PAC-Bayes Tree: weighted subtrees with guarantees

Authors: Tin D. Nguyen, Samory Kpotufe

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We present experiments on real-world datasets, for two common partition-tree approaches, dyadic trees and KD-trees. The various datasets are described in Table 1.
Researcher Affiliation Academia Tin Nguyen MIT EECS tdn@mit.edu Samory Kpotufe Princeton University ORFE samory@princeton.edu
Pseudocode Yes Algorithm 1 Bottom-up pass; Algorithm 2 Top-down pass
Open Source Code No The paper does not include any explicit statement about releasing source code or provide a link to a code repository.
Open Datasets Yes Table 1: UCI datasets (Name (abbreviation) Features count Labels count Train size: Spambase, EEG Eye State, Epileptic Seizure Recognition, Crowdsourced Mapping, Wine Quality, Optical Recognition of Handwritten Digits, Letter Recognition).
Dataset Splits Yes Testing data is fixed to be of size 2000, while each experiment is ran 5 times (with random choice of training data of size reported in Table 1) and average performance is reported. In each experiment, all parameters are chosen by 2-fold cross-validation for each of the procedures.
Hardware Specification No The paper does not specify any particular hardware used for running the experiments (e.g., GPU/CPU models, memory, or cloud instance types).
Software Dependencies No The paper does not provide specific software dependencies or version numbers, such as programming language versions or library versions (e.g., PyTorch, TensorFlow, scikit-learn).
Experiment Setup No The paper describes that parameters are chosen by 2-fold cross-validation with specific search grids (log-grid from 2^-8 to 2^6 and linear-grid search), but it does not provide the concrete final hyperparameter values or other system-level training settings (e.g., batch size, optimizer) used for the reported results.