On Dynamic Programming Decompositions of Static Risk Measures in Markov Decision Processes

Authors: Jia Lin Hau, Erick Delage, Mohammad Ghavamzadeh, Marek Petrik

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical However, we show that these popular decompositions for Conditional-Value-at-Risk (CVa R) and Entropic Value-at-Risk (EVa R) are inherently suboptimal regardless of the discretization level. In particular, we show that a saddle point property assumed to hold in prior literature may be violated. However, a decomposition does hold for Value-at-Risk and our proof demonstrates how this risk measure differs from CVa R and EVa R.
Researcher Affiliation Collaboration Jia Lin Hau University of New Hampshire Durham, NH jialin.hau@unh.edu Erick Delage HEC Montréal Montréal (Québec) erick.delage@hec.ca Mohammad Ghavamzadeh Amazon Palo Alto, CA ghavamza@amazon.com Marek Petrik University of New Hampshire Durham, NH mpetrik@cs.unh.edu
Pseudocode No The paper does not contain any clearly labeled pseudocode or algorithm blocks.
Open Source Code No The paper does not include any explicit statements or links indicating that open-source code for the described methodology is provided.
Open Datasets No The paper is theoretical and does not conduct experiments on datasets, thus no information about publicly available training datasets is provided.
Dataset Splits No The paper is theoretical and does not conduct experiments involving dataset splits for training, validation, or testing.
Hardware Specification No The paper is theoretical and does not describe experiments that would require specific hardware, therefore, no hardware specifications are mentioned.
Software Dependencies No The paper is theoretical and does not specify software dependencies with version numbers for replication.
Experiment Setup No The paper is theoretical and focuses on mathematical analysis and proofs; therefore, it does not describe an experimental setup with specific hyperparameters or system-level training settings.