Distributionally Robust Counterfactual Risk Minimization

Faury, Louis; Tanielian, Ugo; Dohmatob, Elvis; Смирнова, Елена; Vasile, Flavian

doi:10.1609/aaai.v34i04.5797

Cited by 26 publications

(31 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Calafiore [34] also studied a distributionally robust portfolio selection problem in which its KL-divergence-based ambiguity of the return distribution is constructed around a discrete nominal distribution. Chen et al [37] applied the DRO model over KL-divergence-based ambiguity set to the unit commitment problem and Faury et al [63] illustrated how the KL DRO can serve as a principle tool for the counterfactual risk minimization problem. [87] considered the complexity of the new KL-divergence-based DRO approach for general problem with format (12).…”

Section: Kullback-leibler(kl)mentioning

confidence: 99%

Distributionally Robust Optimization: A review on theory and applications

Lin¹,

Fang²,

Gao³

2022

NACO

View full text Add to dashboard Cite

<p style='text-indent:20px;'>In this paper, we survey the primary research on the theory and applications of distributionally robust optimization (DRO). We start with reviewing the modeling power and computational attractiveness of DRO approaches, induced by the ambiguity sets structure and tractable robust counterpart reformulations. Next, we summarize the efficient solution methods, out-of-sample performance guarantee, and convergence analysis. Then, we illustrate some applications of DRO in machine learning and operations research, and finally, we discuss the future research directions.</p>

show abstract

Section: Kullback-leibler(kl)mentioning

confidence: 99%

Distributionally Robust Optimization: A review on theory and applications

Lin¹,

Fang²,

Gao³

2022

NACO

View full text Add to dashboard Cite

show abstract

“…Interestingly, adversarial training can also be understood as solving DRO with a Wasserstein metric-based set [Staib and Jegelka, 2017]. Applications of DRO have been explored in contextual bandits for policy learning [Si et al, 2020, Mo et al, 2020, Faury et al, 2020 and evaluation [Kato et al, 2020, Jeong and. Uncertainty sets based on KL-divergence, L 1 , L 2 and L ∞ norms have been studied [Nilim andEl Ghaoui, 2005, Iyengar, 2005] in robust RL.…”

Section: Related Workmentioning

confidence: 99%

Learning Under Adversarial and Interventional Shifts

Singh¹,

Joshi²,

Doshi‐Velez³

et al. 2021

Preprint

View full text Add to dashboard Cite

Machine learning models are often trained on data from one distribution and deployed on others. So it becomes important to design models that are robust to distribution shifts. Most of the existing work focuses on optimizing for either adversarial shifts or interventional shifts. Adversarial methods lack expressivity in representing plausible shifts as they consider shifts to joint distributions in the data. Interventional methods allow more expressivity but provide robustness to unbounded shifts, resulting in overly conservative models. In this work, we combine the complementary strengths of the two approaches and propose a new formulation, RISe, for designing robust models against a set of distribution shifts that are at the intersection of adversarial and interventional shifts. We employ the distributionally robust optimization framework to optimize the resulting objective in both supervised and reinforcement learning settings. Extensive experimentation with synthetic and real world datasets from healthcare demonstrate the efficacy of the proposed approach.

show abstract

“…Any sample z ∼ γ is now accepted with probability P a (z) = w ϕ (z)/m. Interestingly, by actively capping the importance weights as it is done in counterfactual estimation [5,8], one controls the acceptance rates P a (z) of the rejection sampling algorithm:…”

Section: Sampling From the Latent Importance Weightsmentioning

confidence: 99%

Latent reweighting, an almost free improvement for GANs

Issenhuth¹,

Tanielian²,

Picard³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Standard formulations of GANs, where a continuous function deforms a connected latent space, have been shown to be misspecified when fitting different classes of images. In particular, the generator will necessarily sample some lowquality images in between the classes. Rather than modifying the architecture, a line of works aims at improving the sampling quality from pre-trained generators at the expense of increased computational cost. Building on this, we introduce an additional network to predict latent importance weights and two associated sampling methods to avoid the poorest samples. This idea has several advantages: 1) it provides a way to inject disconnectedness into any GAN architecture, 2) since the rejection happens in the latent space, it avoids going through both the generator and the discriminator, saving computation time, 3) this importance weights formulation provides a principled way to reduce the Wasserstein's distance to the target distribution. We demonstrate the effectiveness of our method on several datasets, both synthetic and high-dimensional.

show abstract

Distributionally Robust Counterfactual Risk Minimization

Cited by 26 publications

References 14 publications

Distributionally Robust Optimization: A review on theory and applications

Distributionally Robust Optimization: A review on theory and applications

Learning Under Adversarial and Interventional Shifts

Latent reweighting, an almost free improvement for GANs

Contact Info

Product

Resources

About