Junction Tree Variational Autoencoder for Molecular Graph Generation

Authors: Wengong Jin, Regina Barzilay, Tommi Jaakkola

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

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate our model on multiple tasks ranging from molecular generation to optimization. Across these tasks, our model outperforms previous state-of-the-art baselines by a significant margin.
Researcher Affiliation Academia 1MIT Computer Science & Artificial Intelligence Lab. Correspondence to: Wengong Jin <wengong@csail.mit.edu>.
Pseudocode Yes Algorithm 1 Tree decoding at sampling time
Open Source Code No The paper does not provide an explicit statement or link indicating that the source code for the methodology is openly available.
Open Datasets Yes We use the ZINC molecule dataset from Kusner et al. (2017) for our experiments, with the same training/testing split. It contains about 250K drug molecules extracted from the ZINC database (Sterling & Irwin, 2015).
Dataset Splits No The paper mentions using 'the same training/testing split' as a cited work, but does not explicitly specify the percentages or counts for training, validation, or test sets within the paper itself for the main VAE model. A 10-fold cross-validation is mentioned for a secondary Sparse Gaussian Process, but not for the primary VAE, and no distinct validation set is described for the VAE training.
Hardware Specification No The paper does not provide any specific details about the hardware used to run the experiments, such as GPU models, CPU types, or memory.
Software Dependencies No The paper mentions tools and components like RDKit and GRU, but does not provide specific version numbers for any software dependencies or libraries used in the implementation or experimentation.
Experiment Setup Yes To be comparable with the above baselines, we set the latent space dimension as 56, i.e., the tree and graph representation h T and h G have 28 dimensions each.