Active Retrosynthetic Planning Aware of Route Quality

Authors: Luotian Yuan, Yemin Yu, Ying Wei, Yongwei Wang, Zhihua Wang, Fei Wu

ICLR 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We apply our framework to different existing approaches on both the benchmark and an expert dataset and demonstrate that it outperforms the existing state-of-the-art approach by 6.2% in route quality while reducing the query cost by 12.8%. [...] 4 EXPERIMENTS
Researcher Affiliation Academia Luotian Yuan1 , Yemin Yu2,4 , Ying Wei3 , Yongwei Wang1,4 , Zhihua Wang4, Fei Wu1,4 1Zhejiang University, 2City University of Hong Kong, 3Nanyang Technological University 4Shanghai Institute for Advanced Study of Zhejiang University
Pseudocode Yes Algorithm 1: Training algorithm
Open Source Code No The paper does not provide an explicit statement about open-sourcing its code or a link to a code repository for its own implementation.
Open Datasets Yes We use two test sets to evaluate our methods. The first one is a widely used USPTO-50k benchmark dataset that has 178 hard molecules raised in Chen et al. (2020). [...] Initially, the method in Guo et al. (2020) is employed to pre-train a model utilizing the USPTO-MIT dataset, followed by the fine-tuning of the model in reactions derived from the high-quality, expert-annotated dataset.
Dataset Splits Yes We partition the expert dataset as 0.8/0.1/0.1 into train/valid/test sets.
Hardware Specification No The paper does not explicitly describe the hardware specifications (e.g., GPU models, CPU types, or cloud instance details) used for its experiments.
Software Dependencies No The paper mentions algorithms (e.g., TD3) and models (e.g., MLP, Morgan fingerprint) but does not provide specific version numbers for software dependencies or libraries used.
Experiment Setup Yes As for the hyper-parameters, we set the maximum route depth as 6 and λ in Eq. 1 as 4. [...] Hyperparameters Values Encoder layers 4 Decoder layers 4 Encoder embedding dimension 2048 Encoder FFN embedding dimension 2048 Encoder attention heads 8 Decoder embedding dimension 2048 Decoder FFN embedding dimension 2048 Decoder attention heads 8 Optimizer Adam Learning rate 1e-4 Weight decay 0.0001 N epochs 12 Clip norm 0.25 Dropout rate 0.1