arXiv:2512.02823v1 Announce Type: cross 
Abstract: Model-based filtering is often carried out while subject to an imperfect model, as learning partially-observable stochastic systems remains a challenge. Recent work on Bayesian inference found that tempering the likelihood or full posterior of an imperfect model can improve predictive accuracy, as measured by expected negative log likelihood. In this paper, we develop the tempered Bayes filter, improving estimation performance through both of the aforementioned, and one newly introduced, modalities. The result admits a recursive implementation with a computational complexity no higher than that of the original Bayes filter. Our analysis reveals that -- besides the well-known fact in the field of Bayesian inference that likelihood tempering affects the balance between prior and likelihood -- full-posterior tempering tunes the level of entropy in the final belief distribution. We further find that a region of the tempering space can be understood as interpolating between the Bayes- and MAP filters, recovering these as special cases. Analytical results further establish conditions under which a tempered Bayes filter achieves improved predictive performance. Specializing the results to the linear Gaussian case, we obtain the tempered Kalman filter. In this context, we interpret how the parameters affect the Kalman state estimate and covariance propagation. Empirical results confirm that our method consistently improves predictive accuracy over the Bayes filter baseline.

تم تقديم نهج جديد للتصفية المعتمدة على النماذج مع فلتر بايز المكيف، الذي يحسن أداء التقدير من خلال تعديل الاحتمالية والاحتمالية الكاملة لنموذج غير كامل. تعالج هذه الطريقة التحديات المتعلقة بتعلم الأنظمة العشوائية القابلة للملاحظة جزئيًا وتحافظ على كفاءة حسابية مقارنة بالفلاتر الأصلية.

Se ha introducido un nuevo enfoque para el filtrado basado en modelos con el filtro Bayes templado, que mejora el rendimiento de estimación al templar la verosimilitud y la posterior completa de un modelo imperfecto. Este método aborda los desafíos de aprender sistemas estocásticos parcialmente observables y mantiene una eficiencia computacional comparable a la del filtro Bayes original.

Une nouvelle approche de filtrage basée sur des modèles a été introduite avec le filtre Bayes tempéré, qui améliore la performance d'estimation en tempérant la vraisemblance et la postérieure complète d'un modèle imparfait. Cette méthode répond aux défis d'apprentissage des systèmes stochastiques partiellement observables tout en maintenant une efficacité computationnelle comparable à celle du filtre Bayes original.

A new approach to model-based filtering has been introduced with the tempered Bayes filter, which enhances estimation performance by tempering the likelihood and full posterior of an imperfect model. This method addresses the challenges of learning partially-observable stochastic systems and maintains computational efficiency comparable to the original Bayes filter.

Tempering the Bayes Filter towards Improved Model-Based Estimation

arXiv:2512.02722v1 Announce Type: new 
Abstract: Uncertainty quantification is essential for deploying reliable Graph Neural Networks (GNNs), where existing approaches primarily rely on Bayesian inference or ensembles. In this paper, we introduce the first credal graph neural networks (CGNNs), which extend credal learning to the graph domain by training GNNs to output set-valued predictions in the form of credal sets. To account for the distinctive nature of message passing in GNNs, we develop a complementary approach to credal learning that leverages different aspects of layer-wise information propagation. We assess our approach on uncertainty quantification in node classification under out-of-distribution conditions. Our analysis highlights the critical role of the graph homophily assumption in shaping the effectiveness of uncertainty estimates. Extensive experiments demonstrate that CGNNs deliver more reliable representations of epistemic uncertainty and achieve state-of-the-art performance under distributional shift on heterophilic graphs.

تم تقديم إطار جديد يسمى الشبكات العصبية البيانية القابلة (CGNNs)، مما يعزز تقدير عدم اليقين في الشبكات العصبية البيانية (GNNs) من خلال تمكين التنبؤات ذات القيم المجموعة عبر المجموعات القابلة. تتناول هذه الطريقة المبتكرة قيود الأساليب الحالية التي تعتمد بشكل أساسي على الاستدلال البايزي أو الفرق، وخاصة في تصنيف العقد تحت ظروف خارج التوزيع.

