• Publications
  • Influence
OneGAN: Simultaneous Unsupervised Learning of Conditional Image Generation, Foreground Segmentation, and Fine-Grained Clustering
We present a method for simultaneously learning, in an unsupervised manner, (i) a conditional image generator, (ii) foreground extraction and segmentation, (iii) clustering into a two-level class
Scale-Localized Abstract Reasoning
TLDR
A modified version of the RAVen dataset is proposed, named RAVEN-FAIR, which outperforms the existing state of the art in this task on all benchmarks by 5-54% and proposes a new way to pool information along the rows and the columns of the illustration-grid of the query.
Scene Graph tO Image Generation with Contextualized Object Layout Refinement
TLDR
This work proposes a method that alleviates generated images with high inter-object overlap, empty areas, blurry objects, and overall compromised quality by generating all object layouts together and reducing the reliance on supervised learning.
Evaluation Metrics for Conditional Image Generation
TLDR
An extensive empirical evaluation is provided, comparing the metrics to their unconditional variants and to other metrics, and utilize them to analyze existing generative models, thus providing additional insights about their performance, from unlearned classes to mode collapse.
Generating Correct Answers for Progressive Matrices Intelligence Tests
TLDR
The proposed neural model combines multiple advances in generative models, including employing multiple pathways through the same network, using the reparameterization trick along two pathways to make their encoding compatible and a dynamic application of variational losses.
Dynamic Dual-Output Diffusion Models
TLDR
This paper reveals some of the causes that affect the generation quality of diffusion models, especially when sampling with few iterations, and comes up with a simple, yet effective, solution to mitigate them.