Realizable $H$-Consistent and Bayes-Consistent Loss Functions for Learning to Defer

Authors: Anqi Mao, Mehryar Mohri, Yutao Zhong

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

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we empirically evaluate our proposed surrogate losses and compare them with existing baselines. In this section, we empirically evaluate our proposed surrogate losses and compare them with existing baselines.
Researcher Affiliation Collaboration Anqi Mao Courant Institute New York, NY 10012 aqmao@cims.nyu.edu Mehryar Mohri Google Research & CIMS New York, NY 10011 mohri@google.com Yutao Zhong Courant Institute New York, NY 10012 yutao@cims.nyu.edu
Pseudocode No The paper does not contain pseudocode or clearly labeled algorithm blocks.
Open Source Code No The paper does not provide an explicit statement or link indicating that the source code for the methodology described is open-source or publicly available.
Open Datasets Yes We follow the setting of Mozannar et al. [2023] and conduct experiments on a synthetic dataset: Mixture-of-Gaussians [Mozannar et al., 2023], and three real-world datasets: CIFAR-10H [Battleday et al., 2020], Hate Speech [Davidson et al., 2017], and COMPASS [Dressel and Farid, 2018].
Dataset Splits Yes Each dataset is randomly split into 70%, 10%, and 20% for training, validation, and testing, respectively.
Hardware Specification No The paper does not specify the exact hardware (e.g., GPU model, CPU type) used for running the experiments.
Software Dependencies No The paper mentions using specific software packages or frameworks, but it does not provide specific version numbers for these dependencies (e.g., 'PyTorch 1.9' or 'Python 3.8').
Experiment Setup No The paper states 'We use the same optimizer, learning rate, and number of epochs as chosen in [Mozannar et al., 2023]', which refers to an external source for hyperparameters rather than listing them explicitly in the main text of this paper.