Distributionally Robust Recourse Action

Authors: Duy Nguyen, Ngoc Bui, Viet Anh Nguyen

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

Reproducibility Variable Result LLM Response
Research Type Experimental Numerical experiments with both synthetic and three real-world datasets demonstrate the benefits of our proposed framework over state-of-the-art recourse methods. (Abstract) and We compare extensively the performance of our Di RRAc model (2) and Gaussian Di RRAc model (8) against four strong baselines: ROAR (Upadhyay et al., 2021), CEPM (Pawelczyk et al., 2020), AR (Ustun et al., 2019) and Wachter (Wachter et al., 2018). We conduct the experiments on three real-world datasets (German, SBA, Student). (Section 5)
Researcher Affiliation Collaboration Duy Nguyen1, Ngoc Bui1, Viet Anh Nguyen2 1Vin AI Research, Vietnam 2The Chinese University of Hong Kong
Pseudocode Yes Algorithm 1 Projected gradient descent algorithm with backtracking line-search (Appendix E)
Open Source Code Yes Source code can be found at https://github.com/duykhuongnguyen/Di RRAc. (Appendix A.1)
Open Datasets Yes We use three real-world datasets which capture different data distribution shifts (Dua & Graff, 2017): (i) the German credit dataset, which captures a correction shift. (ii) the Small Business Administration (SBA) dataset, which captures a temporal shift. (iii) the Student performance dataset, which captures a geospatial shift. (Section 5) and For the German credit dataset from the UCI repository... (Appendix A.1)
Dataset Splits No The paper states We split 80% of the original dataset and train a logistic classifier but does not explicitly mention a separate validation split for hyperparameter tuning or early stopping, or specify a cross-validation setup.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU models, CPU types, or cloud computing specifications) used to run the experiments.
Software Dependencies Yes off-the-shelf solvers such as GUROBI Gurobi Optimization, LLC (2021) or Mosek (MOSEK Ap S, 2019). (Section 3.3) and MOSEK Optimizer API for Python 9.2.10, 2019. (Reference section, MOSEK Ap S entry)
Experiment Setup Yes The detailed settings are provided in Table 4. Table 4: Parameters for the experiments with real-world data in Table 1. Parameters Values K 1 δadd 1.0 bp [1] ρ [0.1] λ 0.7 ζ 1 (Appendix A.2)