arXiv:2511.17372v1 Announce Type: cross 
Abstract: Classical autoencoders are widely used to learn features of input data. To improve the feature learning, classical masked autoencoders extend classical autoencoders to learn the features of the original input sample in the presence of masked-out data. While quantum autoencoders exist, there is no design and implementation of quantum masked autoencoders that can leverage the benefits of quantum computing and quantum autoencoders. In this paper, we propose quantum masked autoencoders (QMAEs) that can effectively learn missing features of a data sample within quantum states instead of classical embeddings. We showcase that our QMAE architecture can learn the masked features of an image and can reconstruct the masked input image with improved visual fidelity in MNIST images. Experimental evaluation highlights that QMAE can significantly outperform (12.86% on average) in classification accuracy compared to state-of-the-art quantum autoencoders in the presence of masks.

قدم الباحثون Autoencoders Masked Quantum (QMAEs)، وهو نهج مبتكر يعزز تعلم الميزات في الحوسبة الكمومية من خلال إعادة بناء الصور المدخلة المmasked بدقة بصرية محسنة، مما تم إثباته بشكل خاص على مجموعات بيانات MNIST. يعتمد هذا التقدم على autoencoders المmasked التقليدية، مستفيدًا من الحالات الكمومية لتعلم الميزات المفقودة بشكل أكثر فعالية.

Investigadores han presentado los Autoencoders Máscarados Cuánticos (QMAEs), un enfoque novedoso que mejora el aprendizaje de características en la computación cuántica al reconstruir imágenes de entrada enmascaradas con una fidelidad visual mejorada, demostrado particularmente en conjuntos de datos MNIST. Este avance se basa en autoencoders enmascarados clásicos, aprovechando los estados cuánticos para aprender características faltantes de manera más efectiva.

Des chercheurs ont introduit les autoencodeurs masqués quantiques (QMAEs), une approche novatrice qui améliore l'apprentissage des caractéristiques en informatique quantique en reconstruisant des images d'entrée masquées avec une fidélité visuelle améliorée, notamment démontrée sur des ensembles de données MNIST. Cette avancée s'appuie sur des autoencodeurs masqués classiques, tirant parti des états quantiques pour apprendre plus efficacement les caractéristiques manquantes.

Researchers have introduced Quantum Masked Autoencoders (QMAEs), a novel approach that enhances feature learning in quantum computing by reconstructing masked input images with improved visual fidelity, particularly demonstrated on MNIST datasets. This advancement builds on classical masked autoencoders, leveraging quantum states to learn missing features more effectively.

Quantum Masked Autoencoders for Vision Learning

arXiv:2511.17426v1 Announce Type: cross 
Abstract: Self-supervised learning (SSL) has recently advanced through non-contrastive methods that couple an invariance term with variance, covariance, or redundancy-reduction penalties. While such objectives shape first- and second-order statistics of the representation, they largely ignore the local geometry of the underlying data manifold. In this paper, we introduce CurvSSL, a curvature-regularized self-supervised learning framework, and its RKHS extension, kernel CurvSSL. Our approach retains a standard two-view encoder-projector architecture with a Barlow Twins-style redundancy-reduction loss on projected features, but augments it with a curvature-based regularizer. Each embedding is treated as a vertex whose $k$ nearest neighbors define a discrete curvature score via cosine interactions on the unit hypersphere; in the kernel variant, curvature is computed from a normalized local Gram matrix in an RKHS. These scores are aligned and decorrelated across augmentations by a Barlow-style loss on a curvature-derived matrix, encouraging both view invariance and consistency of local manifold bending. Experiments on MNIST and CIFAR-10 datasets with a ResNet-18 backbone show that curvature-regularized SSL yields competitive or improved linear evaluation performance compared to Barlow Twins and VICReg. Our results indicate that explicitly shaping local geometry is a simple and effective complement to purely statistical SSL regularizers.

تم تقديم إطار جديد للتعلم الذاتي المراقب يسمى CurvSSL، والذي يتضمن تنظيمًا قائمًا على الانحناء لتحسين عملية التعلم من خلال أخذ الجيومتري المحلي لمناطق البيانات في الاعتبار. تعتمد هذه الطريقة على هياكل قائمة مثل Barlow Twins وتستخدم إعدادًا لمشفّر-مشروع ذو وجهتين، بهدف تحسين تعلم التمثيلات في نماذج التعلم الآلي.

Se ha introducido un nuevo marco de aprendizaje auto-supervisado llamado CurvSSL, que incorpora regularización por curvatura para mejorar el proceso de aprendizaje al considerar la geometría local de las variedades de datos. Este método se basa en arquitecturas existentes como Barlow Twins y emplea una configuración de codificador-proyector de dos vistas, con el objetivo de mejorar el aprendizaje de representaciones en modelos de aprendizaje automático.

Un nouveau cadre d'apprentissage auto-supervisé appelé CurvSSL a été introduit, intégrant une régularisation par courbure pour améliorer le processus d'apprentissage en tenant compte de la géométrie locale des variétés de données. Cette méthode s'appuie sur des architectures existantes comme Barlow Twins et utilise une configuration d'encodeur-projection à deux vues, visant à améliorer l'apprentissage des représentations dans les modèles d'apprentissage automatique.

A new self-supervised learning framework called CurvSSL has been introduced, which incorporates curvature regularization to enhance the learning process by considering the local geometry of data manifolds. This method builds on existing architectures like Barlow Twins and employs a two-view encoder-projector setup, aiming to improve representation learning in machine learning models.

Quantum Masked Autoencoders for Vision Learning

Was this article worth reading? Share it

Attentive AI

BlitzToksAI

Aqaba.ai