FREED++: Improving RL Agents for Fragment-Based Molecule Generation by Thorough Reproduction
Not on the map. Why?
Abstract (arXiv)
A rational design of new therapeutic drugs aims to find a molecular structure with desired biological functionality, e.g., an ability to activate or suppress a specific protein via binding to it. Molecular docking is a common technique for evaluating protein-molecule interactions. Recently, Reinforcement Learning (RL) has emerged as a promising approach to generating molecules with the docking score (DS) as a reward. In this work, we reproduce, scrutinize and improve the recent RL model for molecule generation called FREED (arXiv:2110.01219). Extensive evaluation of the proposed method reveals several limitations and challenges despite the outstanding results reported for three target proteins. Our contributions include fixing numerous implementation bugs and simplifying the model while increasing its quality, significantly extending experiments, and conducting an accurate comparison with current state-of-the-art methods for protein-conditioned molecule generation. We show that the resulting fixed model is capable of producing molecules with superior docking scores compared to alternative approaches.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 41 of its 100 references that are in the map. First posted on arXiv 18 Jan 2024.
Builds on
41 of 100 references.
References · 41
- 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,282
- 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,635
- 2013Playing Atari with Deep Reinforcement Learningcited by 3,211
- 2016Categorical Reparameterization with Gumbel-Softmaxcited by 2,708
- 2018Soft Actor-Critic Algorithms and Applicationscited by 1,179
- 2015Prioritized Experience Replaycited by 1,070
- 2017MoleculeNet: A Benchmark for Molecular Machine Learningcited by 1,037
- 2016Automatic chemical design using a data-driven continuous representation of moleculescited by 762
- 2022Equivariant Diffusion for Molecule Generation in 3Dcited by 666
- 2017Reinforcement Learning with Deep Energy-Based Policiescited by 659
- 2018Junction Tree Variational Autoencoder for Molecular Graph Generationcited by 647
- 2018MolGAN: An implicit generative model for small molecular graphscited by 392
29 more
- 2021Flow Network based Generative Models for Non-Iterative Diverse Candidate Generationcited by 384
- 2018Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generationcited by 363
- 2018GuacaMol: Benchmarking Models for De Novo Molecular Designcited by 294
- 2018Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation Modelscited by 270
- 2017Molecular De Novo Design through Deep Reinforcement Learningcited by 264
- 2022Structure-based Drug Design with Equivariant Diffusion Modelscited by 213
- 2018Optimization of Molecules via Deep Reinforcement Learningcited by 199
- 2022Pocket2Mol: Efficient Molecular Sampling Based on 3D Protein Pocketscited by 182
- 20233D Equivariant Diffusion for Target-Aware Molecule Generation and Affinity Predictioncited by 176
- 2019Symmetry-adapted generation of 3d point sets for the targeted discovery of moleculescited by 171
- 2017Deep Reinforcement Learning for De-Novo Drug Designcited by 161
- 2021Charformer: Fast Character Transformers via Gradient-based Subword Tokenizationcited by 158
- 2022Equivariant 3D-Conditional Diffusion Models for Molecular Linker Designcited by 121
- 2022Generating 3D Molecules for Target Protein Bindingcited by 104
- 2022Multi-Objective GFlowNetscited by 100
- 2022SPICE, A Dataset of Drug-like Molecules and Peptides for Training Machine Learning Potentialscited by 80
- 2017Time Limits in Reinforcement Learningcited by 72
- 2021Hit and Lead Discovery with Explorative RL and Fragment-based Molecule Generationcited by 68
- 2018Molecular Hypergraph Grammar with its Application to Molecular Optimizationcited by 56
- 2020Symmetry-Aware Actor-Critic for 3D Molecular Designcited by 55
- 2015Massively Multitask Networks for Drug Discoverycited by 49
- 2015AtomNet: A Deep Convolutional Neural Network for Bioactivity Prediction in Structure-based Drug Discoverycited by 38
- 2021Learning to design drug-like molecules in three-dimensional space using deep generative modelscited by 37
- 2020An Empirical Study of Tokenization Strategies for Various Korean NLP Taskscited by 33
- 2020AMBERT: A Pre-trained Language Model with Multi-Grained Tokenizationcited by 22
- 2023Searching for High-Value Molecules Using Reinforcement Learning and Transformerscited by 18
- 2014Multi-task Neural Networks for QSAR Predictionscited by 17
- 2023CodeBPE: Investigating Subtokenization Options for Large Language Model Pretraining on Source Codecited by 9
- 2021Structure-aware generation of drug-like moleculescited by 7
Built on by
Citing papers per year: 2025: 1.
1 paper.