Surrogate models that integrate dimensionality reduction with regression techniques are crucial for reducing reliance on costly Computational Fluid Dynamics (CFD) simulations in aerodynamics [1]. This work introduces a β-Variational Autoencoder (β-VAE) framework, enhanced with Principal Component Analysis (PCA), to predict pressure distributions over a transonic wing [2]. The model maps high-dimensional aerodynamic data into a low-dimensional latent space correlated with flight conditions, such as Mach number and angle of attack [3]. Regularization via the β parameter and PCA pre-processing improve reconstruction fidelity and computational efficiency, respectively. In addition, fine-tuning the decoder further enhances the model’s ability to capture complex flow phenomena, including shockwaves, and reduces the model’s sensitivity with β. Our results show that this β-VAE approach outperforms traditional methods like Proper Orthogonal Decomposition combined with neural networks [4] and provides superior accuracy in regions with complex flow phenomena. This methodology not only provides an efficient solution for the estimation of unseen flight conditions but also a scalable one, which would significantly reduce the number of high-cost CFD simulations in the processes of aerodynamic analysis and design. [1] X. Du, P. He, and J. R. Martins. Rapid airfoil design optimization via neural networks-based parameterization and surrogate modeling. Aerospace Science and Technology 113, 106701, 2021. [2] V. Francés-Belda, A. Solera-Rico, J. Nieto-Centenero, E. Andrés, C. Sanmiguel Vila, and R. Castellanos. Towards aerodynamic shape modeling based on β-variational autoencoders. Physics of Fluids (accepted), arXiv:2408.04969, 2024. [3] Y.-E. Kang, S. Yang, and K. Yee. Physics-aware reduced-order modeling of transonic flow via β-variational autoencoder. Physics of Fluids 34, 076103, 2022. [4] R. Castellanos, J. Bowen Varela, A. Gorgues, and E. Andrés, An assessment of reduced-order and machine learning models for steady transonic flow prediction on wings. ICAS 2022.