# A Deep Generative Model for Fragment-Based Molecule Generation

arXiv: https://arxiv.org/abs/2002.12826 · HTML: https://mlsanity.com/paper/2002.12826/

## Key facts

- Cited by 24 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 28 Feb 2020.

## Built on by (most cited first, 24 of 24)

- [3D Equivariant Diffusion for Target-Aware Molecule Generation and Affinity Prediction](https://mlsanity.com/paper/2303.03543/) (2023)
- [A Survey on Deep Graph Generation: Methods and Applications](https://mlsanity.com/paper/2203.06714/) (2022)
- [MEG: Generating Molecular Counterfactual Explanations for Deep Graph Networks](https://mlsanity.com/paper/2104.08060/) (2021)
- [Graph Neural Networks for Molecules](https://mlsanity.com/paper/2209.05582/) (2022)
- [Equivariant Shape-Conditioned Generation of 3D Molecules for Ligand-Based Drug Design](https://mlsanity.com/paper/2210.04893/) (2022)
- [MADGEN: Mass-Spec attends to De Novo Molecular generation](https://mlsanity.com/paper/2501.01950/) (2025)
- [t-SMILES: A Scalable Fragment-based Molecular Representation Framework for De Novo Molecule Generation](https://mlsanity.com/paper/2301.01829/) (2023)
- [Classifier-free graph diffusion for molecular property targeting](https://mlsanity.com/paper/2312.17397/) (2023)
- [LLM-Based Multi-Agent Systems are Scalable Graph Generative Models](https://mlsanity.com/paper/2410.09824/) (2024)
- [GraphGen-Redux: a Fast and Lightweight Recurrent Model for labeled Graph Generation](https://mlsanity.com/paper/2107.08396/) (2021)
- [SynLlama: Generating Synthesizable Molecules and Their Analogs with Large Language Models](https://mlsanity.com/paper/2503.12602/) (2025)
- [Fragment-based Sequential Translation for Molecular Optimization](https://mlsanity.com/paper/2111.01009/) (2021)
- [mathsfG2Retro as a Two-Step Graph Generative Models for Retrosynthesis Prediction](https://mlsanity.com/paper/2206.04882/) (2022)
- [RGCVAE: Relational Graph Conditioned Variational Autoencoder for Molecule Design](https://mlsanity.com/paper/2305.11699/) (2023)
- [Overcoming Order in Autoregressive Graph Generation](https://mlsanity.com/paper/2402.03387/) (2024)
- [Toward Closed-loop Molecular Discovery via Language Model, Property Alignment and Strategic Search](https://mlsanity.com/paper/2512.09566/) (2025)
- [An In-depth Summary of Recent Artificial Intelligence Applications in Drug Design](https://mlsanity.com/paper/2110.05478/) (2021)
- [Bayesian Deep Learning for Graphs](https://mlsanity.com/paper/2202.12348/) (2022)
- [Unsupervised Learning of Graph from Recipes](https://mlsanity.com/paper/2401.12088/) (2024)
- [Generative Model for Small Molecules with Latent Space RL Fine-Tuning to Protein Targets](https://mlsanity.com/paper/2407.13780/) (2024)
- [A Foundation Chemical Language Model for Comprehensive Fragment-Based Drug Discovery](https://mlsanity.com/paper/2509.19586/) (2025)
- [From In Silico to In Vitro: Evaluating Molecule Generative Models for Hit Generation](https://mlsanity.com/paper/2512.22031/) (2025)
- [MolPaQ: Modular Quantum-Classical Patch Learning for Interpretable Molecular Generation](https://mlsanity.com/paper/2604.08575/) (2026)
- [A Systematic Evaluation of Molecule Generation Models for De Novo Drug Design: From Benchmarks to Practical Insights](https://mlsanity.com/paper/2609.10099/) (2026)

## Abstract (arXiv)

Molecule generation is a challenging open problem in cheminformatics. Currently, deep generative approaches addressing the challenge belong to two broad categories, differing in how molecules are represented. One approach encodes molecular graphs as strings of text, and learns their corresponding character-based language model. Another, more expressive, approach operates directly on the molecular graph. In this work, we address two limitations of the former: generation of invalid and duplicate molecules. To improve validity rates, we develop a language model for small molecular substructures called fragments, loosely inspired by the well-known paradigm of Fragment-Based Drug Design. In other words, we generate molecules fragment by fragment, instead of atom by atom. To improve uniqueness rates, we present a frequency-based masking strategy that helps generate molecules with infrequent fragments. We show experimentally that our model largely outperforms other language model-based competitors, reaching state-of-the-art performances typical of graph-based approaches. Moreover, generated molecules display molecular properties similar to those in the training sample, even in absence of explicit task-specific supervision.