Bias Correction of Learned Generative Models using Likelihood-Free Importance Weighting

By Aditya Grover, Jiaming Song, Alekh Agarwal, Kenneth Tran, Ashish Kapoor, Eric Horvitz, Stefano Ermon. 2019

134 citations recorded by the source corpus.

A learned generative model often produces biased statistics relative to the underlying data distribution. A standard technique to correct this bias is importance sampling, where samples from the model are weighted by the likelihood ratio under model and true distributions. When the likelihood ratio is unknown, it can be estimated by training a probabilistic classifier to distinguish samples from the two distributions. We employ this likelihood-free importance weighting method to correct for the bias in generative models. We find that this technique consistently improves standard goodness-of-fit metrics for evaluating the sample quality of state-of-the-art deep generative models, suggesting reduced bias. Finally, we demonstrate its utility on representative applications in a) data augmentation for classification using generative adversarial networks, and b) model-based policy evaluation using off-policy data.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Statistical and Nonlinear Physics, Representation & Self-Supervised Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Statistical and Nonlinear Physics

In Representation & Self-Supervised Learning

Citations in the corpus

Residual Energy-Based Models for Text Generation (2020, methodology): “Lastly, this formulation enables efﬁcient evaluation and generation via importance sampling (Horvitz & Thompson, 1952; Grover et al., 2019).”
Your GAN is Secretly an Energy-based Model and You Should use Discriminator Driven Latent Sampling (2020, methodology): “[23] uses an additional discriminator as a bias corrector for generative models via importance weighting.”
NVAE: A Deep Hierarchical Variational Autoencoder (2020, methodology): “Bias correction in generative learning is an active area of research, and we recommend the interested readers to check this area [77] before building applications using this work.”
Pretrained Transformers as Universal Computation Engines (2021, background): “Mitigating these biases is an active area of research (Grover et al. 2019; Choi et al. 2020).”
Generalized Energy Based Models (2020, background): “While there has been recent interest in using the discriminator to improve the quality of the generator during sampling (Azadi et al., 2019; Turner et al., 2019; Neklyudov et al., 2019; Grover et al., 2019; Tanaka, 2019; Wu et al., 2019b), our approach emerges naturally from the model we consider.”
Towards Fairness in Visual Recognition: Effective Strategies for Bias Mitigation (2019, background): “Classical literature on mitigating bias generally operates on simpler (often linear) models [11, 50, 28], which are easier to understand and control; only recently have researchers begun looking at mitigating bias in end-to-end trained deep learning models [16, 2, 40, 18, 48, 25, 30, 36, 47, 17].”
RankGen: Improving Text Generation with Large Ranking Models (2022, background): “2006) for generative modeling (Grover et al., 2019; Parshakova et al., 2019), speech recognition (Wang and Ou, 2018), open-ended text generation (Bakhtin et al.”
Discriminator Contrastive Divergence: Semi-Amortized Generative Modeling by Exploring Energy of the Discriminator (2020, background): “, mutual information estimation [17] and bias correction of generative 141 models [12].”
Energy-Based Reranking: Improving Neural Machine Translation Using Energy-Based Models (2020, methodology): “3 Related Work Grover et al. (2019) show that importance weights can be used to make generative models better ﬁt the desired data distribution: p θ ( y ) ∝ q ( y ) ω θ ( y ) , where q ( y ) is a generative model that we can efﬁ-ciently take samples from and ω θ ( y ) is the importance weight functio”
A Survey of Controllable Text Generation Using Transformer-based Pre-trained Language Models (2022, background): “For generation-based data augmentation [42], it is necessary to ensure the data distribution balance in different domains.”
Experience Replay with Likelihood-free Importance Weights (2020, methodology): “Likelihood-free density ratio estimation have been adopted in imitation learning (IL, [17]), inverse reinforcement learning (IRL, [12]) and model-based off-policy policy evaluation (OPE, [14]).”
KALE: When Energy-Based Learning Meets Adversarial Training (2020, background): “While there has been a recent interest in using the discriminator/energy to improve the quality of the generator/base (Azadi et al., 2019; Grover et al., 2019; Turner et al., 2019; Tanaka, 2019) (Section 5), our approach emerges naturally from the energy model we consider.”
Steering Language Models With Activation Engineering (2023, background): “…performance (Hase et al. 2023; Qi et al. 2023; Brown et al. 2023); • Intervening at decoding , as with guided or trainable decoding (Gu et al. 2017; Grover et al. 2019; see Zhang et al. 2022a for an overview of controlled generation and Jin et al. 2022 for textual style transfer); • Intervening on…”
EditGAN: High-Precision Semantic Image Editing (2021, background): “It is therefore important to be aware of such biases in the underlying data and counteract them, for example by actively collecting more representative data or by using bias correction methods, an area of active research [86, 87, 88, 89].”
Inclusive GAN: Improving Data and Minority Coverage in Generative Models (2020, background): “[17] proposes an importance weighting strategy to compensate for the biases of learned generative models.”
Fair Generative Modeling via Weak Supervision (2019, methodology): “[58] use importance reweighting for alleviating model bias between pdata and pθ.”
Refining Deep Generative Models via Discriminator Gradient Flow (2020, unknown): “…in generative models, e.g., Grover et al. (2018) proposed Variational Rejection Sampling which performs rejection sampling in the latent space of VAEs to improve the variational posterior and Grover et al. (2019) used likelihood-free importance sampling for bias correction in generative models .”