Convex Polytope Trees

Authors: Mohammadreza Armandpour, Ali Sadeghian, Mingyuan Zhou

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We empirically demonstrate the efficiency of CPT over existing state-of-the-art decision trees in several real-world classification and regression tasks from diverse domains.
Researcher Affiliation Academia Mohammadreza Armandpour Department of Statistics Texas A&M University armand@stat.tamu.edu Ali Sadeghian Department of Computer Science University of Florida asadeghian@ufl.edu Mingyuan Zhou Mc Combs School of Business The University of Texas at Austin mingyuan.zhou@mccombs.utexas.edu
Pseudocode Yes Algorithm 1 Stochastic gradient descent training of the tree splitting parameters for classification task.
Open Source Code Yes Our code to reproduce the results is provided at https://github.com/rezaarmand/Convex_Polytope_Trees.
Open Datasets Yes For multi-class classification, we perform experiments on four benchmark datasets from Lib SVM [10], including MNIST, Connect4, Sens IT, and Letter. We employ the provided training, validation, and testing sets when available; otherwise, we create them under the criterion specified in previous works [35, 20].
Dataset Splits Yes We follow the literature to construct the features [47, 27] and use the same training, validation, and testing split as Lee and Jaakkola [27]. For multi-class classification, we perform experiments on four benchmark datasets from Lib SVM [10], including MNIST, Connect4, Sens IT, and Letter. We employ the provided training, validation, and testing sets when available; otherwise, we create them under the criterion specified in previous works [35, 20].
Hardware Specification No The paper mentions the implementation is in PyTorch and uses Adam optimization but does not specify any hardware details like GPU/CPU models or memory used for experiments in the provided text.
Software Dependencies No The paper states it is implemented in PyTorch and uses Adam optimization but does not provide specific version numbers for these software dependencies.
Experiment Setup Yes A 10-fold cross-validation on the combined train and validation set is used to learn the hyperparameters, namely the maximum number of polytope sides, number of training epochs, learning rate, and batch-size. However, we decide the depth of the tree based on the performance of CPT on the validation set during training, which can be perceived as early stopping for trees. Finally, we report the average and standard error of each method s performance by repeating our experiments for 10 random seed initializations. More details about our implementation and the exact values of hyperparameters for each dataset are presented in the Appendix.