NeurIPS2022Scientific poster design
Langevin Autoencoders for Learning Deep Latent Variable Models
A real NeurIPS 2022 conference poster, shown as a scientific poster design reference. Study how the authors lay out the title, figures, and results, then build your own.
Can a reader answer these from the poster?
A well-designed poster lets a passer-by grasp these in seconds — a useful checklist for your own scientific poster design.
- →What are the two main tools used to approximate intractable distributions in deep latent variable models?
- →What is the primary limitation of traditional MCMC methods in the context of deep latent variable models?
- →What is the primary goal of the proposed amortized Langevin dynamics (ALD)?
- →What is the key novelty of the Langevin autoencoder (LAE) introduced in the paper?
- →How does the amortized Langevin dynamics (ALD) method work?
