5 papers
Generative Refinement for Low-Budget Black-Box Optimization
Edouard R. Dufour, Pascal Fua
Black-box optimization is a fundamental science and engineering tool that makes it possible to optimize objectives without gradient information. Unfortunately, as it often requires…
Gradient-based Nested Co-Design of Aerodynamic Shape and Control for Winged Robots
Daniele Affinita, Mingda Xu, Benoît Valentin Gherardi +1
Designing aerial robots for specialized tasks, from perching to payload delivery, requires tailoring their aerodynamic shape to specific mission requirements. For tasks involving w…
Annealed Co-Generation: Disentangling Variables via Progressive Pairwise Modeling
Hantao Zhang, Jieke Wu, Mingda Xu +3
For multivariate co-generation in scientific applications, we advocate pairwise block rather than joint modeling of all variables. This design mitigates the computational burden an…
Dflow-SUR: Enhancing Generative Aerodynamic Inverse Design using Differentiation Throughout Flow Matching
Aobo Yang, Zhen Wei, Rhea Liem +1
Generative inverse design requires incorporating physical constraints to ensure that generated designs are both reliable and accurate. However, we observe that current state-of-the…
Do you understand epistemic uncertainty? Think again! Rigorous frequentist epistemic uncertainty estimation in regression
Enrico Foglia, Benjamin Bobbia, Nikita Durasov +4
Quantifying model uncertainty is critical for understanding prediction reliability, yet distinguishing between aleatoric and epistemic uncertainty remains challenging. We extend re…