Published in 3DV, 2022
We propose Geometric Neural Parametric Models (GNPM), a learned parametric model that takes into account the local structure of data to learn disentangled shape and pose latent spaces of 4D dynamics, using a geometric-aware architecture on point clouds.
Recommended citation: Mirgahney Mohamed, Lourdes Agapito (2022). "GNPM: Geometric-Aware Neural Parametric Models." 3DV. 1(1). https://arxiv.org/abs/2209.10621.pdf
Published in , 2022
This paper is about the number 1. The number 2 is left for future work.
Recommended citation: Mohamed, Mirgahney et.al. (2021). "Data and compute efficient equivariant convolutional networks; US Patent App. 17/170,745. https://patentimages.storage.googleapis.com/d8/3b/56/ab431403304b8c/US20210248467A1.pdf
Published in ICLR, AI4 Developing Countries., 2020
We design and test an equivariant version of MobileNetV2 and further optimize it with model quantization to enable more efficient inference.
Recommended citation: Mohamed, Mirgahney, et al. "A data and compute efficient design for limited-resources deep learning." ICLR, AI4 Developing Countries. (2020). https://arxiv.org/abs/2004.09691.pdf