Selected work

Representative Papers

IG-GAN: A Generative Adversarial Network for Aerodynamic Data Generation Based on Intrinsic Geometry

Jul 13, 2026

Traditional generative models struggle to effectively capture the non-Euclidean manifold structure inherent in aerodynamic data within Euclidean space. To address this limitation, this work proposes the Intrinsic Geometry Generative Adversarial Network (IG-GAN), which uniquely integrates Bézier surfaces with intrinsic geometry. IG-GAN explicitly constructs a globally smooth manifold by learning the coefficients of piecewise-smooth Bézier surfaces and introduces a radial basis function–based discriminator (RBF-D) for optimization. Evaluated on the Burgers’ equation dataset, the method reduces the mean squared error (MSE) of velocity field prediction by 97.41% compared to SSL-Transformer. On the ONERA M6 aircraft dataset, it achieves an 82.95% reduction in overall MSE across nine aerodynamic coefficients, demonstrating substantially improved generation accuracy for data residing on non-Euclidean manifolds.

0 citationsRead paper

Motion Estimation for Multi-Object Tracking using KalmanNet with Semantic-Independent Encoding

Sep 14, 2025

To address the degraded estimation performance of linear constant-velocity Kalman filtering in multi-object tracking—caused by model mismatch and non-stationary motion—this paper proposes a learning-enhanced Semantic-Independent Kalman Filter (SIC-KF). Our core innovation is a lightweight semantic-independent encoder comprising 1D convolutions (kernel size 1), fully connected layers, and nonlinear activation modules, which jointly learns decoupled representations of homogeneous semantic features and captures nonlinear dependencies among heterogeneous elements, thereby enhancing motion-awareness of the state vector. SIC-KF is seamlessly embedded into the KalmanNet framework to enable end-to-end differentiable training. Experiments on a large-scale, semi-synthetic dataset constructed by us demonstrate that SIC-KF significantly outperforms conventional Kalman filters and state-of-the-art learning-based filters, achieving new SOTA performance in both trajectory prediction accuracy and robustness to motion abruptness.

0 citationsRead paper

Recent publications

Latest Papers

IG-GAN: A Generative Adversarial Network for Aerodynamic Data Generation Based on Intrinsic Geometry

Jul 13, 2026

Traditional generative models struggle to effectively capture the non-Euclidean manifold structure inherent in aerodynamic data within Euclidean space. To address this limitation, this work proposes the Intrinsic Geometry Generative Adversarial Network (IG-GAN), which uniquely integrates Bézier surfaces with intrinsic geometry. IG-GAN explicitly constructs a globally smooth manifold by learning the coefficients of piecewise-smooth Bézier surfaces and introduces a radial basis function–based discriminator (RBF-D) for optimization. Evaluated on the Burgers’ equation dataset, the method reduces the mean squared error (MSE) of velocity field prediction by 97.41% compared to SSL-Transformer. On the ONERA M6 aircraft dataset, it achieves an 82.95% reduction in overall MSE across nine aerodynamic coefficients, demonstrating substantially improved generation accuracy for data residing on non-Euclidean manifolds.

0 citationsRead paper

Motion Estimation for Multi-Object Tracking using KalmanNet with Semantic-Independent Encoding

Sep 14, 2025

To address the degraded estimation performance of linear constant-velocity Kalman filtering in multi-object tracking—caused by model mismatch and non-stationary motion—this paper proposes a learning-enhanced Semantic-Independent Kalman Filter (SIC-KF). Our core innovation is a lightweight semantic-independent encoder comprising 1D convolutions (kernel size 1), fully connected layers, and nonlinear activation modules, which jointly learns decoupled representations of homogeneous semantic features and captures nonlinear dependencies among heterogeneous elements, thereby enhancing motion-awareness of the state vector. SIC-KF is seamlessly embedded into the KalmanNet framework to enable end-to-end differentiable training. Experiments on a large-scale, semi-synthetic dataset constructed by us demonstrate that SIC-KF significantly outperforms conventional Kalman filters and state-of-the-art learning-based filters, achieving new SOTA performance in both trajectory prediction accuracy and robustness to motion abruptness.

0 citationsRead paper