Notice: This website is continuously updated as new conference proceedings and journal volumes are published. Consequently, the figures presented here may differ from those reported in our publications.
A brief overview of our methodology is provided on the Methods page, along with references to publications documenting our approach and links to the code and data used in each study.
The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates.
Average Sensitivity of Decision Tree Learning
Authors: Satoshi Hara, Yuichi Yoshida
ICLR 2023 | Venue PDF | LLM Run Details | Input Tokens: 23,502 Total number of tokens sent to the LLM as input for this paper's analysis. | Output Tokens: 4,130 Total number of tokens produced by the LLM (including reasoning/thinking tokens) for this paper's analysis.
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | 7 EXPERIMENTS We demonstrate that the proposed algorithm can output stable decision trees. Datasets We used datasets shown in Table 1. |
| Researcher Affiliation | Academia | Satoshi Hara Osaka University EMAIL Yuichi Yoshida National Institute of Informatics EMAIL |
| Pseudocode | Yes | Algorithm 1: Procedure PREDICT; Algorithm 2: Procedure DISTANCE; Algorithm 3:; Algorithm 4: |
| Open Source Code | Yes | The code is available at https://github.com/sato9hara/Stable Decision Tree |
| Open Datasets | Yes | These datasets are obtained from https://github.com/chenhongge/RobustTrees |
| Dataset Splits | No | For training, we randomly sampled 80% of the data points and 1000 data points for small and large datasets, respectively. In the experiments, we evaluated the test accuracy of the learned decision trees using the entire test data. While cross-validation is mentioned for tree depth selection, explicit validation dataset splits are not provided. |
| Hardware Specification | No | No specific hardware details (like CPU/GPU models, memory) are mentioned in the paper. |
| Software Dependencies | No | We implemented both the greedy and proposed algorithms in Python 3 using the JIT compiler of Numba. Python 3 is not specific enough (e.g., 3.x), and no version is given for Numba. |
| Experiment Setup | Yes | We set the tree depth shown in Table 1 so that the greedy algorithm exhibits the highest accuracy in cross-validation." and "we set Q = 500." |