用可追溯的反应记忆库，提升化学反应条件推荐的准确性和可解释性。
HiRes: Inspectable Precedent Memory for Reaction Condition Recommendation
- 构建分层反应表示，融合图编码与检索机制，实现条件推荐与依据追溯一体化。
- 催化剂、溶剂、试剂的准确率分别达92.9%、53.4%、53.0%，优于现有模型。
- 适合需要可解释性化学合成规划的研究者和工业研发人员。
反应条件推荐紧随逆合成断键选择之后，实践中化学家既需精准预测，也需可解释的先例支持。我们提出HiRes（分层反应表示），一种基于检索增强的条件推荐系统，其学习到的反应空间兼具分类特征与可追溯的先例记忆功能。模型结合图编码器、变换感知交叉注意力、多流反应融合与k-NN检索层。HiRes在主流USPTO-Condition数据集上达到领先性能，催化剂、溶剂、试剂的Top-1准确率（Acc@1）分别为0.929、0.534、0.530。在催化剂上与最佳基线持平，在溶剂与试剂上优于REACON等模型。配对自助分析表明，融合检索与学习式条件头相比纯参数方法，在溶剂与试剂选择上具有统计显著提升。最终，HiRes弥合了预测准确性与化学可解释性之间的差距，提供单一表示同时实现高性能推荐与具体化学先例支持。
原文摘要 · Abstract (English)
Reaction condition recommendation sits immediately after retrosynthetic disconnection selection, and in practice, chemists require both accurate predictions and the precedents that justify them. We present HiRes (Hierarchical Reaction Representations), a retrieval-augmented condition recommendation system whose learned reaction space serves as both a classifier feature and an inspectable precedent memory. The model combines a graph encoder, transformation-aware cross-attention, multi-stream reaction fusion, and a k-NN retrieval layer. HiRes achieves state-of-the-art performance among primary-slot USPTO-Condition models, reaching Catalyst, Solvent, and Reagent top-1 accuracies (Acc@1) of 0.929, 0.534, and 0.530 respectively. It ties the best reported baseline on Catalyst while outperforming models such as REACON on Solvent and Reagent. Furthermore, paired bootstrap analysis demonstrates that integrating retrieval with learned condition heads provides statistically significant gains for solvent and reagent selection over purely parametric approaches. Ultimately, HiRes bridges the gap between predictive accuracy and chemical interpretability, offering a single representation that supplies both competitive recommendations and the concrete chemical precedents necessary for practical synthesis planning.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。