Published papers

Understanding expressions that refer to the physical world is crucial for such human-assisting systems in the real world, as robots that must perform actions that are expected by users. In real-world reference resolution, a system must ground the verbal information that appears in user interactions to the visual information observed in egocentric views. To this end, we propose a multimodal reference resolution task and construct a Japanese Conversation dataset for Real-world Reference Resolution (J-CRe3). Our dataset contains egocentric video and dialogue audio of real-world conversations between two people acting as a master and an assistant robot at home. The dataset is annotated with crossmodal tags between phrases in the utterances and the object bounding boxes in the video frames. These tags include indirect reference relations, such as predicate-argument structures and bridging references as well as direct reference relations. We also constructed an experimental model and clarified the challenges in multimodal reference resolution tasks.

Linear regression is a basic and widely-used methodology in data analysis. It is known that some quantum algorithms efficiently perform least squares linear regression of an exponentially large data set. However, if we obtain values of the regression coefficients as classical data, the complexity of the existing quantum algorithms can be larger than the classical method. This is because it depends strongly on the tolerance error $ε$: the best one among the existing proposals is $O(ε^{-2})$. In this paper, we propose the new quantum algorithm for linear regression, which has the complexity of $O(ε^{-1})$ and keeps the logarithmic dependence on the number of data points $N_D$. In this method, we overcome bottleneck parts in the calculation, which take the form of the sum over data points and therefore have the complexity proportional to $N_D$, using quantum amplitude estimation, and other parts classically. Additionally, we generalize our method to some class of convex optimization problems.

This paper introduces LLM-jp, a cross-organizational project for the research and development of Japanese large language models (LLMs). LLM-jp aims to develop open-source and strong Japanese LLMs, and as of this writing, more than 1,500 participants from academia and industry are working together for this purpose. This paper presents the background of the establishment of LLM-jp, summaries of its activities, and technical reports on the LLMs developed by LLM-jp. For the latest activities, visit https://llm-jp.nii.ac.jp/en/.

We studied the band structure of NdFeAs$_{1-x}$P$_{x}$O$_{0.9}$F$_{0.1}$ ($x$ = 0, 0.2, 0.4 and 0.6) using angle-resolved photoemission spectroscopy (ARPES) measurements. Two of the hole bands, $α_1$ $(d_{xz})$ and $α_3$ $(d_{z^2})$, were observed at the Brillouin zone center in the $P$-polarized light configuration, while the other two hole bands, $α_2$ $(d_{yz})$ and $γ$ $(d_{xy})$, were observed in the $S$-polarized alternative. The observed $γ$ band shifts downwards as $x$ increases, which is consistent with the theoretical prediction for the change in bond angle of As/P-Fe-As/P. Furthermore, a small amount of the $d_{xy}$ orbital component was observed at the same binding energy as that of the top of the $α_1$ band, thus indicating the band reconstruction of the originally degenerate $α_1$ and $α_2$ $(d_{xz}/d_{yz})$ bands by the unoccupied $d_{xy}$ band. The change in the energy level of the $α_1$ band top with $d_{xy}$ orbital character is accompanied by a $T_{c}$ upturn at $0.2 < x < 0.4$. The $T_{c}$ continues to increase as the $α_1$ band shifts downward, crossing the Fermi level. The incipient band with the $d_{xy}$ orbital character on its top could be an important ingredient for high $T_{c}$ 1111-type iron-based superconductors.

Polycyclic codes are a generalization of cyclic and constacyclic codes. Even though they have been known since 1972 and received some attention more recently, there have not been many studies on polycyclic codes. This paper presents an in-depth investigation of polycyclic codes associated with trinomials. Our results include a number of facts about trinomials, some properties of polycyclic codes, and many new quantum codes derived from polycyclic codes. We also state several conjectures about polynomials and polycyclic codes. Hence, we show useful features of polycyclic codes and present some open problems related to them.

