Training Deep Learning classifiers for stroke-based applications requires collecting lots of samples, which is often expensive and time-consuming. Data augmentation (DA) techniques can mitigate this issue by artificially increasing the number of training samples, eventually improving model performance and robustness. Since the effectiveness of DA techniques mostly depends on the task and dataset, researchers have proposed automatic DA methods, mostly for computer vision tasks. Unfortunately stroke-based data remain underexplored. To address this research gap, we propose AutoChainer, an automatic DA technique suitable for stroke-based data, that consists of applying random chains of augmentation transformations. We perform classification tasks on a variety of datasets (including gestures, letters and signatures) and models, showing that AutoChainer achieves state-of-the-art results. It also has the potential to enhance the visual quality of augmented samples, making them more interpretable, and offers easy customization to task-specific requirements, such as balancing classification accuracy and execution time.

Scientists play a pivotal role in advancing our understanding of the world, contributing not only through objective and rational work but also by incorporating their emotions into scientific discourse. The affective side of science is gaining increased attention. Building on this line of inquiry, this study explores the impact of emotional elements and sentiment analysis within abstracts published in the journal Nature over the past 10 years, adopting an affective-centric perspective on scientific endeavors. The research aims to explore the presence and trends of emotional elements and sentiments in these abstracts and to investigate the correlation between these emotional elements and citation counts, considering the influence of emotions on decision-making. The novelty of this research lies in not only considering sentiment analysis but also employing two distinct emotion detection techniques on the abstracts. The study employs a hybrid approach that integrates Lexicon-based, Deep Learning (DL), and Natural Language Toolkit (NLTK) sentiment analysis methods to explore diverse emotional dimensions through text emotion detection. The results demonstrate the presence of emotions and positive sentiment trends in the abstracts and positive sentiments in scientific abstracts were more prevalent and are associated with higher citation rates. It also highlights the observation that diverse emotions are prevalent in the most cited papers.