mlsanityData to 14 Oct 2026

Gotta be SAFE: A New Framework for Molecular Design

arXiv 2310.10773 ↗First posted 16 Oct 2023cs.LG q-bio.BMcited by 48

Not on the map. Why?

Abstract (arXiv)

Traditional molecular string representations, such as SMILES, often pose challenges for AI-driven molecular design due to their non-sequential depiction of molecular substructures. To address this issue, we introduce Sequential Attachment-based Fragment Embedding (SAFE), a novel line notation for chemical structures. SAFE reimagines SMILES strings as an unordered sequence of interconnected fragment blocks while maintaining compatibility with existing SMILES parsers. It streamlines complex generative tasks, including scaffold decoration, fragment linking, polymer generation, and scaffold hopping, while facilitating autoregressive generation for fragment-constrained design, thereby eliminating the need for intricate decoding or graph-based models. We demonstrate the effectiveness of SAFE by training an 87-million-parameter GPT2-like model on a dataset containing 1.1 billion SAFE representations. Through targeted experimentation, we show that our SAFE-GPT model exhibits versatile and robust optimization performance. SAFE opens up new avenues for the rapid exploration of chemical space under various constraints, promising breakthroughs in AI-driven molecular design.

Key facts. Cited by 48 papers in the map (data through 14 Oct 2026). Builds on 16 of its 42 references that are in the map. First posted on arXiv 16 Oct 2023.

Builds on

16 of 42 references.

References · 16

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2017Proximal Policy Optimization Algorithmscited by 15,215
  3. 2016Automatic chemical design using a data-driven continuous representation of moleculescited by 765
  4. 2018Junction Tree Variational Autoencoder for Molecular Graph Generationcited by 652
  5. 2018Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation Modelscited by 275
  6. 2021Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Developmentcited by 264
  7. 2018Learning Deep Generative Models of Graphscited by 229
  8. 2018Molecular Transformer - A Model for Uncertainty-Calibrated Chemical Reaction Predictioncited by 229
  9. 2018Fast Lexically Constrained Decoding with Dynamic Beam Allocation for Neural Machine Translationcited by 182
  10. 2018Multi-Objective De Novo Drug Design with Conditional Graph Generative Modelcited by 101
  11. 2022SELFIES and the future of molecular string representationscited by 76
  12. 2021Learning to Extend Molecular Scaffolds with Structural Motifscited by 71
4 more
  1. 2022MolGenSurvey: A Systematic Survey in Machine Learning Models for Molecule Designcited by 66
  2. 2022Group SELFIES: A Robust Fragment-Based Molecular String Representationcited by 33
  3. 2020Scaffold-constrained molecular generationcited by 18
  4. 2022Probabilistic Generative Transformer Language models for Generative Design of Moleculescited by 8

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 10; 2025: 11; 2026: 27.

48 papers.

Sort:

By area:

