Normalizing Flows for Probabilistic Modeling and Inference
Authors: George Papamakarios, Eric Nalisnick, Danilo Jimenez Rezende, Shakir Mohamed, Balaji Lakshminarayanan
JMLR 2021 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | In this review, we attempt to provide such a perspective by describing flows through the lens of probabilistic modeling and inference. We place special emphasis on the fundamental principles of flow design, and discuss foundational topics such as expressive power and computational trade-offs. We also broaden the conceptual framing of flows by relating them to more general probability transformations. Lastly, we summarize the use of flows for tasks such as generative modeling, approximate inference, and supervised learning. |
| Researcher Affiliation | Industry | George Papamakarios EMAIL Eric Nalisnick EMAIL Danilo Jimenez Rezende EMAIL Shakir Mohamed EMAIL Balaji Lakshminarayanan EMAIL Deep Mind |
| Pseudocode | Yes | Using a masked conditioner c, the above procedure can be implemented in pseudocode as follows: Initialize z to an arbitrary value for i = 1, . . . , D (h1, . . . , h D) = c(z) zi = τ 1(z i; hi). |
| Open Source Code | No | The paper is a review article summarizing existing work on normalizing flows. It does not describe a novel methodology from the authors of this paper for which open-source code would be provided. There are no statements about code release or links to repositories for the content presented in this review. |
| Open Datasets | Yes | Dinh et al. (2017) showed compelling samples from models trained on Celeb A, Image Net (64 64), CIFAR-10, and LSUN. |
| Dataset Splits | No | This paper is a review article summarizing existing work on normalizing flows and does not conduct its own experiments. Therefore, it does not provide specific dataset split information for data partitioning. |
| Hardware Specification | No | This paper is a review article and does not present new experimental results or methodology implementation. Therefore, it does not provide specific hardware details used for running experiments. |
| Software Dependencies | No | This paper is a review article and does not present new experimental results or methodology implementation. Therefore, it does not provide specific ancillary software details with version numbers for replication. |
| Experiment Setup | No | This paper is a review article and does not present new experimental results or methodology implementation. Therefore, it does not provide specific experimental setup details or hyperparameters in the main text. |