Latent Variable Modeling for Generative Concept Representations and Deep Generative Models

2018/12/26 by Daniel T. Chang, Chang, Daniel T.
Computer Science · #FOS: Computer and information sciences #Generative Adversarial Networks and Image Synthesis #Image Processing and 3D Reconstruction #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Topic Modeling

paper · pdf · doi:10.48550/arxiv.1812.11856

openalex publication_date 2018/12/26 · openalex created_date 2019/01/11 · openalex updated_date 2026/07/28

Abstract

Latent representations are the essence of deep generative models and determine their usefulness and power. For latent representations to be useful as generative concept representations, their latent space must support latent space interpolation, attribute vectors and concept vectors, among other things. We investigate and discuss latent variable modeling, including latent variable models, latent representations and latent spaces, particularly hierarchical latent representations and latent space vectors and geometry. Our focus is on that used in variational autoencoders and generative adversarial networks.

Citations

• Auto-Encoding Variational Bayes
• Adversarial Feature Learning
• Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling
• Precise Recovery of Latent Vectors from Generative Adversarial Networks
• A Review of Learning with Deep Generative Models from Perspective of Graphical Modeling
• Multi-Level Variational Autoencoder: Learning Disentangled\n Representations from Grouped Observations

Related

• Sampling Generative Networks
• Comparing the latent space of generative models
• Generative Adversarial Networks: recent developments
• No Representation without Transformation
• LASERS: LAtent Space Encoding for Representations with Sparsity for Generative Modeling