Scintillator-based X-ray imaging detectors are pivotal in numerous scientific and practical domains, including medical imaging, and security monitoring. Recent advancements have spurred interest in 4D X-ray imaging using synchrotron radiation, necessitating higher temporal resolutions. Consequently, this places stringent demands on X-ray detector technology, especially when X-ray energy exceeds 20 keV. The selection of a suitable scintillator material is crucial for achieving optimal timing resolution, yet it poses a significant challenge in dynamic X-ray imaging. This study delves into the optimization of scintillator properties and their impact on spatial resolution and light output, elucidating the performance of Ce-doped Gd3Ga3Al2O12 (GAGG:Ce) scintillators for X-ray imaging applications. We developed a micro X-ray imaging detector using a 100 $μ$m-thick GAGG:Ce scintillator plate and conducted X-ray imaging tests at the Aichi SR facility. The results demonstrated that the resolution, quantified as the chart slit width at a contrast transfer function (CTF) value of 10%, reached 2 - 3 $μ$m with a 4x lens, 0.52 $μ$m $\pm$ 0.03 $μ$m with a 20x lens, and 0.42 $μ$m $\pm$ 0.01 $μ$m with a 40x lens. Although the results of this study did not achieve a spatial resolution nearing the effective pixel size of the 40x lens, the text also elucidates the underlying reasons for this limitation. Furthermore, we compared the X-ray sensitivity of our GAGG:Ce scintillator plate with that of a commercial LuAG:Ce scintillator, revealing an approximately 1.5-fold increase in light output. As a demonstration, transmission images of dried small fish were captured using the GAGG:Ce scintillator plate and the developed X-ray imaging system. These findings highlight the potential of the X-ray imaging detector devised in this study for future generations of X-ray imaging applications.

Multimodal reference resolution, including phrase grounding, aims to understand the semantic relations between mentions and real-world objects. Phrase grounding between images and their captions is a well-established task. In contrast, for real-world applications, it is essential to integrate textual and multimodal reference resolution to unravel the reference relations within dialogue, especially in handling ambiguities caused by pronouns and ellipses. This paper presents a framework that unifies textual and multimodal reference resolution by mapping mention embeddings to object embeddings and selecting mentions or objects based on their similarity. Our experiments show that learning textual reference resolution, such as coreference resolution and predicate-argument structure analysis, positively affects performance in multimodal reference resolution. In particular, our model with coreference resolution performs better in pronoun phrase grounding than representative models for this task, MDETR and GLIP. Our qualitative analysis demonstrates that incorporating textual reference relations strengthens the confidence scores between mentions, including pronouns and predicates, and objects, which can reduce the ambiguities that arise in visually grounded dialogues.

The present study proposes a novel full-field extensional rheo-optical technique to investigate the relationship between the rheological properties and internal structural deformation of complex fluids under uniaxial extensional flow. Macroscopic viscoelasticity from rheological measurements and microscopic birefringence from optical measurements are integrated to evaluate the microstructural deformation and orientation of the fluids under extensional stress. The proposed technique integrates a liquid dripping method with a high-speed polarization camera to measure the extensional stress and flow birefringence simultaneously. In the liquid dripping method, temporal evolution images of the liquid filament diameter for fluids dripping from a nozzle are measured to obtain the extensional stress loading on the filament. These images are acquired using the high-speed polarization camera that captures full-field two-dimensional (2D) birefringence with high spatiotemporal resolution. Wormlike and networked micellar solutions of cetyltrimethylammonium bromide (CTAB) and sodium salicylate (NaSal) with varying concentrations of CTAB and NaSal are employed as the measurement targets. Consequently, we successfully visualized temporally developing images of the flow birefringence field of uniaxially extending micellar solutions induced by the orientation of micelles. Furthermore, the proposed technique supports investigating the conditions for establishing the stress-optical rule, which is the linear relationship between stress and birefringence. The stress-optical coefficient, which is a proportionality constant indicating the sensitivity of birefringence to stress, is analyzed from these measurements. The stress-optical coefficient under uniaxial extensional flow, obtained using the proposed technique, is confirmed to be comparable to that under shear flow.

Quantum error correcting codes (QECC) is becoming an increasingly important branch of coding theory. For classical block codes, a \href{codetables.de} {comprehensive database of best known codes} exists which is available online at \cite{codetables}. The same database contains data on best known quantum codes as well, but only for the binary field. There has been an increased interest in quantum codes over larger fields with many papers reporting such codes in the literature. However, to the best of our knowledge, there is no database of best known quantum codes for most fields. We established a new database of QECC that includes codes over $\mathbb{F}_{q^2}$ for $q\leq 29$. We also present several methods of constructing quantum codes from classical codes based on the CSS construction. We have found dozens of new quantum codes that improve the previously known parameters and also hundreds new quantum codes that did not exist in the literature.

