3 papers
cs.MM2026
LLM2Manim: Pedagogy-Aware AI Generation of STEM Animations
Aastha Joshi, Hongyi Ke, Meet Gajjar +3
High-quality STEM animations can be useful for learning, but they are still not common in daily teaching, mostly because they take time and special skills to make. In this paper, w…
physics.flu-dyn2026
Preconditioned Adjoint Data Assimilation for Two-Dimensional Decaying and Forced Turbulence
Hongyi Ke, Zejian You, Qi Wang
Adjoint-based data assimilation for turbulent Navier-Stokes flows is limited by backward adjoint growth and increasing dominance of small-scale structures, which degrade reconstruc…
cs.LG2025
Fourier-Invertible Neural Encoder (FINE) for Homogeneous Flows
Anqiao Ouyang, Hongyi Ke, Qi Wang
We present the Fourier-Invertible Neural Encoder (FINE), a compact and interpretable architecture for dimension reduction in translation-equivariant datasets. FINE integrates rever…