# Representation-Guided Discrete Molecular Graph Retrosynthesis

arXiv: https://arxiv.org/abs/2605.24428 · HTML: https://mlsanity.com/paper/2605.24428/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 19 of its 36 references that are in the map.
- First posted on arXiv 23 May 2026.

## Builds on (most cited first, 19 of 19)

- [Deep Unsupervised Learning using Nonequilibrium Thermodynamics](https://mlsanity.com/paper/1503.03585/) (2015)
- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [How Powerful are Graph Neural Networks?](https://mlsanity.com/paper/1810.00826/) (2018)
- [Instance Normalization: The Missing Ingredient for Fast Stylization](https://mlsanity.com/paper/1607.08022/) (2016)
- [DiGress: Discrete Denoising diffusion for graph generation](https://mlsanity.com/paper/2209.14734/) (2022)
- [Molecular Transformer - A Model for Uncertainty-Calibrated Chemical Reaction Prediction](https://mlsanity.com/paper/1811.02633/) (2018)
- [Learning to Plan Chemical Syntheses](https://mlsanity.com/paper/1708.04202/) (2017)
- [Retrosynthesis Prediction with Conditional Graph Logic Network](https://mlsanity.com/paper/2001.01408/) (2020)
- [Molecule Edit Graph Attention Network: Modeling Chemical Reactions as Sequences of Graph Edits](https://mlsanity.com/paper/2006.15426/) (2020)
- [A Graph to Graphs Framework for Retrosynthesis Prediction](https://mlsanity.com/paper/2003.12725/) (2020)
- [What matters for Representation Alignment: Global Information or Spatial Structure?](https://mlsanity.com/paper/2512.10794/) (2025)
- [RetroXpert: Decompose Retrosynthesis Prediction like a Chemist](https://mlsanity.com/paper/2011.02893/) (2020)
- [Permutation invariant graph-to-sequence model for template-free retrosynthesis and reaction prediction](https://mlsanity.com/paper/2110.09681/) (2021)
- [Learning Graph Models for Retrosynthesis Prediction](https://mlsanity.com/paper/2006.07038/) (2020)
- [RetroBridge: Modeling Retrosynthesis with Markov Bridges](https://mlsanity.com/paper/2308.16212/) (2023)
- [Relative Molecule Self-Attention Transformer](https://mlsanity.com/paper/2110.05841/) (2021)
- [Equivariant Denoisers Cannot Copy Graphs: Align Your Graph Diffusion Models](https://mlsanity.com/paper/2405.17656/) (2024)
- [RetroDiff: Retrosynthesis as Multi-stage Distribution Interpolation](https://mlsanity.com/paper/2311.14077/) (2023)
- [Bridging the Gap between Chemical Reaction Pretraining and Conditional Molecule Generation with a Unified Model](https://mlsanity.com/paper/2303.06965/) (2023)

## Abstract (arXiv)

Stochastic process-based molecular graph generators have become the state of the art for template-free single-step retrosynthesis. However, these models are typically trained only on product-reactant pairs, thereby acquiring chemistry-relevant representations in an indirect and implicit manner. Meanwhile, recent advances in computer vision demonstrate that offering representation guidance to a generator can effectively distill semantics from pretrained encoders into DiTs, substantially improving both convergence and generation quality. Whether similar gains extend to the retrosynthesis task, and what graph-specific design choices can make them work, remains an open question. To address these questions, we conduct a systematic empirical study over a unified design space spanning teacher molecular representations, endpoint and granularity choices, injection depths in the denoiser, correspondence strategies and guidance scheme. Guided by these considerations, we develop Graph-oriented Representation Guidance (GRG), which achieves 58.6 / 77.2 / 83.4 / 87.1 top-1 / 3 / 5 / 10 accuracy on USPTO-50k, while increasing diversity to 15.5, both substantially outperforming the adopted base generator. Notably, GRG consistently improves all top-k metrics in out-of-distribution settings, suggesting that representation guidance facilitates the acquisition of intrinsic chemical semantics. Meanwhile, the introduced representation guidance reduces the number of epochs by 35% and the wall-clock time by 30% to reach comparable performance. In addition, we introduce a simple yet effective representation-similarity-based reranking mechanism, which further improves the top of the ranked list without training an additional verifier.