您所在的位置: 首頁?>?新聞動(dòng)態(tài)?>?行業(yè)新聞
受材料基因組計(jì)劃、算法發(fā)展和數(shù)據(jù)驅(qū)動(dòng)的研究在其他領(lǐng)域取得巨大成功的推動(dòng),材料科學(xué)研究中的信息學(xué)方法已逐漸成形。該方法采用機(jī)器學(xué)習(xí)模型,僅依賴已有的數(shù)據(jù)便可快速做出預(yù)測,既不需通過直接的實(shí)驗(yàn),也不需要求解基本方程來計(jì)算/模擬。該方法對于難以用傳統(tǒng)方法測量或計(jì)算的材料性能研究將會十分有效,給材料信息學(xué)添加了機(jī)器學(xué)習(xí)的翅膀。來自美國康涅狄格州立大學(xué)材料科學(xué)與工程系及材料科學(xué)研究所的Rampi Ramprasad教授,綜述了過去十年來基于數(shù)據(jù)驅(qū)動(dòng)的“材料信息學(xué)”的成功策略,特別強(qiáng)調(diào)了材料“指紋”(也稱作“描述符”,可有多種類型和多個(gè)尺度)的選擇。綜述還指出了該領(lǐng)域所面臨的一些挑戰(zhàn)以及近期要克服的困難。該文近期發(fā)表于npj Computational Materials 3:54 (2017); doi:10.1038/s41524-017-0056-5。
原文Abstract及其翻譯
Machine Learning in Materials Informatics: Recent Applications and Prospects(材料信息學(xué)中的機(jī)器學(xué)習(xí):最新應(yīng)用與前景)
Rampi Ramprasad, Rohit Batra, Ghanshyam Pilania, Arun Mannodi-Kanakkithodi & Chiho Kim
Abstract Propelled partly by the Materials Genome Initiative, andpartly by the algorithmic developments and the resounding successes ofdata-driven efforts in other domains, informatics strategies are beginning totake shape within materials science. These approaches lead to surrogate machinelearning models that enable rapid predictions based purely on past data ratherthan by direct experimentation or by computations/simulations in whichfundamental equations are explicitly solved. Data-centric informatics methodsare becoming useful to determine material properties that are hard to measureor compute using traditional methods—due to the cost, time or effortinvolved—but for which reliable data either already exists or can be generatedfor at least a subset of the critical cases. Predictions are typicallyinterpolative, involving fingerprinting a material numerically first, and thenfollowing a mapping (established via a learning algorithm) between thefingerprint and the property of interest. Fingerprints, also referred to as“descriptors”, may be of many types and scales, as dictated by the applicationdomain and needs. Predictions may also be extrapolative—extending into newmaterials spaces—provided prediction uncertainties are properly taken intoaccount. This article attempts to provide an overview of some of the recentsuccessful data-driven “materials informatics” strategies undertaken in thelast decade, with particular emphasis on the fingerprint or descriptor choices.The review also identifies some challenges the community is facing and thosethat should be overcome in the near future.
摘要 受材料基因組計(jì)劃、算法發(fā)展和數(shù)據(jù)驅(qū)動(dòng)的研究在其他領(lǐng)域取得巨大成功的推動(dòng),材料科學(xué)研究中的信息學(xué)方法已逐漸成形。該方法采用機(jī)器學(xué)習(xí)模型,僅依賴已有的數(shù)據(jù)便可快速做出預(yù)測,不需通過直接的實(shí)驗(yàn)以及求解基本方程來計(jì)算/模擬。對于難以用傳統(tǒng)方法測量或計(jì)算的材料性能研究(因受傳統(tǒng)方法的人力物力成本所限),以數(shù)據(jù)為中心的材料信息學(xué)方法會十分有效,前提是已經(jīng)存在相關(guān)材料的可靠數(shù)據(jù)或是可根據(jù)一些關(guān)鍵事例生成出部分密切相關(guān)的數(shù)據(jù)。這些預(yù)測通常是內(nèi)插的(interpolative),即首先從數(shù)值上賦予材料“指紋”,然后通過學(xué)習(xí)算法來建立材料“指紋”與其性能的關(guān)系。“指紋”,也稱作“描述符”,可有多種類型和多個(gè)尺度,可由應(yīng)用領(lǐng)域和需求來決定。在對預(yù)測的不確定性已有充分考慮的前提下,預(yù)測也可以是外推的,即延伸到新的材料空間。本文嘗試對過去十年間基于數(shù)據(jù)驅(qū)動(dòng)的“材料信息學(xué)”成功的策略進(jìn)行綜述,特別強(qiáng)調(diào)了指紋或描述符的選擇。綜述還指出了該領(lǐng)域所面臨的一些挑戰(zhàn)以及近期要克服的困難。
隱石檢測擁有一批在業(yè)內(nèi)取得顯著成就的專業(yè)技術(shù)人員,在行業(yè)內(nèi)有著豐富的檢測經(jīng)驗(yàn)。秉承著專注、專業(yè)、高效、想客戶所想的理念,公司積極增加項(xiàng)目和完善更先進(jìn)的測試儀器設(shè)備,保障每一個(gè)檢測,分析,研發(fā)任務(wù)優(yōu)質(zhì)高效的完成。同時(shí)通過專業(yè)所長,為全球數(shù)萬家優(yōu)質(zhì)客戶提供最及時(shí)的行業(yè)技術(shù)標(biāo)準(zhǔn)信息,和更高精尖的分析檢測解決方案。
隱石檢測分別成立了閥門實(shí)驗(yàn)室,腐蝕實(shí)驗(yàn)室,金相實(shí)驗(yàn)室,力學(xué)實(shí)驗(yàn)室,無損實(shí)驗(yàn)室,耐候老化實(shí)驗(yàn)室。從事常壓儲罐檢測,鍋爐能效檢測,金屬腐蝕檢測,SSC應(yīng)力腐蝕檢測,HIC抗氫致開裂檢測,閥門檢測,應(yīng)力應(yīng)變檢測,無損探傷檢測,機(jī)械設(shè)備檢測,金相分析,石墨烯納米材料檢測,水質(zhì)檢測,油品檢測涉及的服務(wù)范圍已廣泛覆蓋到鋼鐵材料,有色金屬材料,石油化工設(shè)備,通用機(jī)械設(shè)備,冶金礦石,建筑工程材料、航空航天材料,高鐵船舶材料,汽車用零部件、非金屬材料,電子電工產(chǎn)品等各個(gè)領(lǐng)域,并獲得了CMA和CNAS;雙重認(rèn)可。
江蘇省無錫市錫山區(qū)華夏中路3號文華國際
Copyright ? 2017-2024 江蘇隱石實(shí)驗(yàn)科技有限公司 All Rights Reserved 備案號:蘇ICP備2021030923號-2 蘇公網(wǎng)安備32020502001473 技術(shù)支持:迅誠科技