[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82286-en":3,"doc-seo-82286-105":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},82286,13056703019662,"Evangeline","https://ap-avatar.wpscdn.com/avatar/be000253a8e92610077?_k=1778726343310543188",8,"Research & Report","Autoregressive Latent Diffusion for 3D Molecule Generation","Three-dimensional (3D) molecule generation is commonly addressed with diffusion models, which deliver high-quality structures but often require the molecular size to be fixed in advance. Autoregressive methods better support variable-length generation and conditioning on partial molecular context, yet balancing unconditional and context-conditioned learning remains difficult. KRONOS is introduced as a latent autoregressive diffusion framework operating in a pretrained autoencoder’s latent space, jointly modeling molecular topology and geometry. Experiments on QM9 and GEOM-Drugs show leading unconditional performance among autoregressive methods and strong fragment-conditioned generation with negligible loss in unconditional quality.","arXiv :2607 .09277v 1 [ cs .LG] 10 Jul 2026  \nAUTOREGRESSIVE LATENT DIFFUSION FOR 3D MOLECULE GENERATION  \nFederico Ottomano, Gaopeng Ren, Kim E. Jelfs, Alex M. Ganose  \nDepartment of Chemistry  \nImperial College London [f.ottomano@imperial.ac.uk](f.ottomano@imperial.ac.uk)  \nYingzhen Li  \nDepartment of Computing Imperial College London  \nABSTRACT  \nThree-dimensional (3D) molecule generation has been dominated by diffusion models, which achieve strong generation quality but typically require the molecular size to be specified a priori. Recent autoregressive approaches have substantially narrowed the performance gap while naturally supporting variable-length generation and conditioning on partial molecular context. However, balancing unconditional and context-conditioned generation remains challenging. We introduce KRONOS, a latent autoregressive diffusion framework that generates molecules in the latent space of a pre-trained autoencoder, jointly modeling molecular graph topology and geometry, while retaining the flexibility of autoregressive generation. We further introduce a mixed training strategy inspired by Fill-in-the Middle (FIM) paradigm, enabling both unconditional and fragment-conditioned molecular generation within a single left-to-right autoregressive model. Experiments on QM9 and GEOM-Drugs demonstrate that KRONOS achieves leading unconditional generation performance among autoregressive methods, while remaining competitive with diffusion models. Moreover, fragment-conditioned generation is achieved with negligible impact on unconditional generation performance, demonstrating that both generation paradigms can be supported within a single architecture.  \n1 INTRODUCTION  \nGenerative modeling of three-dimensional (3D) molecules is a central problem in machine learning for drug discovery and molecular design. Recent progress has largely been driven by diffusion (Hoogeboom et al., 2022b ; Huang et al., 2024 ; Luo et al., 2025 ; Joshi et al., 2025) and flow matching (Irwin et al., 2024), which achieve excellent generation quality but typically require the molecular size to be specified a priori.  \nAutoregressive models provide a complementary paradigm by generating molecules sequentially (Gebauer et al., 2019 ; Luo & Ji, 2022 ; Daigavane et al., 2024), naturally supporting variablelength generation and conditioning on partial molecular structures. These capabilities are particularly valuable in medicinal chemistry, where new molecules are often designed by elaborating known scaffolds or functional fragments. Recent autoregressive models (Cheng et al., 2025 ; Rose et al., 2025) have substantially narrowed the performance gap to diffusion-based approaches, making autoregressive generation an increasingly competitive alternative.  \nAt the same time, unified latent representations (Joshi et al., 2025 ; Park & Walsh, 2025) have emerged as an effective interface for generative modeling by jointly encoding discrete structural information and continuous geometry. Existing latent molecular generators, however, have almost exclusively relied on diffusion-based generation, leaving autoregressive generation largely unexplored.  \nIn this work, we introduce KRONOS, a latent autoregressive diffusion framework for 3D molecule generation. KRONOS performs autoregressive generation directly in the latent space of a pretrained Unified AutoEncoder (UAE) (Luo et al., 2025), modeling the joint distribution over latent molecular tokens autoregressively, while parameterizing each conditional distribution with a diffusion objective (Li et al., 2024) . Furthermore, an explicit stop-prediction mechanism enables variable-length molecular generation without requiring the molecular size in advance. To support fragment-conditioned generation, we further introduce a mixed training strategy inspired by Fillin-the-Middle (FIM) paradigm (Bavarian et al., 2022) enabling both unconditional and fragmentconditioned molecular generation within a single left-t","cbCaipVfPo3QSnvP","https://ap.wps.com/l/cbCaipVfPo3QSnvP","pdf",3815549,1,22,"English","en",105,"# Introduction\n# Related Work","[{\"question\":\"What problem does KRONOS address in 3D molecule generation?\",\"answer\":\"KRONOS targets the difficulty of generating 3D molecules with both unconditional synthesis and fragment-conditioned generation, while avoiding limitations of diffusion models that typically require pre-specifying molecular size.\"},{\"question\":\"How does KRONOS generate molecules?\",\"answer\":\"KRONOS performs autoregressive generation in the latent space of a pretrained unified autoencoder, modeling latent molecular tokens while using a diffusion objective for the conditional latent token predictions.\"},{\"question\":\"How is variable-length and fragment-conditioned generation supported?\",\"answer\":\"Variable-length generation is enabled through an explicit stop-prediction mechanism. Fragment-conditioned generation is enabled with a mixed training strategy inspired by the Fill-in-the-Middle (FIM) paradigm within a single left-to-right autoregressive model.\"}]",1784179399,55,{"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},"autoregressive-latent-diffusion-for-3d-molecule-generation","",{"@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/autoregressive-latent-diffusion-for-3d-molecule-generation/82286/",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 KRONOS address in 3D molecule generation?","Question",{"text":75,"@type":76},"KRONOS targets the difficulty of generating 3D molecules with both unconditional synthesis and fragment-conditioned generation, while avoiding limitations of diffusion models that typically require pre-specifying molecular size.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does KRONOS generate molecules?",{"text":80,"@type":76},"KRONOS performs autoregressive generation in the latent space of a pretrained unified autoencoder, modeling latent molecular tokens while using a diffusion objective for the conditional latent token predictions.",{"name":82,"@type":73,"acceptedAnswer":83},"How is variable-length and fragment-conditioned generation supported?",{"text":84,"@type":76},"Variable-length generation is enabled through an explicit stop-prediction mechanism. Fragment-conditioned generation is enabled with a mixed training strategy inspired by the Fill-in-the-Middle (FIM) paradigm within a single left-to-right autoregressive model.","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"]