arXiv:2510.24262v1 Announce Type: new 
Abstract: Data augmentation using generative models has emerged as a powerful paradigm for enhancing performance in computer vision tasks. However, most existing augmentation approaches primarily focus on optimizing intrinsic data attributes -- such as fidelity and diversity -- to generate visually high-quality synthetic data, while often neglecting task-specific requirements. Yet, it is essential for data generators to account for the needs of downstream tasks, as training data requirements can vary significantly across different tasks and network architectures. To address these limitations, we propose UtilGen, a novel utility-centric data augmentation framework that adaptively optimizes the data generation process to produce task-specific, high-utility training data via downstream task feedback. Specifically, we first introduce a weight allocation network to evaluate the task-specific utility of each synthetic sample. Guided by these evaluations, UtilGen iteratively refines the data generation process using a dual-level optimization strategy to maximize the synthetic data utility: (1) model-level optimization tailors the generative model to the downstream task, and (2) instance-level optimization adjusts generation policies -- such as prompt embeddings and initial noise -- at each generation round. Extensive experiments on eight benchmark datasets of varying complexity and granularity demonstrate that UtilGen consistently achieves superior performance, with an average accuracy improvement of 3.87% over previous SOTA. Further analysis of data influence and distribution reveals that UtilGen produces more impactful and task-relevant synthetic data, validating the effectiveness of the paradigm shift from visual characteristics-centric to task utility-centric data augmentation.

UtilGen هو نهج مبتكر لزيادة البيانات يركز على نماذج توليدية موجهة نحو الفائدة، مما يعزز الأداء في مهام رؤية الكمبيوتر. على عكس الطرق التقليدية التي تعطي الأولوية للجودة البصرية، يتكيف UtilGen مع متطلبات المهام المحددة، مما يضمن أن البيانات المولدة ليست فقط متنوعة ولكن أيضًا ذات صلة بالتطبيقات اللاحقة. هذه الابتكار مهم لأنه يجسر الفجوة بين توليد البيانات والاستخدام العملي، مما قد يؤدي إلى تحسين النتائج في مختلف تطبيقات الذكاء الاصطناعي.

UtilGen es un enfoque innovador para la augmentación de datos que se centra en modelos generativos orientados a la utilidad, mejorando el rendimiento en tareas de visión por computadora. A diferencia de los métodos tradicionales que priorizan la calidad visual, UtilGen se adapta a los requisitos específicos de las tareas, asegurando que los datos generados sean no solo diversos, sino también relevantes para aplicaciones posteriores. Esta innovación es significativa ya que cierra la brecha entre la generación de datos y su uso práctico, lo que podría llevar a mejores resultados en diversas aplicaciones de IA.

UtilGen est une approche révolutionnaire de l'augmentation de données qui se concentre sur des modèles génératifs centrés sur l'utilité, améliorant ainsi les performances dans les tâches de vision par ordinateur. Contrairement aux méthodes traditionnelles qui privilégient la qualité visuelle, UtilGen s'adapte aux exigences spécifiques des tâches, garantissant que les données générées sont non seulement diverses mais aussi pertinentes pour les applications en aval. Cette innovation est importante car elle comble le fossé entre la génération de données et l'utilisation pratique, ce qui pourrait conduire à de meilleurs résultats dans diverses applications d'IA.

UtilGen is a groundbreaking approach to data augmentation that focuses on utility-centric generative models, enhancing performance in computer vision tasks. Unlike traditional methods that prioritize visual quality, UtilGen adapts to specific task requirements, ensuring that the generated data is not only diverse but also relevant for downstream applications. This innovation is significant as it bridges the gap between data generation and practical use, potentially leading to improved outcomes in various AI applications.

UtilGen: Utility-Centric Generative Data Augmentation with Dual-Level Task Adaptation

arXiv:2601.07982v1 Announce Type: new 
Abstract: We develop a new statistical ideal observer model that performs holistic visual search (or gist) processing in part by placing thresholds on minimum extractable image features. In this model, the ideal observer reduces the number of free parameters thereby shrinking down the system. The applications of this novel framework is in medical image perception (for optimizing imaging systems and algorithms), computer vision, benchmarking performance and enabling feature selection/evaluations. Other applications are in target detection and recognition in defense/security as well as evaluating sensors and detectors.

تم تطوير نموذج جديد للمراقب المثالي الإحصائي لتحسين معالجة البحث البصري الشامل من خلال وضع حدود على الميزات القابلة للاستخراج من الصورة. يهدف هذا النموذج إلى تبسيط النظام من خلال تقليل المعلمات الحرة، مع تطبيقات في إدراك الصور الطبية ورؤية الكمبيوتر والدفاع/الأمن.

Se ha desarrollado un nuevo modelo estadístico de observador ideal para mejorar el procesamiento de búsqueda visual holística al establecer umbrales sobre las características de imagen mínimas extraíbles. Este modelo busca optimizar el sistema reduciendo los parámetros libres, con aplicaciones en percepción de imágenes médicas, visión por computadora y defensa/seguridad.

Un nouveau modèle d'observateur idéal statistique a été développé pour améliorer le traitement de la recherche visuelle holistique en établissant des seuils sur les caractéristiques d'image minimales extractibles. Ce modèle vise à rationaliser le système en réduisant les paramètres libres, avec des applications dans la perception d'images médicales, la vision par ordinateur et la défense/sécurité.

