Search papers, labs, and topics across Lattice.
University of Melbourne
6
0
8
10
TC-UAP is the first method to effectively safeguard videos from both reference- and tuning-based customization, achieving unprecedented identity protection.
Fine-tuning a 3D GAN with human preferences directly on radiance fields yields face geometries that users overwhelmingly prefer, without the need for complex mesh extraction.
FedReLa achieves significant accuracy gains for minority classes in federated learning without requiring knowledge of global class distributions.
TimeLAVA reveals that a learning-agnostic approach can significantly enhance data valuation in time series, outperforming traditional methods by effectively capturing temporal dynamics.
Forget tedious optimization – LightHarmony3D generates realistic lighting and shadows for inserted 3D objects in a single pass, making scene augmentation feel truly real.
Generating 3D scenes with diffusion models just got a whole lot more consistent across views, thanks to a new 3D-native approach that skips the 2D latent space bottleneck.
