Improving Molecular Design by Stochastic Iterative Target Augmentation
Authors: Kevin Yang, Wengong Jin, Kyle Swanson, Dr.Regina Barzilay, Tommi Jaakkola
ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate significant gains over strong baselines for both unconditional and conditional molecular design. In particular, our approach outperforms the previous state-of-the-art in conditional molecular design by over 10% in absolute gain. Finally, we show that our approach is useful in other domains as well, such as program synthesis. As shown in Table 1, our iterative augmentation paradigm significantly improves the performance of VSeq2Seq and Hier GNN. |
| Researcher Affiliation | Academia | 1UC Berkeley 2MIT 3University of Cambridge. Correspondence to: Kevin Yang <yangk@berkeley.edu>. |
| Pseudocode | Yes | Algorithm 1 Stochastic iterative target augmentation |
| Open Source Code | No | The paper does not include an unambiguous statement that the authors are releasing their code for the work described, nor does it provide a direct link to a source-code repository for their methodology. |
| Open Datasets | Yes | Additionally, we evaluate our augmentation of VSeq2Seq in a transductive setting, as well as in a semi-supervised setting where we provide 100K additional source-side precursors from the ZINC database (Sterling & Irwin, 2015). Our task is based on the educational Karel programming language (Pattis, 1981) used for evaluation in Bunel et al. (2018) and Chen et al. (2019). |
| Dataset Splits | No | The paper mentions "validation set performance" in Figure 5, but it does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) needed to reproduce the data partitioning in the provided text. |
| Hardware Specification | No | The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments. |
| Software Dependencies | No | The paper mentions 'Chemprop package', but does not provide specific ancillary software details, such as library or solver names with version numbers, needed to replicate the experiment. |
| Experiment Setup | Yes | Full hyperparameters are provided in Appendix E.1. |