5 papers
AutoIndexer: Flexible-Order Decoding by Training Causal Models on Chains of Edits
Shu Ishida, Aliasghar Khani, Tianyu Zhang +3
Autoregressive decoding is less suited to non-causally ordered data, e.g., cross-references or future dependencies, since their outputs cannot be revised post hoc, unlike iterative…
FunPhase: A Periodic Functional Autoencoder for Motion Generation via Phase Manifolds
Marco Pegoraro, Evan Atherton, Bruno Roy +2
Learning natural body motion remains challenging due to the strong coupling between spatial geometry and temporal dynamics. Embedding motion in phase manifolds, latent spaces that…
Motion Generation: A Survey of Generative Approaches and Benchmarks
Aliasghar Khani, Arianna Rampini, Bruno Roy +5
Motion generation, the task of synthesizing realistic motion sequences from various conditioning inputs, has become a central problem in computer vision, computer graphics, and rob…
UniMoGen: Universal Motion Generation
Aliasghar Khani, Arianna Rampini, Evan Atherton +1
Motion generation is a cornerstone of computer graphics, animation, gaming, and robotics, enabling the creation of realistic and varied character movements. A significant limitatio…
Wavelet Latent Diffusion (Wala): Billion-Parameter 3D Generative Model with Compact Wavelet Encodings
Aditya Sanghi, Aliasghar Khani, Pradyumna Reddy +5
Large-scale 3D generative models require substantial computational resources yet often fall short in capturing fine details and complex geometries at high resolutions. We attribute…