让AI理解分子失败原因，实现精准化学设计
Closing the Prior-Posterior Loop: Self-Reflective Molecular Design with Analysis-Driven LLM Iteration
- 用物理原理分析结果替代单一评分反馈，让AI真正理解设计逻辑
- 在2.0-5.0 eV目标下误差低至0.0014 eV，成功率100%
- 适用于多种分子设计任务，适配7种主流大模型
通用大语言模型能否像资深化学家一样精确设计分子？现有基于LLM的框架采用生成-评分-拒绝的单数值反馈循环，本质是受控的试错。本文提出将单一数值反馈升级为第一性原理计算提供的完整理据（如轨道能、原子电荷、电子密度），使模型从随机采样跃变为因果推理者。系统结合检索增强生成与自反思模块，将上述物理量而非压缩评分回流至设计循环。在HOMO-LUMO间隙2.0至5.0 eV的目标下，该结构-性质关系（SPR）反馈配置实现了最低0.0014 eV的偏差，且成功率100%，在中位数与均值偏差上持续优于标量反馈及无反思基线。该框架可无缝扩展至偶极矩设计、合成可及性优化和分子对接任务，并在7种不同大模型主干上表现稳健。结果确立了新范式：当模型不仅知道分子为何失败，更理解其原因时，迭代分子设计才真正具备机制性。
原文摘要 · Abstract (English)
Can a general-purpose large language model design molecules with the precision of a seasoned chemist? Current LLM-based frameworks answer this question with scalar feedback loops - generate, score, reject - that amount to informed trial-and-error. Here we show that replacing a single number with the full physicochemical rationale from first-principles calculations transforms the LLM from a stochastic sampler into a causal reasoner. Our system couples retrieval-augmented generation with a self-reflection module that feeds orbital energies, atomic charges, and electron densities - rather than compressed scores - back into the design loop. On HOMO-LUMO gap targets from 2.0 to 5.0 eV, this structure-property-relationship (SPR) reflection achieves a deviation as low as 0.0014 eV with a 100% success rate under the SPR+RAG configuration, consistently outperforming scalar-feedback and non-reflective baselines in median and mean deviation. The framework generalizes seamlessly to dipole-moment design, synthetic accessibility optimization, and molecular docking, and proves robust across 7 distinct LLM backbones. These results establish a new paradigm: when the model understands not only that a molecule fails, but why, iterative molecular design becomes genuinely mechanistic.
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。