[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-120677-en":3,"doc-seo-120677-105":29,"detail-sidebar-cat-0-en-105":90},{"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":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":11,"language":21,"language_code":22,"site_id":23,"html_lang":22,"table_of_contents":24,"faqs":25,"seo_title":26,"seo_description":14,"update_tm":27,"read_time":28},120677,1099513958762,"Logic","https://ap-avatar.wpscdn.com/avatar/1000023916a998db790?x-image-process=image/resize,m_fixed,w_180,h_180&k=1784791008015729253",6,"Technology","SIMBAML - Connecting Mechanistic Models and Machine Learning with Augmented Data","Training advanced machine learning models often requires large datasets that are costly and difficult to obtain. When system dynamics prior knowledge is available, mechanistic representations can supplement real measurements. SimbaML (Simulation-Based ML) is an open-source tool that unifies realistic synthetic dataset generation from ODE-based models with direct analysis and integration into ML pipelines. It supports transfer learning from synthetic to real data, data augmentation, identifying data-collection needs, and benchmarking physics-informed ML approaches.","arXiv :2304 .04000v 1 [ cs .LG] 8 Apr 2023  \nSIMBAML: CONNECTING MECHANISTIC MODELS AND MACHINE LEARNING WITH AUGMENTED DATA  \nMaximilian Kleissl 1 ;􀀃 , Lukas Drews 1 ;􀀃 , Benedict B. Heyder 1 ;􀀃 ;†, Julian Zabbarov 1 ;􀀃 ffirstname.lastnameg@student .hpi .de  \n†bjoern .heyder@student .hpi .de  \nPascal Iversen 1 , Simon Witzke 1 , Bernhard Y. Renard 1 ;2 , Katharina Baum 1 ;3 ;4 [f](ffirstname.lastnameg@hpi.de)[firstname.lastname](ffirstname.lastnameg@hpi.de)[g](ffirstname.lastnameg@hpi.de)[@hpi.de](ffirstname.lastnameg@hpi.de)  \n1Hasso Plattner Institute, Digital Engineering Faculty, University of Potsdam,  \n14482 Potsdam, Germany  \n2Department of Mathematics and Computer Science, Free University Berlin,  \n14195 Berlin, Germany  \n3Windreich Department of Artificial Intelligence and Human Health, Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA  \n4Hasso Plattner Institute for Digital Health at Mount Sinai, Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA  \n􀀃 These authors contributed equally.  \nABSTRACT  \nTraining sophisticated machine learning (ML) models requires large datasets that are difficult or expensive to collect for many applications. If prior knowledge about system dynamics is available, mechanistic representations can be used to supplement real-world data. We present SimbaML (Simulation-Based ML), an open-source tool that unifies realistic synthetic dataset generation from ordinary differential equation-based models and the direct analysis and inclusion in ML pipelines. SimbaML conveniently enables investigating transfer learning from synthetic to real-world data, data augmentation, identifying needs for data collection, and benchmarking physics-informed ML approaches. SimbaML is available from [https://pypi.org/project/simba-ml/](https://pypi.org/project/simba-ml/) .  \n1 INTRODUCTION AND RELATED WORK  \nThe success of machine learning (ML) models highly depends on the quality and quantity of available data. However, collecting real-world data is costly and time-consuming. Recent advances in generative models that produce synthetic data, such as generative adversarial networks (Goodfellow et al., 2020), variational autoencoders (Wan et al., 2017), and diffusion models (Sohl-Dickstein et al., 2015), do not fully alleviate this issue as these models need to be trained on large data corpora themselves and cannot extrapolate out-of-distribution. Scientific communities have commonly developed a wealth of domain knowledge that should be leveraged (Baker et al., 2018; Alber et al., 2019) . Detailed prior knowledge of interactions between modeled entities can be represented by mechanistic models, such as ordinary differential equations (ODEs) . They have been used to simulate the dynamical behavior of various systems (Herty et al., 2007; Harush & Barzel, 2017; Hass et al., 2019; Baum et al., 2019) . Consequently, they allow physics-informed learning via observational bias (Karniadakis et al., 2021) and generating datasets for ML benchmarks. Multiple frameworks combining data generation from mechanistic models with ML have recently been proposed (Otness et al., 2021; Takamoto et al., 2022; Hoffmann et al., 2021), see Table 1 in the supplement for details. However, they do not allow for generating realistic data by simulating measurement errors or missing data. In addition, they are not designed for easy extension to other mechanistic models, mainly focus on benchmarking ML models, and commonly do not provide transfer learning functionalities.  \nA  \nB C  \nFigure 1: (A) Overview of the SimbaML framework. (B) Performance of various ML models for different synthetic dataset sizes simulated using a biochemical pathway model. SimbaML conveniently allows end-to-end evaluation of required dataset properties, given prior knowledge of the system dynamics. (C) Comparison of two Covid-19 7-day forecasts from March 13, 2020, using a probabilistic neural network augmented with synthetic data generated by SimbaML","cbCaip7oZfKNQHEO","https://ap.wps.com/l/cbCaip7oZfKNQHEO","pdf",1030635,1,"English","en",105,"# Introduction and Related Work\n# SIMBAML: Features and Use Cases","[{\"question\":\"Why does SimbaML focus on augmenting real-world data with mechanistic models?\",\"answer\":\"Collecting real-world data is costly and time-consuming, and many generative models still struggle out of distribution. SimbaML uses mechanistic ODE knowledge to generate realistic synthetic data that complements real observations in ML pipelines.\"},{\"question\":\"What kinds of synthetic data does SimbaML generate from mechanistic models?\",\"answer\":\"SimbaML solves user-defined ODE systems to produce time-series data, then applies realistic effects by adding noise and sparsifying the series by randomly removing time points.\"},{\"question\":\"What ML tasks can SimbaML support beyond data generation?\",\"answer\":\"SimbaML provides customizable end-to-end pipelines for preprocessing, training, and evaluation, enabling transfer learning, benchmarking, and identifying needs for additional data collection.\"}]","SIMBAML - Connecting Mechanistic Models and Machine Learning with Augmented Data | PDF",1785731331,15,{"code":4,"msg":30,"data":31},"ok",{"site_id":23,"language":22,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":27},"simbaml-connecting-mechanistic-models-and-machine-learning-with-augmented-data","",{"@graph":35,"@context":84},[36,53,67],{"@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/technology/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/simbaml-connecting-mechanistic-models-and-machine-learning-with-augmented-data/120677/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":22,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"Why does SimbaML focus on augmenting real-world data with mechanistic models?","Question",{"text":74,"@type":75},"Collecting real-world data is costly and time-consuming, and many generative models still struggle out of distribution. SimbaML uses mechanistic ODE knowledge to generate realistic synthetic data that complements real observations in ML pipelines.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"What kinds of synthetic data does SimbaML generate from mechanistic models?",{"text":79,"@type":75},"SimbaML solves user-defined ODE systems to produce time-series data, then applies realistic effects by adding noise and sparsifying the series by randomly removing time points.",{"name":81,"@type":72,"acceptedAnswer":82},"What ML tasks can SimbaML support beyond data generation?",{"text":83,"@type":75},"SimbaML provides customizable end-to-end pipelines for preprocessing, training, and evaluation, enabling transfer learning, benchmarking, and identifying needs for additional data collection.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":23},{"code":4,"msg":5,"data":91},[92,96,100,104,109,112,117,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":110,"slug":111},50,"technology",{"id":113,"doc_module":4,"doc_module_name":45,"category_name":114,"show_sort_weight":115,"slug":116},7,"Healthcare",40,"healthcare",{"id":118,"doc_module":4,"doc_module_name":45,"category_name":119,"show_sort_weight":120,"slug":121},8,"Research & Report",30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":105,"slug":137},19,"General","general"]