Search

The integration of multimodality in natural language processing (NLP) tasks seeks to exploit the complementary information contained in two or more modalities, such as text, audio and video. This paper investigates the integration of often under-researched audio features with text, using the task of argumentation mining (AM) as a case study. We take a previously reported dataset and present an audio-enhanced version (the Multimodal USElecDeb60To16 dataset). We report the performance of two text models based on BERT and GloVe embeddings, one audio model (based on CNN and Bi-LSTM) and multimodal combinations, on a dataset of 28,850 utterances. The results show that multimodal models do not outperform text-based models when using the full dataset. However, we show that audio features add value in fully supervised scenarios with limited data. We find that when data is scarce (e.g. with 10% of the original dataset) multimodal models yield improved performance, whereas text models based on BERT considerably decrease performance. Finally, we conduct a study with artificially generated voices and an ablation study to investigate the importance of different audio features in the audio models.

Try to use machine learning method to accurately classify different types of TEM images of atom lattice of mono layer Graphenes, see Figure1.

Demographic patterns in our world (e.g., aging processes, birth and death rates) are increasingly influenced by migration movements. A rising number of people is ‘on the move’, in search of a better fortune elsewhere. It is noteworthy that nowadays many migration movements do not show anymore stable patterns, but reflect a high degree of dynamics, for instance, in the form of return migration, circular and temporary migration, or chain migration. There is also a great heterogeneity in the motivations of many migrants that may have significant impacts on the migration choice, the destination place, the migrant’s status, and the duration of stay. Consequently, return migration, temporary migration and circular migration have in recent years become important research and policy issues. This note offers a short review of the dilemma’s and assessment issues inherent in the effects of non-structural or temporary migrants (so-called ‘movers’) on host economies. Particular attention will be paid to circular migration policy in Europe as a vehicle to both mitigate temporary tensions on regional labor markets of host economics and to provide a solid base for sustainable growth in the sending countries. Various research and policy challenges are outlined as well.

In the Netherlands the share of immigrants in the total population has steadily increased in recent decades. The present paper takes a look at wage differences between natives and migrants who are equally educated. This reduces potential skills biases in our analysis. We apply a Mincer equation in estimating the wage differences between natives and migrants. In our study we analyze only young graduates, so that conventional human capital factors cannot explain the differences in monthly gross wages. Therefore, we focused on "otherness" factors, such as parents' roots to find an alternative explanation. Our empirical results show that acquiring Dutch human capital, Dutch-specific skills, language proficiency, and integration in the long-term (second-generation with non-OECD background) are not sufficient to overcome wage differences in the Dutch labor market, especially for migrants with parents from non-OECD countries.

This paper presents an operationalization of a mixed Bourdieu–Mincer-type model that seeks to find evidence for individual and local cultural capital effects on human capital 'ability'. We aim to compare these effects for native workers and immigrants (as well as between immigrants themselves) in a locality. The main objective of the paper is twofold: 1) to examine how ethnic background affects immigrants' schooling results; and 2) to explore the link between the wage differential of immigrant young workers entering the labour market in the context of a locally varying cultural milieu. Our study utilises the 2007–2009 data set for higher professional education (termed HBO in Dutch) graduates from Maastricht University. We use a two-stage least squares (2SLS) estimation method to analyse empirically a system of two equations. In the first Bourdieu-type equation, individual cultural capital, together with school type/quality, explains the individual's schooling achievement. Next, this 'schooling achievement' is employed as an explanatory variable in the second Mincer-type equation, which examines wage differential effects. Our Mincer-type equation is next augmented with a control for the local cultural milieu. We find evidence of ethnic segregation with regard to the quality of educational institution to which immigrants have access, which naturally explains part of the wage differential effect. Moreover, we find that local cultural capital determines the size of the wage gap.

Since the early 1990s many empirical studies have been conducted on the impact of international migration on international trade, predominantly from the host country perspective. Because most studies have adopted broadly the same specification, namely a log-linear gravity model of export and import flows augmented with the logarithm of the stock of immigrants from specific source countries as an additional explanatory variable, the resulting elasticities are broadly comparable and yield a set of estimates that is well suited to meta-analysis. We therefore compile and analyze in this paper the distribution of immigration elasticities of imports and exports across 48 studies that yielded 300 observations. The results show that immigration complements rather than substitutes for trade flows between host and origin countries. Correcting for heterogeneity and publication bias, an increase in the number of immigrants by 10 percent may be expected to increase the volume of trade on average by about 1.5 percent. However, the impact is lower for trade in homogeneous goods. Over time, the growing stock of immigrants decreases the elasticities. The estimates are affected by the choice of some covariates, the nature of the data (cross-section or panel) and the estimation technique. Elasticities vary between countries in ways that cannot be fully explained by study characteristics; trade restrictions and immigration policies matter for the impact of immigration on trade. The migrant elasticity of imports is larger than that of exports in about half the countries considered, but the publication bias and heterogeneity-corrected elasticity is slightly larger for exports than for imports.

The rapid growth in the foreign-born population in many high and middle-income countries in the past decades has prompted much research on the socio-economic impacts of immigration. The migration issue has become one of the most debated subjects in many developed countries. Since the early 1990s, many applied studies have been conducted on the impact of international migration on international trade, foreign direct investment, and tourism. These studies have largely adopted the same specification, viz. the log-linear gravity model in combination with the knowledge capital model, where the (logarithm of the) stock of migrants from a specific source country may be included as an additional explanatory variable. Our study provides a concise review of the relationship between migrants and their international economic linkages. It then focuses on Foreign Direct Investment (FDI), for both inward and outward FDI. The main aim of our study is to offer a synthesis by means of a meta-analysis of various studies undertaken worldwide, in order to test the robustness of the relationship between migration and foreign direct investment. Our primary results confirm that immigration has a positive impact on FDI investment in both directions (inward and outward), and that these impacts are higher when migrants are highly-educated and skilled.

This PhD thesis tries to show the impact of transport infrastructure in economic development in least developed countries and in particular in the case of Afghanistan. Some least developed countries during 1990 to 1999 experienced lack of investment in transportation. Lack of investment further increased the economic development gap between developed and least developed countries. Moreover, lack of literature and research in poor countries such as Afghanistan encouraged me to do my research in this country in order to unveil the problems, facing poor people who are living in inaccessible places and suffer from lack of economic opportunities and long term unemployment. This thesis shows the effect of inaccessibility and immobility in economic opportunities and basic social services in Afghanistan. This thesis is important because it covers the role of transport infrastructures at the moment that international community promised to rebuild the infrastructures of post conflict Afghanistan.