Se ha introducido un nuevo marco llamado Redes Neuronales Gráficas Credales (CGNNs), que mejora la cuantificación de la incertidumbre en las Redes Neuronales Gráficas (GNNs) al permitir predicciones con valores conjuntos a través de conjuntos credales. Este enfoque innovador aborda las limitaciones de los métodos existentes que se basan principalmente en la inferencia bayesiana o en conjuntos, especialmente en la clasificación de nodos en condiciones fuera de distribución.

Un nouveau cadre appelé Réseaux de Neurones Graphiques Crédaux (CGNNs) a été introduit, améliorant la quantification de l'incertitude dans les Réseaux de Neurones Graphiques (GNNs) en permettant des prédictions à valeurs d'ensemble via des ensembles crédaux. Cette approche innovante répond aux limites des méthodes existantes qui s'appuient principalement sur l'inférence bayésienne ou les ensembles, en particulier dans la classification des nœuds dans des conditions hors distribution.

A new framework called Credal Graph Neural Networks (CGNNs) has been introduced, enhancing uncertainty quantification in Graph Neural Networks (GNNs) by enabling set-valued predictions through credal sets. This innovative approach addresses the limitations of existing methods that primarily rely on Bayesian inference or ensembles, particularly in node classification under out-of-distribution conditions.

Credal Graph Neural Networks

arXiv:2512.01366v1 Announce Type: cross 
Abstract: Failing to be aware of speeding vehicles approaching from behind poses a huge threat to the road safety of pedestrians and cyclists. In this paper, we propose BlinkBud, which utilizes a single earbud and a paired phone to online detect hazardous objects approaching from behind of a user. The core idea is to accurately track visually identified objects utilizing a small number of sampled camera images taken from the earbud. To minimize the power consumption of the earbud and the phone while guaranteeing the best tracking accuracy, a novel 3D object tracking algorithm is devised, integrating both a Kalman filter based trajectory estimation scheme and an optimal image sampling strategy based on reinforcement learning. Moreover, the impact of constant user head movements on the tracking accuracy is significantly eliminated by leveraging the estimated pitch and yaw angles to correct the object depth estimation and align the camera coordinate system to the user's body coordinate system, respectively. We implement a prototype BlinkBud system and conduct extensive real-world experiments. Results show that BlinkBud is lightweight with ultra-low mean power consumptions of 29.8 mW and 702.6 mW on the earbud and smartphone, respectively, and can accurately detect hazards with a low average false positive ratio (FPR) and false negative ratio (FNR) of 4.90% and 1.47%, respectively.

تم تقديم BlinkBud كحل مبتكر لتعزيز سلامة المشاة وراكبي الدراجات من خلال اكتشاف الأشياء الخطرة التي تقترب من الخلف باستخدام سماعة أذن واحدة وهاتف مرتبط. يستخدم النظام خوارزمية فريدة لتتبع الأجسام ثلاثية الأبعاد تدمج بين مرشح كالمان والتعلم المعزز لتحسين دقة التتبع مع تقليل استهلاك الطاقة.

BlinkBud se ha presentado como una solución innovadora para mejorar la seguridad de peatones y ciclistas al detectar objetos peligrosos que se acercan por detrás utilizando un solo auricular y un teléfono emparejado. El sistema emplea un algoritmo único de seguimiento de objetos en 3D que integra un filtro de Kalman y aprendizaje por refuerzo para optimizar la precisión del seguimiento mientras minimiza el consumo de energía.

BlinkBud a été introduit comme une solution novatrice pour améliorer la sécurité des piétons et des cyclistes en détectant les objets dangereux approchant par derrière à l'aide d'une seule oreillette et d'un téléphone associé. Le système utilise un algorithme de suivi d'objets 3D unique qui intègre un filtre de Kalman et l'apprentissage par renforcement pour optimiser la précision du suivi tout en minimisant la consommation d'énergie.

BlinkBud has been introduced as a novel solution to enhance pedestrian and cyclist safety by detecting hazardous objects approaching from behind using a single earbud and a paired phone. The system employs a unique 3D object tracking algorithm that integrates a Kalman filter and reinforcement learning to optimize tracking accuracy while minimizing power consumption.

Tempering the Bayes Filter towards Improved Model-Based Estimation

Was this article worth reading? Share it

finlight.me

LucidQuery AI

MyFramework