mlsanityData to 14 Oct 2026

Language-Informed Flow Matching for Trend-Guided Structure-Based 3D Molecular Generation

arXiv 2608.31009 ↗First posted 31 Aug 2026cs.LGcited by 0

On the map: Diffusion models → Flow matching, time series and science

Abstract (arXiv)

Structure-based drug design (SBDD) requires ligands that satisfy both 3D target affinity and 1D chemical validity. Existing controllable generation methods often rely on task-specific fine-tuning or externally imposed sampling-time guidance, adding cost and potentially conflicting with evolving 3D geometric constraints. We propose LiFT, a language-informed cross-modal framework built on Flow Matching for trend-guided 3D molecular generation across both de novo design and scaffold hopping. LiFT uses a "Sense-Evolve-Assemble" agent to generate target-aware SMILES as intermediate chemical conditions, from which a pre-trained chemical foundation model extracts continuous semantic priors. These priors are integrated into geometric generation through a lightweight semantic projector with zero-initialized adaptive normalization for stable cross-modal conditioning. We further introduce a Self-Conditioned Decoupled Router (SCDR), which modulates the velocity field according to intermediate structural states during ODE integration. Experiments on Cross-Docked2020 show that LiFT achieves competitive distribution matching while improving medicinal chemistry metrics and maintaining competitive structural validity under task-steering settings without additional generator fine-tuning. Our results suggest that language-derived chemical priors provide effective trend-level guidance for 3D molecular generation. Code and released artifacts are available at https://github.com/kasurl/LiFT.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 21 of its 34 references that are in the map. First posted on arXiv 31 Aug 2026.

Builds on

21 of 34 references.

Closest ancestors · 5

Same direction, newest first

  1. 2025MolGuidance: Advanced Guidance Strategies for Conditional Molecular Generation with Flow Matching13 Dec
  2. 2025MolSculpt: Sculpting 3D Molecular Geometries from Chemical Syntax9 Dec
  3. 2025Multi-domain Distribution Learning for De Novo Drug Design25 Aug
  4. 2025NExT-Mol: 3D Diffusion Meets 1D Language Modeling for 3D Molecule Generation18 Feb
  5. 2022Structure-based Drug Design with Equivariant Diffusion Models24 Oct

Other references · 16

  1. 2017FiLM: Visual Reasoning with a General Conditioning Layercited by 2,333
  2. 2017Deep Setscited by 1,421
  3. 2021E(n) Equivariant Graph Neural Networkscited by 965
  4. 2020Learning from Protein Structure with Geometric Vector Perceptronscited by 313
  5. 2022Pocket2Mol: Efficient Molecular Sampling Based on 3D Protein Pocketscited by 183
  6. 20233D Equivariant Diffusion for Target-Aware Molecule Generation and Affinity Predictioncited by 177
  7. 2022A 3D Generative Model for Structure-Based Drug Designcited by 149
  8. 2023MolCA: Molecular Graph-Language Modeling with Cross-Modal Projector and Uni-Modal Adaptercited by 128
  9. 2025Training-Free Guidance Beyond Differentiability: Scalable Path Steering with Tree Search in Diffusion and Flow Modelscited by 24
  10. 2024CBGBench: Fill in the Blank of Protein-Molecule Complex Binding Graphcited by 19
  11. 2025Chem3DLLM: 3D Multimodal Large Language Models for Chemistrycited by 7
  12. 2025Prior-Guided Flow Matching for Target-Aware Molecule Design with Learnable Atom Numbercited by 6
4 more
  1. 2023Geometric Deep Learning for Structure-Based Drug Design: A Surveycited by 2
  2. 2025Controllable 3D Molecular Generation for Structure-Based Drug Design Through Bayesian Flow Networks and Gradient Integrationcited by 2
  3. 2026Empowering LLMs for Structure-Based Drug Design via Exploration-Augmented Latent Inferencecited by 2
  4. 2026Unveiling the Secret of AdaLN-Zero in Diffusion Transformercited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025MODA: A Unified 3D Diffusion Framework for Multi-Task Target-Aware Molecular Generation
  2. 2025Toward Closed-loop Molecular Discovery via Language Model, Property Alignment and Strategic Search
  3. 2025SynCoGen: Synthesizable 3D Molecule Generation via Joint Reaction and Coordinate Modeling
  4. 2026SynLaD: Latent Diffusion for Generating Synthesizable Molecules Conditioned on 3D Pharmacophore Profiles
  5. 2025FlowMol3: Flow Matching for 3D De Novo Small-Molecule Generation
  6. 2026CAGenMol: Condition-Aware Diffusion Language Model for Goal-Directed Molecular Generation
  7. 2025Compositional Flows for 3D Molecule and Synthesis Pathway Co-design
  8. 2025Uncertainty-Aware Multi-Objective Reinforcement Learning-Guided Diffusion Models for 3D De Novo Molecular Design