paper-with-me

Papers

Conditional Sampling with Monotone GANs: from Generative Models to Likelihood-Free Inference

2020-06-11 · Ricardo Baptista, Bamdad Hosseini, Nikola B. Kovachki, Youssef Marzouk

We present a novel framework for conditional sampling of probability measures, using block triangular transport maps. We develop the theoretical foundations of block triangular transport in a Banach space setting, establishing general conditions under which conditional sampling can be achieved and drawing connections between monotone block triangular maps and optimal transport. Based on this theory, we then introduce a computational approach, called monotone generative adversarial networks (M-GANs), to learn suitable block triangular maps. Our algorithm uses only samples from the underlying joint probability measure and is hence likelihood-free. Numerical experiments with M-GAN demonstrate accurate sampling of conditional measures in synthetic examples, Bayesian inverse problems involving ordinary and partial differential equations, and probabilistic image in-painting.

📄 PDF Abstract BibTeX arXiv:2006.06755

Code (1)

baptistar/mgan 공식 구현 pytorch

Similar Papers 제목 키워드 기반

ColdGANs: Taming Language GANs with Cautious Sampling Strategies

2020-06-08 · NeurIPS 2020 12 · Thomas Scialom, Paul-Alexis Dray, Sylvain Lamprier, Benjamin Piwowarski 외

Training regimes based on Maximum Likelihood Estimation (MLE) suffer from known limitations, often leading to poorly generated text sequences. At the root of these limitations is the mismatch between training and inferen…

Abstractive Text SummarizationQuestion GenerationQuestion-GenerationText Generation

DiffFlow: A Unified SDE Framework for Score-Based Diffusion Models and Generative Adversarial Networks

2023-07-05 · Jingwei Zhang, Han Shi, Jincheng Yu, Enze Xie 외

Generative models can be categorized into two types: explicit generative models that define explicit density forms and allow exact likelihood inference, such as score-based diffusion models (SDMs) and normalizing flows; …

Denoising

Learning Energy-Based Models by Diffusion Recovery Likelihood

2020-12-15 · ICLR 2021 1 · Ruiqi Gao, Yang song, Ben Poole, Ying Nian Wu 외

While energy-based models (EBMs) exhibit a number of desirable properties, training and sampling on high-dimensional datasets remains challenging. Inspired by recent progress on diffusion probabilistic models, we present…

Image Generation

MaskGAN: Better Text Generation via Filling in the______

2018-01-23 · William Fedus, Ian Goodfellow, Andrew M. Dai

Neural text generation models are often autoregressive language models or seq2seq models. These models generate text by sampling words sequentially, with each word conditioned on the previous word, and are state-of-the-a…

Multivariate Time Series ImputationText Generation

Virtual Conditional Generative Adversarial Networks

2019-01-25 · Haifeng Shi, Guanyu Cai, Yuqin Wang, Shaohua Shang 외

When trained on multimodal image datasets, normal Generative Adversarial Networks (GANs) are usually outperformed by class-conditional GANs and ensemble GANs, but conditional GANs is restricted to labeled datasets and en…

ClusteringConditional Image GenerationDecoderImage Generation