Josep Lladós

Several studies show that financial vulnerability is not determined solely by low levels of wealth, but also by behavioural and social factors that shape financial behaviour. From this perspective, the social environment and financial knowledge can influence how investors evaluate their investment experiences. However, most of the literature has focused on how these aspects affect participation in financial markets, rather than on how they shape perceptions of the investment experience itself. This study explores how interactions with one’s social environment and both objective and subjective levels of financial knowledge contribute to how people evaluate the outcomes of their investments. To do so, we analyse a sample of undergraduate students using multivariate regression and Oaxaca–Blinder decompositions across three social environments—family, workplace, and banking advisors—and three types of financial assets: stocks, investment funds, and pension funds. The results show that perceptions of investment experience are shaped not only by individual factors but also by financial knowledge and the social environment—and these effects differ between men and women. There are also differences across types of financial assets, suggesting varying levels of vulnerability. These findings highlight the importance of personal characteristics, financial knowledge, and social context in explaining investment perceptions and differences in financial vulnerability.

Achieving robustness in recognition systems across diverse domains is crucial for their practical utility. While ample data availability is usually assumed, low-resource languages, such as ancient manuscripts and non-western languages, tend to be kept out of the equations of massive pretraining and foundational techniques due to an under representation. In this work, we aim for building models which can generalize to new distributions of data, such as alphabets, faster than centralized fine-tune strategies. Despite the recent advancements in model editing to enhance the incorporation of low-resource data, the test for its applicability in a real environment remains unmet. In contrast to state-of-the-art meta-learning, we showcase the challenges of domain merging in sparse distributions of data, with agnosticity of its relation to the overall distribution or any other prototyping necessity. Even when using the same exact training data, our experiments showcase significant performance boosts in transfer learning to new alphabets and out-of-domain evaluation in challenging domain shifts, including historical ciphered texts and non-Latin scripts. This research contributes a novel approach into building models that can easily adopt under-represented alphabets and, therefore, enable document recognition to a wider set of contexts and cultures.