[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82010-en":3,"doc-seo-82010-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},82010,687197207639,"Asher","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","Accurate, Interdisciplinary and Transparent Structure-property Understanding with Deep Native Structural Reasoning","Structure–property relationships underpin biology, chemistry, and materials science, but mechanistic explanation demands reasoning from structural evidence under physical constraints. SciReasoner is a multimodal scientific foundation model that supports native structure-aware reasoning across proteins, small molecules, and inorganic crystals. It discretizes coordinates, topologies, and periodic connectivities into a unified structure vocabulary, enabling addressable structural tokens as evidence in autoregressive trajectories. Evaluation across 86 benchmarks shows state-of-the-art results on 67 tasks, with improved accuracy and expert preference in 98% of double-blind cases.","arXiv :2607 .07708v 1 [ cs .CL] 8 Jul 2026  \nAccurate, Interdisciplinary and Transparent Structure-property Understanding with Deep Native Structural Reasoning  \nChen Tang 1,2†, Yizhou Wang 1,2†, Jianyu Wu 1,3†, Lintao Wang 1,5†, Shixiang Tang2,1 ,  \nPengze Li 1,4 , Encheng Su 1,8 , Jun Yao 1,8 , Jiabei Xiao 1,2 , Yuqi Shi9,10 , Jielan Li 1 , Hongxia Hao 1 , Zhangyang Gao 1 , Fang Wu 11 , Ben Fei 1,2 , Xiangyu Yue2 , Pan Tan 1 , Bozitao Zhong3 , Jinouwen Zhang 1 , Aoran Wang 1 , Yan Lu2,1 , Jiaheng Liu6,1 , Xinzhu Ma 1 , Liang Hong3 , Mingyue Zheng9,10 , Phil Torr7 , Bowen Zhou 1 , Wanli Ouyang 1,2 , Lei Bai 1  \n1 Shanghai Artificial Intelligence Laboratory, China.  \n2 The Chinese University of Hong Kong, Hong Kong.  \n3 Shanghai Jiao Tong University, China.  \n4 Fudan University, China.  \n5 University of Sydney, Australia.  \n6 Nanjing University, China.  \n7 University of Oxford, UK.  \n8 The University of Science and Technology of China, China.  \n9 Drug Discovery and Design Center, State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, China.  \n10 University of Chinese Academy of Sciences, China.  \n11 Stanford University, USA.  \nCorresponding authors: [shixiangtang@cuhk.edu.hk](shixiangtang@cuhk.edu.hk) ; [ouyangwanli@pjlab.org.cn](ouyangwanli@pjlab.org.cn) ;  \n[bailei@pjlab.org.cn](bailei@pjlab.org.cn) ;  \n†These authors contributed equally to this work.  \n􀂀 [SciReasoner.github.io](SciReasoner.github.io)  \nAbstract  \nStructure–property relationships are foundational to biology, chemistry and materials science, where function, reactivity and physical response emerge from spatial, chemical and periodic organization. Mechanistically explaining these relationships requires interpreting structural evidence through scientific principles and physical constraints, from stereochemistry and bonding to symmetry, energetics and periodic order. However, applying artificial intelligence to this process presents a joint challenge of representation and reasoning: models must preserve domain-native structural information while showing how specific evidence supports predictions under these constraints. Here we introduce SciReasoner, a multimodal scientific foundation model for native structural reasoning across proteins, small molecules and inorganic crystals. SciReasoner discretizes coordinates, topologies and periodic connectivities into a unified structure-aware vocabulary, treating structural tokens as addressable evidence units within autoregressive reasoning trajectories. We evaluate SciReasoner in settings where shortcut correlations are weakened and structure-grounded inference is essential. In homology-controlled Gene Ontology prediction, SciReasoner improves Cellular Component annotation for low-homology and orphan-like proteins, increasing Fmax from 0.42 to 0.55 . In chemistry, it raises single-step retrosynthesis accuracy from 0.63 to 0.72 while generating fragment-level disconnection and precursor-verification traces. In materials science, its representations separate elemental and compound phases and resolve high- and low-band-gap regimes. Across 86 benchmarks, SciReasoner achieves state-of-the-art performance on 67 tasks. Double-blind expert evaluation rates its reasoning traces as preferred or at least comparable to those of a frontier large language model in 98% of cases. By making structure an inspectable substrate for reasoning under scientific constraints, SciReasoner connects accurate prediction with interpretable scientific inference.  \n1 Introduction  \nStructure–property relationships [1–6] are foundational to the physical and biological sciences. Across proteins, small molecules and crystalline materials, observable functions and properties arise from the spatial, chemical and periodic organization of matter. In proteins, protein structures and conformations, long-range interactions and active-site geometry shape their biological functions [7, 8]; in chemicals, bonding, f","cbCaimP2Hu3P2oVM","https://ap.wps.com/l/cbCaimP2Hu3P2oVM","pdf",11540845,1,42,"English","en",105,"# Introduction\n## Structure–property relationships and the need for mechanistic reasoning\n## Challenges for AI representation and evidence-grounded reasoning","[{\"question\":\"What problem does SciReasoner address in structure–property understanding?\",\"answer\":\"It addresses the challenge of explaining structure–property relationships by using structural evidence under scientific and physical constraints, rather than relying on text-based correlations.\"},{\"question\":\"How does SciReasoner represent structural information for reasoning?\",\"answer\":\"It discretizes coordinates, topologies, and periodic connectivities into a unified structure-aware vocabulary, treating structural tokens as addressable evidence units in autoregressive reasoning.\"},{\"question\":\"How is SciReasoner evaluated and what improvements are reported?\",\"answer\":\"It is evaluated on 86 benchmarks, achieving state-of-the-art performance on 67 tasks. Reported gains include higher Fmax in homology-controlled Gene Ontology prediction, improved retrosynthesis accuracy, and better resolution of material band-gap regimes, with expert evaluation preferring or matching frontier LLM reasoning traces in 98% of cases.\"}]",1784177558,106,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"accurate-interdisciplinary-and-transparent-structure-property-understanding-with-deep-native-structural-reasoning","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/accurate-interdisciplinary-and-transparent-structure-property-understanding-with-deep-native-structural-reasoning/82010/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does SciReasoner address in structure–property understanding?","Question",{"text":75,"@type":76},"It addresses the challenge of explaining structure–property relationships by using structural evidence under scientific and physical constraints, rather than relying on text-based correlations.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does SciReasoner represent structural information for reasoning?",{"text":80,"@type":76},"It discretizes coordinates, topologies, and periodic connectivities into a unified structure-aware vocabulary, treating structural tokens as addressable evidence units in autoregressive reasoning.",{"name":82,"@type":73,"acceptedAnswer":83},"How is SciReasoner evaluated and what improvements are reported?",{"text":84,"@type":76},"It is evaluated on 86 benchmarks, achieving state-of-the-art performance on 67 tasks. Reported gains include higher Fmax in homology-controlled Gene Ontology prediction, improved retrosynthesis accuracy, and better resolution of material band-gap regimes, with expert evaluation preferring or matching frontier LLM reasoning traces in 98% of cases.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]