Functional Transparency for Structured Data: a Game-Theoretic Approach

Authors: Guang-He Lee, Wengong Jin, David Alvarez-Melis, Tommi Jaakkola

ICML 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct experiments on chemical and time-series datasets.
Researcher Affiliation Academia 1MIT Computer Science and Artificial Intelligence Laboratory. Correspondence to: Guang-He Lee <guanghe@csail.mit.edu>.
Pseudocode No The paper describes its methodology in narrative text and mathematical formulations but does not include any explicitly labeled 'Pseudocode' or 'Algorithm' blocks.
Open Source Code No The paper does not provide an explicit statement or a link indicating that the source code for the described methodology is publicly available.
Open Datasets Yes We conduct experiments on molecular toxicity prediction on the Tox21 dataset from Molecule Net benchmark (Wu et al., 2018b), which contains 12 binary labels and 7, 831 molecules. and with the bearing dataset from NASA (Lee et al., 2016) and trained on the ZINC dataset (Sterling & Irwin, 2015)
Dataset Splits No The paper mentions total dataset sizes for some datasets (e.g., 7,831 molecules for Tox21, 20K test molecules for ZINC) but does not provide explicit training, validation, and test split percentages or sample counts for all datasets used, nor does it reference predefined splits with citations for all of them.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., GPU models, CPU types, or memory) used for running the experiments.
Software Dependencies No The paper mentions software like TensorFlow (Abadi et al., 2016) and scikit-learn (Pedregosa et al., 2011) but does not provide specific version numbers for the ancillary software dependencies used in the experiments.
Experiment Setup Yes We parametrize µ( ) and Λ( ) jointly by stacking 1 layer of CNN, LSTM, and 2 fully connected layers. We set the neighborhood radius ϵ to 9... The Markov order K is set to 2... and We use GCN as the predictor and decision trees as the witnesses as in 5.2... we set the maximum tree depth as max{ log2(m) 1, 1} for each neighborhood