paper-with-me

Papers

Discrete Flows: Invertible Generative Models of Discrete Data

2019-05-24 · NeurIPS 2019 12 · Dustin Tran, Keyon Vafa, Kumar Krishna Agrawal, Laurent Dinh, Ben Poole

While normalizing flows have led to significant advances in modeling high-dimensional continuous distributions, their applicability to discrete distributions remains unknown. In this paper, we show that flows can in fact be extended to discrete events---and under a simple change-of-variables formula not requiring log-determinant-Jacobian computations. Discrete flows have numerous applications. We consider two flow architectures: discrete autoregressive flows that enable bidirectionality, allowing, for example, tokens in text to depend on both left-to-right and right-to-left contexts in an exact language model; and discrete bipartite flows that enable efficient non-autoregressive generation as in RealNVP. Empirically, we find that discrete autoregressive flows outperform autoregressive baselines on synthetic discrete distributions, an addition task, and Potts models; and bipartite flows can obtain competitive performance with autoregressive baselines on character-level language modeling for Penn Tree Bank and text8.

📄 PDF Abstract BibTeX arXiv:1905.10347

Code (2)

google/edward2 공식 구현 tf
TrentBrick/PyTorchDiscreteFlows pytorch

Tasks

Language ModelingLanguage Modelling

Methods 이 논문이 사용한 방법론

Affine Coupling 설명 없음
Batch Normalization 설명 없음
RealNVP RealNVP is a generative model that utilises real-valued non-volume preserving (real NVP) transformations for density estimation. The model can perform efficient and exact…
Normalizing Flows Normalizing Flows are a method for constructing complex distributions by transforming a probability density through a series of invertible mappings. By repeatedly applying…

Similar Papers 제목 키워드 기반

Discrete Denoising Flows

2021-07-24 · ICML Workshop INNF 2021 7 · Alexandra Lindt, Emiel Hoogeboom

Discrete flow-based models are a recently proposed class of generative models that learn invertible transformations for discrete random variables. Since they do not require data dequantization and maximize an exact likel…

Denoising

General Invertible Transformations for Flow-based Generative Modeling

2020-11-30 · Jakub M. Tomczak

In this paper, we present a new class of invertible transformations with an application to flow-based generative models. We indicate that many well-known invertible transformations in reversible logic and reversible neur…

General Invertible Transformations for Flow-based Generative Modeling

2021-06-02 · ICML Workshop INNF 2021 7 · Jakub Mikolaj Tomczak

In this paper, we present a new class of invertible transformations with an application to flow-based generative models. We indicate that many well-known invertible transformations in reversible logic and reversible neur…

IDF++: Analyzing and Improving Integer Discrete Flows for Lossless Compression

2020-06-22 · ICLR 2021 1 · Rianne van den Berg, Alexey A. Gritsenko, Mostafa Dehghani, Casper Kaae Sønderby 외

In this paper we analyse and improve integer discrete flows for lossless compression. Integer discrete flows are a recently proposed class of models that learn invertible transformations for integer-valued random variabl…

Quantization

Fast Lossless Neural Compression with Integer-Only Discrete Flows

2022-06-17 · Siyu Wang, Jianfei Chen, Chongxuan Li, Jun Zhu 외

By applying entropy codecs with learned data distributions, neural compressors have significantly outperformed traditional codecs in terms of compression ratio. However, the high inference latency of neural networks hind…

Quantization