[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119843-en":3,"doc-seo-119843-105":29,"detail-sidebar-cat-0-en-105":93},{"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":20,"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},119843,16904993612988,"Olivia Brown","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","Machine Learning Prediction of HEA Properties","The research investigates how machine learning can predict phase-related properties of high-entropy alloys (HEAs), which are difficult to study due to complex atomic-scale phase structures and a large, complicated composition space. Traditional CALPHAD-based thermodynamic approaches can model phases but struggle to capture broader structure–composition relationships needed for testing novel alloys. Neural networks and autoencoders are explored to learn these relationships from CALPHAD-derived data, though data limitations require improved modeling. Future work aims to use generative models with latent-space composition encoding for experimentally validated predictions.","| Machine Learning Prediction of HEA Properties\u003Cbr>Nick Beaver, Nathaniel Melisso, Travis Murphy, Mohsen Kivy\u003Cbr>Materials Engineering Department |  |\n| --- | --- |\n|  |  |\n\nResearch Goal Challenges with HEA Research The Machine Learning Approach  \nThe goal of this research project is to study the ability of HEAs are challenging to study for a few key reasons . A promising method to explore HEAs is to use neural nets. Neural  \n~~ machine learning models for the prediction of phase~~s~~ in Common thermodynamic trend~~s~~ and models ~~are~~ misleading nets ~~can~~ approximate ~~any~~ function given enough data and ~~  \nhigh-entropy alloys (HEA) . A model that could reliably predict phases of HEAs could supplant conventional computational approaches to phase prediction.  \nWhat is a High-Entropy Alloy?  \nHigh-entropy alloys are a very new and promising development in the field of materials science. They are less than 20 years old, and come as an extension of the well-studied alloy systems that have been used in engineering for decades. A typical alloy system has one main element as its base, with small amounts of additional elements added in to tune its physical properties in desirable ways (Figure 1a) . HEAs are different in that they are composed of multiple elements in similar amounts, which contributes to their more mixed (entropic) structure (Figure 1b) .  \nFigure 1a, 1b: engineering alloy,(b) a  \npossible for an HEA [1]  \nFigure 1a, 1b: (a) A typical engineering alloy,(b) a  \npossible structure for an HEA [1]  \nThere can be multiple possible atomic-scale structures (phases) in an HEA just like in other alloys, including single-phase structures, multi-phase structures (Figure 2), and intermetallic compounds. The unique phase effects seen in HEAs can give them very useful properties [1] .  \nFigure 2: Illustrative example of a 2-phase HEA  \nwhen applied to these alloys [1] . A result of this is the‘cocktail effect’, where HEAs have a different microstructure (and therefore properties) than expected. This is partly because HEAs represent a much more complicated composition space than regular alloys due to them containing multiple components. Getting enough data to make meaningful models is one of the central challenges in the study of HEAs.  \nComputational Approach  \nAn extremely common method to get data beyond previous experimental results is the use of CALPHAD (calculation of phase diagrams) software (Figure 3) . This computational method uses thermodynamic equations in conjunction with experimental databases to predict the phases present in HEAs. Since the structure and mixture of phases in an alloy are fundamental to its physical properties, having the ability to simulate them based on our current understanding of thermodynamics is critical. However, this fails to describe the larger relationships between structure and composition necessary for making testable predictionson novel alloys.  \nFigure 3: Single-composition phase diagram for TiVZrNbHf  \ntraining time, which in theory makes them ideal for finding the complex relationships between composition and structure present in HEAs. It was hypothesized that there was a relationship between the simple structure and the complex structure that a neural net could learn when given CALPHAD data. However, the relationship proved to be too complex to learn reliably given the amount of data we were able to obtain, so a new method was needed to model the data.  \nFuture Research Figure 4: Example neural net and its code [2]  \nSpecialized machine learning models called autoencoders can be used to model systems with high-dimensional dynamics. They are used lower the dimensionality of these systems such that low-dimensional dynamics can capture the behavior of the more complex original behavior. This research will continue the coming academic year as a materials engineering senior project.  \nFigure 5a, 5b: (a) Latent encoding of Invar alloys,(b)  \nlabelled space with desired property [3] ","cbCaidGLTD9icsa2","https://ap.wps.com/l/cbCaidGLTD9icsa2","pdf",793559,1,"English","en",105,"# Research Goal and Challenges with HEA Research\n## The Machine Learning Approach\n# What Is a High-Entropy Alloy?