Learning Social Welfare Functions

Authors: Kanad Pardeshi, Itai Shapira, Ariel D. Procaccia, Aarti Singh

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, despite the problem s non-convexity, we demonstrate a practical algorithm for learning weighted power mean functions across tasks using simulated data and a real-world food resource allocation dataset from Lee et al. [11]. Our empirical results validate theoretical bounds and highlight algorithm performance across parameter settings.
Researcher Affiliation Academia Kanad Shrikar Pardeshi Carnegie Mellon University kpardesh@andrew.cmu.edu Itai Shapira Harvard University itaishapira@g.harvard.edu Ariel D. Procaccia Harvard University arielpro@seas.harvard.edu Aarti Singh Carnegie Mellon University aarti@andrew.cmu.edu
Pseudocode Yes The algorithm s pseudocode is presented in Algorithm 1.
Open Source Code Yes We have also provided code for the semi-synthetic experiments.
Open Datasets No The dataset we rely on (which is not publicly available) comes from the work of Lee et al. [11] with a US-based nonprofit that operates an on-demand donation transportation service supported by volunteers.
Dataset Splits No We use the default error bars provided by the Seaborn library for the plots in Section 6 and Appendices C and E. These error bars describe the variation in observed metrics for different random splits between test and train data.
Hardware Specification Yes We ran the experiments on an NVIDIA RTX A5000 GPU.
Software Dependencies No The paper mentions using the 'Seaborn library' but does not specify version numbers for any software dependencies.
Experiment Setup Yes In our experiments we choose p using grid search. However, optimization over p can also be done using other methods like simulated annealing. We minimize the ℓ2 loss in the cardinal case with weighted power mean and the logistic loss in the ordinal case with log weighted power mean. The algorithm s pseudocode is presented in Algorithm 1. ... For our experiments, we set plower = 3.5 and pupper = 3.5. We use a grid resolution of ϵ = 0.1.