Adapts back translation from NLP to molecular generation, using unlabeled molecules from ZINC to create synthetic training pairs that improve property optimization and retrosynthesis prediction across Transformer and graph-based architectures.
Identifies failure modes in molecular generative models, showing that trivial edits fool distribution-learning benchmarks and that ML-based scoring functions introduce exploitable model-specific and data-specific biases during goal-directed optimization.
Introduces the Frechet ChemNet Distance (FCD), a single metric that captures chemical validity, biological relevance, and diversity of generated molecules by comparing distributions of learned ChemNet representations.
A graph-based genetic algorithm (GB-GA) and a graph-based generative model with Monte Carlo tree search (GB-GM-MCTS) for molecular optimization that match or outperform ML-based generative approaches while being orders of magnitude faster.
GuacaMol provides an open-source benchmarking framework with 5 distribution-learning and 20 goal-directed tasks to standardize evaluation of de novo molecular design models.
MolGenBench introduces a comprehensive benchmark for evaluating molecular generative models in realistic drug discovery settings, spanning de novo design and hit-to-lead optimization across 120 protein targets with 220,005 experimentally validated actives.
MolScore is an open-source framework that unifies scoring functions, evaluation metrics, and benchmarks for generative molecular design, with configurable objectives and GUI support.
This study applies perplexity, a model-intrinsic metric from NLP, to rank de novo molecular designs generated by SMILES-based chemical language models and introduces a delta score to detect pretraining bias in transfer-learned CLMs.
A large-scale benchmark of 25 molecular optimization methods on 23 oracles under constrained oracle budgets, showing that sample efficiency is a critical and often neglected dimension of evaluation.
STONED introduces simple string manipulation algorithms on SELFIES for molecular design, achieving competitive results with deep generative models while requiring no training data or GPU resources.
Introduces TamGen, a target-aware molecular generation method using a pre-trained GPT-like chemical language model with protein structure conditioning. A Design-Refine-Test pipeline discovers 14 inhibitors against tuberculosis ClpP protease, with IC50 values as low as 1.9 uM.
This position paper demonstrates that genetic algorithms (GAs) perform surprisingly well on molecular generation benchmarks, often outperforming complex deep learning methods. The authors propose the GA criterion: new molecule generation algorithms should demonstrate a clear advantage over GAs.