\n# Computational Approach\n# Future Research\n# Acknowledgements\n# References","[{\"question\":\"Why are high-entropy alloys (HEAs) challenging to study?\",\"answer\":\"HEAs are difficult to study because they can contain multiple atomic-scale phases and have a much more complicated composition space than conventional alloy systems.\"},{\"question\":\"How does the research use machine learning for HEA phase prediction?\",\"answer\":\"The work explores neural networks to learn complex relationships between composition and structure using CALPHAD-generated data, with the expectation that a reliable model could replace slower computational phase-prediction approaches.\"},{\"question\":\"What limitation affected the initial neural-net approach?\",\"answer\":\"The relationship between simple structure inputs and complex HEA structures proved too complex to learn reliably with the amount of available data, motivating a new modeling strategy.\"},{\"question\":\"What directions does future research focus on?\",\"answer\":\"Future research plans to use generative models, encoding the HEA compositional space into a latent space and labeling it with independent properties, then using low-dimensional dynamical models to make predictions to be tested experimentally.\"}]","Machine Learning Prediction of HEA Properties | PDF",1785726604,3,{"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":88,"head_meta":90,"extra_data":92,"updated_unix":27},"machine-learning-prediction-of-hea-properties","",{"@graph":35,"@context":87},[36,52,66],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,49],{"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":28},"https://docshare.wps.com/document/research-report/",{"item":50,"name":13,"@type":42,"position":51},"https://docshare.wps.com/document/machine-learning-prediction-of-hea-properties/119843/",4,{"url":50,"name":13,"@type":53,"author":54,"headline":13,"publisher":56,"fileFormat":59,"inLanguage":22,"description":14,"dateModified":60,"datePublished":60,"encodingFormat":59,"isAccessibleForFree":61,"interactionStatistic":62},"DigitalDocument",{"name":9,"@type":55},"Person",{"url":40,"name":57,"@type":58},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":63,"interactionType":64,"userInteractionCount":4},"InteractionCounter",{"@type":65},"ViewAction",{"@type":67,"mainEntity":68},"FAQPage",[69,75,79,83],{"name":70,"@type":71,"acceptedAnswer":72},"Why are high-entropy alloys (HEAs) challenging to study?","Question",{"text":73,"@type":74},"HEAs are difficult to study because they can contain multiple atomic-scale phases and have a much more complicated composition space than conventional alloy systems.","Answer",{"name":76,"@type":71,"acceptedAnswer":77},"How does the research use machine learning for HEA phase prediction?",{"text":78,"@type":74},"The work explores neural networks to learn complex relationships between composition and structure using CALPHAD-generated data, with the expectation that a reliable model could replace slower computational phase-prediction approaches.",{"name":80,"@type":71,"acceptedAnswer":81},"What limitation affected the initial neural-net approach?",{"text":82,"@type":74},"The relationship between simple structure inputs and complex HEA structures proved too complex to learn reliably with the amount of available data, motivating a new modeling strategy.",{"name":84,"@type":71,"acceptedAnswer":85},"What directions does future research focus on?",{"text":86,"@type":74},"Future research plans to use generative models, encoding the HEA compositional space into a latent space and labeling it with independent properties, then using low-dimensional dynamical models to make predictions to be tested experimentally.","https://schema.org",{"og:url":50,"og:type":89,"og:title":13,"og:site_name":57,"og:description":14},"article",{"robots":91,"canonical":50},"index,follow",{"doc_id":7,"site_id":23},{"code":4,"msg":5,"data":94},[95,99,103,107,112,117,122,125,130,133,137],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":96,"show_sort_weight":97,"slug":98},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":100,"show_sort_weight":101,"slug":102},"Literature",80,"literature",{"id":51,"doc_module":4,"doc_module_name":45,"category_name":104,"show_sort_weight":105,"slug":106},"Exam",70,"exam",{"id":108,"doc_module":4,"doc_module_name":45,"category_name":109,"show_sort_weight":110,"slug":111},5,"Comic",60,"comic",{"id":113,"doc_module":4,"doc_module_name":45,"category_name":114,"show_sort_weight":115,"slug":116},6,"Technology",50,"technology",{"id":118,"doc_module":4,"doc_module_name":45,"category_name":119,"show_sort_weight":120,"slug":121},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":123,"slug":124},30,"research-report",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":128,"slug":129},9,"Religion & Spirituality",20,"religion-spirituality",{"id":128,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":128,"slug":132},"World Cup","world-cup",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":134,"slug":136},10,"Lifestyle","lifestyle",{"id":138,"doc_module":4,"doc_module_name":45,"category_name":139,"show_sort_weight":108,"slug":140},19,"General","general"]