11 problems

Minimax lower-bound conjecture. For distributionally robust reinforcement learning, the minimax lower bound on the number of samples is still

Uniformization conjecture. Under additional regularity assumptions such as continuous demand with a known positive lower bound on the probability density function, the class of wor…

Extension conjecture. All of the paper's analytical and algorithmic techniques should go through under common additional assumptions on the demand distribution, as noted in the dis…

Kuhn et al.'s conjecture. The construction of $\P_j^\star$ is NP-hard. The conjecture concerns the computational complexity of constructing distributions supported on the regions w…

The paper considers the primal and dual formulations of an optimal-transport distributionally robust optimization problem with conditional moment constraints, represented by martin…

Uniqueness conjecture. The infimum has a unique minimizer without assuming that the law of $\left\langle \frac{w}{|w|},X\right\rangle$ under $\mathds{P}$ is absolutely continuous w…

Let samples be drawn from distributions that admit a density and have certain concentration properties. For a sample size $K_S$, let $\hat{P}_{K_S}$ denote the resulting empirical…

Interaction conjecture. The effectiveness of a subset of scenario paths might be affected by its interaction with other subsets of scenario paths.

Let $\alpha$ be a finite positive parameter, and consider the generalized $\alpha$-unimodal Gauss bound in equation (10) for the worst-case probability of the event…

Let $\omega_{t+1}$ be a scenario that is suppressed under the modified $\chi^2$ distance, so that $p^{\omega_{t+1}\mid\omega_t*}=0$ for that divergence. For the same scenario, comp…

Distributionally robust interpretation conjecture. The one-norm regularization in the regularized DeePC problem is related to distributionally robust optimization problems, in whic…