and academia”

Next on the program were presentations on the vision of the Institute for Digital Observatory given by Deputy Director Masashi Toyoda and Hitachi Vice President, Executive Officer, and CTO Itaru Nishizawa.

Deputy Director Toyoda spoke about the digital observatory’s research goal of building a resilient society through the early identification and avoidance of serious risks, stating that the three key pillars of this work were the development of core technologies for monitoring social and economic activity, the establishment of platforms for putting the observed data to use, and the use of data as a basis for creating value (supply chain resilience, diversity and inclusion, and being nature positive and opportunity seeking). He described how research at the institute combined both the arts and sciences and was established on the basis of collaboration not only among the eight groups at The University of Tokyo, but also with external organizations such as Hitachi, the University of Tsukuba, and the National Institute of Informatics. He also mentioned the ongoing discussion of future plans being held with relevant government agencies (the Ministry of Economy, Trade and Industry and the Ministry of Agriculture, Forestry and Fisheries among them) with a view to putting the work done by the institute to use across a wide range of fields, including the analysis of data on social activity, multimedia analysis, and global social risk analysis. The presentation also covered the observation platform for supply chain resilience, one of the major research topics at the institute (see Figure 1 and Figure 2).

Noting that the goal of the Institute for Digital Observatory is to observe and utilize diverse forms of data, Sadao Kurohashi, Director General of the National Institute of Informatics and leader of the institute’s natural language processing group, made the point that, by translating numeric data into a language that humans can understand, the trends and other meaningful content embedded in it on things like transaction volumes and production quantities can be turned into important sources of information. Expanding on this topic, he talked about the dramatic advances in language understanding by LLMs brought about by generative AI, observing that, were the potential for using Generative Pre-trained Transformer 4 (GPT-4)^*2 to extract and structure information to be realized, it would become a major strength. Accompanying this, he also mentioned current concerns about AI being a black box and how some of the research institutions and companies engaged in the development of generative AI have established an oligopolistic position. The National Institute of Informatics has been building its own language models to ensure the reliability and transparency of generative AI and is working on elucidating the principals involved.

*2

An artificial intelligence model for natural language processing developed by OpenAI. GPT-4 was designed to be capable of more natural interaction and text generation than GPT-3 and GPT-3.5.

Based on their mission of analyzing the risks for global society, Professor Takuto Sakamoto of the Graduate School of Arts and Sciences at The University of Tokyo has been working with his team made up largely of graduate students to understand and predict sociopolitical events with the potential to threaten global supply chains. He recounted specific examples of conflict event data and explained the data generation process where the goal is to collect and map their own data together with existing open data on events such as conflicts, human rights violations, and political instability. They then draw on the expertise of different team members to produce quantitative predictions of global risks that are complemented by contextual information for interpreting what they mean. Given the effort and expense required to produce this data, he explained how he wanted to automate the process of generating it to make it more efficient and timelier, as well as making up for the gaps where current data is inadequate.

Professor George Shishido of the Graduate Schools for Law and Politics at The University of Tokyo spoke about how the medium- to long-term research goals of the Institute for Digital Observatory would contribute to enhancing the performance of the legal system, improving the efficiency of corporate activities, and advancing the sustainability of the public environment by presenting new insights and perspectives based on data analysis results that have been collated and made available through the establishment a comprehensive data analysis platform for legal and political studies. He also mentioned the institute’s more immediate goals of identifying, evaluating, and collating the risks evident in international rules on free trade. In this regard, he gave an overview of current projects that include research on trade policy and on the impacts that factors such as the global situation are having on trade, explaining how, by using data to highlight the different risks, they were seeking to offer meaningful perspectives on what companies are doing about risk reduction and the formulation of relevant policies by government.

The next presentation was by Zhao Han, a civil engineering major at The University of Tokyo School of Engineering who is engaged in research under the supervision of Professor Akiyuki Kawasaki of the university’s Institute for Future Initiatives. Talking about the connections between worsening climate change and economic activity, she made the point that, based on the distribution of national CO₂ emissions, the lower a regions’ greenhouse gas emissions, the more vulnerable it tends to be to climate change, and that importing and consuming regions (primarily developed nations) are passing off their environmental load onto producing and exporting regions (primarily emerging nations). On this basis, she argued that accurately assessing the relationships between economic activity and climate change would be essential to building fairer and more sustainable supply chains. She presented the results of analysis using the Environmentally Extended Multiregional Input-output (EE-MRIO) tables that was aimed at bringing to light the impacts that each nation’s trading activities are having on the global environment. She also explained how, by pursuing further analysis and visualization efforts that combine datasets relevant to climate change risk, she hopes to identify regions where the risks are high, share information with stakeholders, prioritize policy making, and supply predictions to minimize losses.

In the next session, Associate Professor Shinichiro Kumagaya of the Research Center for Advanced Science and Technology at The University of Tokyo, who himself has cerebral palsy, spoke on the theme of diversity and inclusion with a particular focus on disabilities. He reported on the current situation and challenges of inclusion that exist with reference to the percentage of disabled students at higher education institutions and how their numbers are trending. Pointing out that inclusion means creating an environment in which those people around them make allowances for the ways in which people with diversity struggle, where they can also express the capabilities in which they excel, Associate Professor Kumagaya went on to make the point that measuring progress on inclusion posed difficult challenges. Touching on changes in the environment for disabled people and the correlation between diversity and performance found in the results of past surveys on inclusion, he also expressed his desire to engage in the digital observatory processes as part of activities that bring greater transparency to the state of Diversity, Equity and Inclusion (DEI) at national universities and use people’s concerns as a force for inclusion.

Associate Professor Rie Yamaguchi, who studies information security and privacy protection at the Graduate School of Information Science and Technology at The University of Tokyo, spoke about her desire to use techniques for anomaly detection and pattern analysis based on people’s lifestyles in the research work of the digital observatory. She showed how there are links between various global events and divergences in time-series data. As an example of analyzing such data, she showed how many of the events going on at the time showed up in data, such as the movements of vessels arriving and departing the Ukrainian port of Odessa and the trend In Ukraine’s net external assets. Noting that this characteristic was a close fit with the goals of the Institute for Digital Observatory, she talked about how she wanted to match anomalies in time-series data with events in the news, to eliminate noise in the data with the help of advice from other team members, and to uncover correlations between global events.

The closing address was given by Nobuhito Saito, Executive Vice President of The University of Tokyo, who recounted the background to the establishment of the Institute for Digital Observatory and the forum’s program, saying that the institute’s mission was not just to observe, but also to collect and analyze data and use it to identify what actions to take next. Expressing his desire to create an organization that can respond quickly and dynamically to an unpredictable global situation, he had the following to say.

“Social and economic activities such as manufacturing, production, distribution, and services are subject to a wide variety of risks relating to the environment, society, and governance. Moreover, these risks have become an international issue over recent years, having been further exacerbated by COVID-19 and geopolitical risk along with human rights issues and the destruction of natural capital in emerging nations. With supply chains having been disrupted by a range of developments, including the Russia-Ukraine conflict, Japan as a whole should benefit if we can shed light on the complex ecosystems of industry by leveraging digital capabilities for broad-based monitoring that addresses questions such as where to source parts and other inputs, what to produce, and where to sell it.”

Building on this, he talked about initiatives for practical industrial applications whereby industry and academia work in partnership to explore this new academic field, monitoring huge and diverse volumes of data that extend from statistics published by national and local government, companies, and other organizations to numerical data collected from sensors in the real world and the large amount of text that flows daily across SNSs.