基于小样本数据的晶体合成工艺智能推荐研究OA
Research on An Intelligent Recommendation Model for Crystal Synthesis Procedures Based on Small-Sample Data
[背景]晶体合成是制备新材料的关键环节,但工艺条件复杂多变,实验数据稀缺,工艺确定困难.[问题]如何基于有限的"结构-工艺"数据实现晶体合成工艺的智能生成与可行性评估,是当前亟待解决的问题.[方法]本文基于晶体结构理解增强模型CrysBert,构建了适用于小样本的结构-工艺可行性判别模型;基于晶体结构生成模型CrysGPT,构建了晶体合成工艺生成模型;通过CrysBert和CrysGPT的协同,实现晶体工艺的自动生成与筛选.[结果]基于162条小样本数据训练,工艺判别模型准确率达到0.90,明显优于传统方法;工艺生成模型推荐的候选工艺,经判别模型评价,可行工艺比例达到60.7%,与领域专家推荐工艺的成功率(62.3%)接近.[结论]本研究验证了通过CrysBert和CrysGPT的协同,可为晶体合成工艺的智能设计提供新途径.
[Background]Crystal synthesis is vital for developing new materials but challenging due to complex,variable conditions and limited experimental data.[Problem]Intelligently generating and evaluating crystal synthesis pro-cesses from limited"structure-process"datasets is urgently needed.[Methods]We constructed a feasibility evalu-ation model based on the structure-understanding model(CrysBert)optimized for small datasets,and a generative synthesis model utilizing the large structure-generation model(CrysGPT).Integrating both models enabled auto-matic process generation and screening.[Results]Trained on 162 small-sample data points,the discriminative model achieved an accuracy of 0.90,outperforming traditional methods.The generative model efficiently pro-duced candidate processes,achieving a feasibility rate of 60.7%after discriminative screening,approaching the expert benchmark(62.3%).[Conclusions]This study demonstrates that integrating CrysBert and CrysGPT pro-vides a promising new pathway for intelligent crystal synthesis design.
朱冬;杨小渝;唐述杰;朱锋锋;孔潇;郭艳峰;李兵;秦志鹏
中国科学院计算机网络信息中心,北京 100083中国科学院大学,北京 100049中国科学院计算机网络信息中心,北京 100083中国科学院大学,北京 100049中国科学院大学,北京 100049中国科学院上海微系统与信息技术研究所,上海 200050中国科学院大学,北京 100049中国科学院上海微系统与信息技术研究所,上海 200050
CrysBertCrysGPT工艺智能推荐小样本数据晶体工艺设计
CrysBertCrysGPTcrystal synthesis process intelligent recommendationsmall-sample datacrystal synthesis process design
《数据与计算发展前沿》 2026 (1)
219-231,13
国家自然科学基金面上项目"生成式AI驱动的高分子智能设计方法与技术研究"(62376258)国家自然科学基金联合基金项目"人工智能驱动的海洋防腐涂层材料按需逆向设计"(U24B20126)云南省重点研发计划-材料基因工程项目"高强塑积、轻质高强韧耐磨钢及其大型耐磨件关键成形技术研发(202403AA08001)
评论