A sustainable financing strategy for SMEs should aim to enhance a low-cost collateral-free supply of loans to SMEs with good track records of repayments to banks. In this paper, we suggest two alternative financing models for SMEs that address certain borrowing constraints of SMEs. First, the model incorporates institutional mechanisms involving the government, banks, and SMEs. The strategy employs a two-pronged approach: (i) the government enhances the supply of loanable funds to banks, and (ii) identifies good SME borrowers through skills development programs and introduces them to banks. This model will reduce default risk and allow banks to offer lower-interest and collateral-free credit to SMEs, thereby improving their access to finance and performance. Second, the model could be extended to accommodate digital finance using a data-driven credit risk score of the borrowers to reduce banks’ default risks and transaction costs with or without government funds. The proposed model could resolve the moral hazard and selection bias problems. Our proposed models are based on a public-private partnership approach and therefore could solve certain borrowing constraints of SMEs. Our empirical results support the model outcomes and therefore are consistent with the predictions of our theoretical models.

This study examines the nexus between money demand and cryptocurrencies by estimating two simultaneous equations using Divisia indices as a proxy for money demand and volume traded for cryptocurrencies. The study examines the linkage between cryptocurrencies and money demand and their potential influence over monetary policy actions. It finds that the volume of cryptocurrencies traded negatively influences money demand. Moreover, we see a positive association between money demand and cryptocurrencies, implying that as the demand for money increases, the demand for cryptocurrencies increases. Further, we examine the determinates of cryptocurrencies and report that the return of cryptocurrencies, the financial development index, GDP, inflation, and stock market indices are significant predictors of the demand for cryptocurrencies.

This study examines the nexus between money demand and cryptocurrencies by estimating two simultaneous equations using Divisia indices as a proxy for money demand and volume traded for cryptocurrencies. The study examines the linkage between cryptocurrencies and money demand and their potential influence over monetary policy actions. It finds that the volume of cryptocurrencies traded negatively influences money demand. Moreover, we see a positive association between money demand and cryptocurrencies, implying that as the demand for money increases, the demand for cryptocurrencies increases. Further, we examine the determinates of cryptocurrencies and report that the return of cryptocurrencies, the financial development index, GDP, inflation, and stock market indices are significant predictors of the demand for cryptocurrencies.

The goal of our study is to examine the impact of natural disasters on the South Pacific Stock Exchange. We use daily time-series data for Fiji's stock market for the period 2000-2019. Our empirical framework is based on three factor regression models, namely the market model, the Fama and French three-factor model, and the Fama and French five-factor model. We find evidence that natural disasters in Fiji reduce abnormal returns in the most relevant five-factor model. Additionally, we provide evidence that different types of natural disasters have heterogeneous effects on Fiji's stock market. Our findings are further supported by a robustness check.

© 2020-Center for Economic Integration, Sejong Institution, Sejong University, All Rights Reserved. Regional economic integration is one of the most effective ways of fostering stability, addressing regional challenges, and increasing rates of economic growth. Strengthening regional cooperation and integration in different Asian sub-regions offers great potential for eliminating poverty and achieving inclusive and sustainable development. Hurdles remain, however, for further integration. The willingness to commit to “unity in diversity” is important to overcoming this, requiring the acceptance of cultural, linguistic, social, religious, and political differences. Such complex unities have been achieved in the European Union (EU) and elsewhere (Fella, 2002). Currently, in most Asian sub-regions, such political ambitions for economic integration have been scarce. This implies that a more cautious approach is needed in Asia than in other places. However, by studying the European context, we can derive valuable lessons for better fostering economic integration. Regional economic integration should first ensure the maximization of benefits for all participants. With this, complete cooperation is required to develop an effective and uniform regional business environment. Second, a well-developed infrastructure with an absence of trade barriers is needed. Various Asian sub-regions (e.g., Central and South Asia), however, suffer from a lack of infrastructure, which hinders trade opportunities. Difficulties in accessing capital also presents a key obstacle. In this regard, the role of infrastructure financing is very important. In the EU, their European Investment Bank maintains equitable financing for intra-EU projects. Open engagements with partner countries also play a key role. Third, over the past 30 years, different institutional mechanisms have sprung into existence in Europe that have created the basis for strategic economic integration. Such institutional mechanisms and initiatives are needed in Asia.

