Yassine Bechqito | Searcharxiv

arXiv2026

Designing Within the Lines: Practitioners' Perspectives and Visualisation Tool Evaluation in the Arabic Context

Design guidelines and best practices serve as references that support designers throughout the visualisation design process. While considerable effort has identified the elements that contribute to effective data visualisations, little attention has been paid to how language (scripts and reading direction), tool support, and cultural context also shape design decisions. As a result, assumptions of homogeneity persist, with visualisation practices predominantly benefiting users of English and left-to-right (LTR) scripts while overlooking the needs of over two billion Arabic script users. We investigate how Arabic-speaking visualisation practitioners design for right-to-left (RTL) scripts. We report on an analytical evaluation of seven popular GUI-based visualisation authoring tools using an Arabic dataset complemented by interviews with 11 Arabic-speaking practitioners across journalism, design, and data analysis. Our findings reveal that visualisation practitioners constantly negotiate tensions between Arabic reading conventions, "universal" LTR visual norms, and limited tool support for Arabic text, Eastern numerals, and maps. They engage in substantial labour, such as manually mirroring charts, fixing alignment issues, and stitching together multi-tool workflows, while making strategic compromises in language choice, interactivity, and chart type. Our tool analysis further reveals fragmented, inconsistent support for RTL mirroring, poor numeral rendering, and map defaults that encode geopolitical assumptions. In light of these findings, we discuss how RTL visualisation work is carried out under many constraints that affect agency and creativity. We argue that visualisation tools and defaults operationalise linguistic and geopolitical power in RTL contexts, and offer research directions and design implications that more robustly support RTL practitioners.

cs.HC↗

arXiv2025

Task-based Loss Functions in Computer Vision: A Comprehensive Review

Loss functions are at the heart of deep learning, shaping how models learn and perform across diverse tasks. They are used to quantify the difference between predicted outputs and ground truth labels, guiding the optimization process to minimize errors. Selecting the right loss function is critical, as it directly impacts model convergence, generalization, and overall performance across various applications, from computer vision to time series forecasting. This paper presents a comprehensive review of loss functions, covering fundamental metrics like Mean Squared Error and Cross-Entropy to advanced functions such as Adversarial and Diffusion losses. We explore their mathematical foundations, impact on model training, and strategic selection for various applications, including computer vision (Discriminative and generative), tabular data prediction, and time series forecasting. For each of these categories, we discuss the most used loss functions in the recent advancements of deep learning techniques. Also, this review explore the historical evolution, computational efficiency, and ongoing challenges in loss function design, underlining the need for more adaptive and robust solutions. Emphasis is placed on complex scenarios involving multi-modal data, class imbalances, and real-world constraints. Finally, we identify key future directions, advocating for loss functions that enhance interpretability, scalability, and generalization, leading to more effective and resilient deep learning models.

cs.LG↗