A new statistical ideal observer model has been developed to enhance holistic visual search processing by establishing thresholds on minimum extractable image features. This model aims to streamline the system by reducing free parameters, with applications in medical image perception, computer vision, and defense/security.

Likelihood ratio for a binary Bayesian classifier under a noise-exclusion model

arXiv:2601.07976v1 Announce Type: cross 
Abstract: This study advances task-based image quality assessment by developing an anthropomorphic thresholded visual-search model observer. The model is an ideal observer for thresholded data inspired by the human visual system, allowing selective processing of high-salience features to improve discrimination performance. By filtering out irrelevant variability, the model enhances diagnostic accuracy and computational efficiency.
  The observer employs a two-stage framework: candidate selection and decision-making. Using thresholded data during candidate selection refines regions of interest, while stage-specific feature processing optimizes performance. Simulations were conducted to evaluate the effects of thresholding on feature maps, candidate localization, and multi-feature scenarios. Results demonstrate that thresholding improves observer performance by excluding low-salience features, particularly in noisy environments. Intermediate thresholds often outperform no thresholding, indicating that retaining only relevant features is more effective than keeping all features.
  Additionally, the model demonstrates effective training with fewer images while maintaining alignment with human performance. These findings suggest that the proposed novel framework can predict human visual search performance in clinically realistic tasks and provide solutions for model observer training with limited resources. Our novel approach has applications in other areas where human visual search and detection tasks are modeled such as in computer vision, machine learning, defense and security image analysis.

قدمت دراسة حديثة نموذجًا للمراقب البصري الأنثروبومورفي المعتمد على العتبة، مما يعزز تقييم جودة الصورة القائم على المهام من خلال تقليد النظام البصري البشري. يقوم هذا النموذج بمعالجة الميزات ذات الأهمية العالية بشكل انتقائي، مما يحسن أداء التمييز والدقة التشخيصية بينما يقوم بتصفية التباين غير ذي الصلة.

Un estudio reciente ha presentado un modelo de observador visual antropomórfico umbralizado, mejorando la evaluación de la calidad de imagen basada en tareas al imitar el sistema visual humano. Este modelo procesa selectivamente características de alta saliencia, mejorando el rendimiento de discriminación y la precisión diagnóstica mientras filtra la variabilidad irrelevante.

Une étude récente a introduit un modèle d'observateur visuel anthropomorphique à seuil, améliorant l'évaluation de la qualité d'image basée sur des tâches en imitant le système visuel humain. Ce modèle traite sélectivement des caractéristiques de haute saillance, améliorant la performance de discrimination et la précision diagnostique tout en filtrant la variabilité non pertinente.

A recent study has introduced an anthropomorphic thresholded visual-search model observer, enhancing task-based image quality assessment by mimicking the human visual system. This model selectively processes high-salience features, improving discrimination performance and diagnostic accuracy while filtering out irrelevant variability.

Application of Ideal Observer for Thresholded Data in Search Task

arXiv:2505.18227v3 Announce Type: replace 
Abstract: In Transformer architectures, tokens\textemdash discrete units derived from raw data\textemdash are formed by segmenting inputs into fixed-length chunks. Each token is then mapped to an embedding, enabling parallel attention computations while preserving the input's essential information. Due to the quadratic computational complexity of transformer self-attention mechanisms, token reduction has primarily been used as an efficiency strategy. This is especially true in single vision and language domains, where it helps balance computational costs, memory usage, and inference latency. Despite these advances, this paper argues that token reduction should transcend its traditional efficiency-oriented role in the era of large generative models. Instead, we position it as a fundamental principle in generative modeling, critically influencing both model architecture and broader applications. Specifically, we contend that across vision, language, and multimodal systems, token reduction can: (i) facilitate deeper multimodal integration and alignment, (ii) mitigate "overthinking" and hallucinations, (iii) maintain coherence over long inputs, and (iv) enhance training stability, etc. We reframe token reduction as more than an efficiency measure. By doing so, we outline promising future directions, including algorithm design, reinforcement learning-guided token reduction, token optimization for in-context learning, agentic framework design, and broader ML and scientific domains.

يبرز مقال حديث أن تقليل الرموز في هياكل المحولات يجب أن يتجاوز مجرد الكفاءة، داعيًا إلى دوره كمبدأ أساسي في النمذجة التوليدية عبر مجالات متعددة، بما في ذلك الرؤية واللغة.

Un artículo reciente enfatiza que la reducción de tokens en arquitecturas Transformer debería ir más allá de la mera eficiencia, abogando por su papel como un principio fundamental en la modelización generativa en varios dominios, incluyendo visión y lenguaje.

Un article récent souligne que la réduction des tokens dans les architectures Transformer devrait aller au-delà de la simple efficacité, plaidant pour son rôle en tant que principe fondamental dans la modélisation générative à travers divers domaines, y compris la vision et le langage.

A recent paper emphasizes that token reduction in Transformer architectures should extend beyond mere efficiency, advocating for its role as a fundamental principle in generative modeling across various domains, including vision and language.

Token Reduction Should Go Beyond Efficiency in Generative Models -- From Vision, Language to Multimodality

One More Thing in AI – Your Shortcut to AI Mastery

UtilGen: Utility-Centric Generative Data Augmentation with Dual-Level Task Adaptation

Was this article worth reading? Share it

One More Thing in AI

Magicley AI

Airparser

Omnifact

GPTHumanizer

Uwear

Ready to build your own newsroom?