Increased consumption of nonrenewable energy sources may lead to more air pollution, resulting in negative health impacts in a society. The main purpose of this study is to investigate the relationship between fossil fuel energy consumption and health issues using generalized method of moments estimation technique for data from 18 Asian countries (both low-and middle-income) over the period 1991-2018. The findings demonstrate that fossil fuel energy consumption increases the risk of lung and respiratory diseases. In addition, the results demonstrate the significant effect of CO 2 emissions and fossil fuel consumption on undernourishment and death rates. Furthermore, we find that increases in the gross domestic product per capita and healthcare expenditure may help reduce undernourishment and death ratio. The conclusion recommends that diversification of energy in low-and middle-income countries from too much reliance on fossil fuels to more renewable energy sources can improve energy insecurity, at the same time reduce greenhouse gas emissions and minimize the negative impacts on human health.

International remittances by immigrants to their home countries is key in the funding development in migrant-sending countries. This study investigates the combined impact of international remittance inflows and financial inclusion on economic growth using a sample of 60 low- and middle-income countries over 1996–2017. The study constructs a composite financial inclusion index using principal component analysis. The results show that financial inclusion could strengthen the growth-enhancing effect of remittances. Thus, the study helps explain the development dilemma of remittance inflows and financial inclusion in migrant-sending countries to boost their economic growth.

Establishment of green credit guarantee schemes (GCGSs) and returning a portion of the tax revenue originally generated from spillover effect of green energy supply to investors. It can reduce the risk of green finance and increase the rate of return of green energy projects, respectively. In addition, technical developments in the sphere of distributed ledger technologies provide the opportunity to increase the transparency in green finance and investments. This paper contributes to literature by proposing two applied frameworks, backed by theoretical models on green finance and investment based on projects size. The objective is to induce the private participation in green finance and investment.

The main objective of the Bank of Japan's monetary policy is inflation targeting; however, it may be associated with the recently increased income disparity in Japan. Using the autoregressive distributed lag (ARDL) estimation model, this study conducted in the context of an aging society revealed that the bank's monetary policy increased income inequality through increased financial asset prices, resulting in a widening of the income gap. We also find that in the long run, the aging population will increase income disparity in Japan; however, tax policy will partially mitigate the adverse effect on income inequality. Unconventional monetary policy will not bring a desirable result for the Japanese economy and the nation as a whole. In addition, the government needs to revise the incremental tax system so that wealthier groups pay more tax than before. In addition, expanding the tax system by adopting the carbon taxation scheme would be beneficial. The government with this kind of tax can pursue two major goals: reducing poverty and environmental pollution, which are ongoing challenges in Japan.

© 2017, Duncker und Humblot GmbH. All rights reserved. In April 2013, the Bank of Japan (BOJ), in order to overcome deflation and achieve sustainable economic growth, introduced an inflation target of 2 %. However due to the lower oil prices in the global market, this target could not be achieved for a long-term period. In Feb 2016, the BOJ took further steps, and started a negative interest rate policy, by increasing massive money supply through purchasing long-term Japanese government bonds (JGB). Previously the BOJ only purchased short-term government bonds. This policy has flattened the yield curve of JGBs. Banks started to reduce purchasing government bonds, because the yield of short-term government bonds’ became negative, and even for long-term government bonds up to 15 years the interest rates became negative. On the other hand, bank loans to the corporate sector did not grow, due to the vertical investment-saving (IS) curve in the Japanese economy. This paper discusses that first, in the low oil price era, the BOJ has to modify the inflation target of 2 % and reduce it, secondly it argues that the role of monetary policy and the negative interest rate policy in Japan cannot recover from the current recession and deflation situation for the long term. What is important is to make the IS curve downward sloping rather than vertical. That means the rate of return on investment must be positive and companies must be willing to invest if interest rates were set too low. The recession in Japan is coming from structural problems that cannot be solved by the current monetary policy. The last section reports a simulation results of tackling aging population of Japan by introducing productivity based wage rate and postpone retirement age will help the recovery of Japanese economy.

This paper asks why Japan has not yet suffered from a sovereign debt crisis, although its gross public debt as a percentage of GDP is much higher than in Greece. We use a simple stylized model to explain the occurrence of both a fundamental and a speculative debt crisis. We apply this model to both countries and derive some hypotheses about why investors are still ready to hold Japanese Government Bonds. In particular, we point to the significance of domestic debt holdings, to the central bank’s government debt purchases, to investors’ access to “safe havens,” and to the role of an autonomous monetary policy. We also analyze potential challenges to Japan’s long-term fiscal situation, resulting from its aging population.