  1. 2025GenMol: A Drug Discovery Generalist with Discrete Diffusioncited by 50
  2. 2024Molecule Generation with Fragment Retrieval Augmentationcited by 23
  3. 2024ACEGEN: Reinforcement learning of generative chemical agents for drug discoverycited by 14
  4. 2024Less can be more for predicting properties with large language modelscited by 11
  5. 2025mCLM: A Modular Chemical Language Model that Generates Functional and Makeable Moleculescited by 8
  6. 2025FragFM: Hierarchical Framework for Efficient Molecule Generation via Fragment-Level Discrete Flow Matchingcited by 7
  7. 2025Virtual Cells: Predict, Explain, Discovercited by 7
  8. 2024Zero Shot Molecular Generation via Similarity Kernelscited by 6
  9. 2025Refine Drugs, Don't Complete Them: Uniform-Source Discrete Flows for Fragment-Based Drug Discoverycited by 5
  10. 2025FROGENT: An End-to-End Full-process Drug Design Multi-Agent Systemcited by 4
  11. 2026FRIGID: Scaling Diffusion-Based Molecular Generation from Mass Spectra at Training and Inference Timecited by 4
  12. 2025Toward Closed-loop Molecular Discovery via Language Model, Property Alignment and Strategic Searchcited by 3
36 more
  1. 2024Advancements in Molecular Property Prediction: A Survey of Single and Multimodal Approachescited by 2
  2. 2024SAFE setup for generative molecular designcited by 2
  3. 2026CAGenMol: Condition-Aware Diffusion Language Model for Goal-Directed Molecular Generationcited by 2
  4. 2026A2D2: Fine-Tuning Any-Length Discrete Diffusion for Adaptive Decodingcited by 2
  5. 2024Hierarchical Structure Enhances the Convergence and Generalizability of Linear Molecular Representationcited by 1
  6. 2025MolEditRL: Structure-Preserving Molecular Editing via Discrete Diffusion and Reinforcement Learningcited by 1
  7. 2025Conditional Chemical Language Models are Versatile Tools in Drug Discoverycited by 1
  8. 2026TSSR: Two-Stage Swap-Reward-Driven Reinforcement Learning for Character-Level SMILES Generationcited by 1
  9. 2026Unveiling Scaling Behaviors in Molecular Language Models: Effects of Model Size, Data, and Representationcited by 1
  10. 2026Insertion Based Sequence Generation with Learnable Order Dynamicscited by 1
  11. 2026De novo molecular structure elucidation from mass spectra via flow matchingcited by 1
  12. 2026Augmenting Molecular Language Models with Local n-gram Memorycited by 1
  13. 2024A Hitchhiker's Guide to Deep Chemical Language Processing for Bioactivity Predictioncited by 0
  14. 2024Generative Model for Small Molecules with Latent Space RL Fine-Tuning to Protein Targetscited by 0
  15. 2024T- Hop: A framework for studying the importance path information in molecular graphs for chemical property predictioncited by 0
  16. 2025MODA: A Unified 3D Diffusion Framework for Multi-Task Target-Aware Molecular Generationcited by 0
  17. 2025NovoMolGen: Rethinking Molecular Language Model Pretrainingcited by 0
  18. 2026From Tokens to Blocks: A Block-Diffusion Perspective on Molecular Generationcited by 0
  19. 2026Amortized Molecular Optimization via Group Relative Policy Optimizationcited by 0
  20. 2026SIGMA: Semantic Identifier Grouping for Molecular Autoregressioncited by 0
  21. 2026Generative Chemical Language Models for Energetic Materials Discoverycited by 0
  22. 2026MolDeTox: Evaluating Language Model's Stepwise Fragment Editing for Molecular Detoxificationcited by 0
  23. 2026Provably Learning Diffusion Models under the Manifold Hypothesis: Collapse and Refinecited by 0
  24. 2026Beyond Drug Discovery: The Nanotechnology Molecular Optimization (NMO) Benchmarkcited by 0
  25. 2026On the Design Space of Discrete Diffusion Online Adaptation for Molecular Optimizationcited by 0
  26. 2026MARLIN: De Novo Molecular Structure Elucidation from Tandem Mass Spectra without a Ground-Truth Formulacited by 0
  27. 2026MS-GPT: Rethinking MS/MS De Novo Structure Elucidation as Spectrum-Induced Posterior Querying of a Molecule-Language Modelcited by 0
  28. 2026Oracle-Budgeted Molecular Optimization with Short-Term Graph Memorycited by 0
  29. 2026RxnCLF: Contrastive Transformation-Aware Reaction Foundation Model for Improved Reactivity Predictioncited by 0
  30. 2026Molecular LLM Agents: From Architectural Design to Scientific Autonomycited by 0
  31. 2026Elite-Weighted Supervised Fine-tuning for Goal-Directed Molecular Optimizationcited by 0
  32. 2026Frontier LLMs are effective batch optimizers: Assessing reasoning models in continuous and discrete settingscited by 0
  33. 2026One Sequence, Many Decodings: CAGenMol-2 Recasts Drug Design as Masked Molecular Inferencecited by 0
  34. 2026STAMP as a topology-aware molecular language for fragment-level learning and editingcited by 0
  35. 2026PMO-Dock: Benchmarking Docking, Specificity, and Generalization in Molecular Optimizationcited by 0
  36. 2026SNACK: A Sequential Notation Framework for Probabilistic Graph Generationcited by 0