3 citations · 9 across the 6 of their papers we have counts for
6 papers
GIST: Towards Photorealistic Style Transfer via Multiscale Geometric Representations
Renan A. Rojas-Gomez, Minh N. Do
State-of-the-art Style Transfer methods often leverage pre-trained encoders optimized for discriminative tasks, which may not be ideal for image synthesis. This can result in signi…
Augmentations vs Algorithms: What Works in Self-Supervised Learning
Warren Morningstar, Alex Bijamov, Chris Duvarney +8
We study the relative effects of data augmentations, pretraining algorithms, and model architectures in Self-Supervised Learning (SSL). While the recent literature in this space le…
SASSL: Enhancing Self-Supervised Learning via Neural Style Transfer
Renan A. Rojas-Gomez, Karan Singhal, Ali Etemad +3
Existing data augmentation in self-supervised learning, while diverse, fails to preserve the inherent structure of natural images. This results in distorted augmented samples with…
Making Vision Transformers Truly Shift-Equivariant
Renan A. Rojas-Gomez, Teck-Yian Lim, Minh N. Do +1
For computer vision, Vision Transformers (ViTs) have become one of the go-to deep net architectures. Despite being inspired by Convolutional Neural Networks (CNNs), ViTs' output re…
Learnable Polyphase Sampling for Shift Invariant and Equivariant Convolutional Networks
Renan A. Rojas-Gomez, Teck-Yian Lim, Alexander G. Schwing +2
We propose learnable polyphase sampling (LPS), a pair of learnable down/upsampling layers that enable truly shift-invariant and equivariant convolutional networks. LPS can be train…
Inverting Adversarially Robust Networks for Image Synthesis
Renan A. Rojas-Gomez, Raymond A. Yeh, Minh N. Do +1
Despite unconditional feature inversion being the foundation of many image synthesis applications, training an inverter demands a high computational budget, large decoding capacity…