|
|
|
题名
|
作者
|
年代
|
出处
|
被引量
|
| 1 | Reliable and explainable machine-learning methods for accelerated material discovery显示文摘Despite ML’s impressive performance in commercial applications,several unique challenges exist when applying ML in materials science applications.In such a context,the contributions of this work are twofold.First,we identify common pitfalls of existing ML techniques when learning from underrepresented/imbalanced material data.Specifically,we show that with imbalanced data,standard methods for assessing quality of ML models break down and lead to misleading conclusions.Furthermore,we find that the model’s own confidence score cannot be trusted and model introspection methods(using simpler models)do not help as they result in loss of predictive performance(reliability-explainability trade-off).Second,to overcome these challenges,we propose a general-purpose explainable and reliable machine-learning framework.Specifically,we propose a generic pipeline that employs an ensemble of simpler models to reliably predict material properties.We also propose a transfer learning technique and show that the performance loss due to models’simplicity can be overcome by exploiting correlations among different material properties.A new evaluation metric and a trust score to better quantify the confidence in the predictions are also proposed.To improve the interpretability,we add a rationale generator component to our framework which provides both model-level and decision-level explanations.Finally,we demonstrate the versatility of our technique on two applications:(1)predicting properties of crystalline compounds and(2)identifying potentially stable solar cell materials.We also point to some outstanding issues yet to be resolved for a successful application of ML in material science. | Bhavya Kailkhura Brian Gallagher Sookyung Kim Anna Hiszpanski T.Yong-Jin Han | 2019 | npj Computational Materials2019，，1： | 3 |
| 2 | A passivewireless multi-sensor SAW technology device and system per-spectives 显示文摘 | Malocha Donald C Gallagher Mark Fisher Brian | 2013 | Sensors2013，13，5： | 1 |
| 3 | A Passive Wireless Multi-Sensor SAW Technology Device and System Per- spectives 显示文摘 | Donald C Malocha Mark Gallagher Brian Fisher | 2013 | Sensors2013，13，5： | 1 |
| 4 | Unusual cone conformation retention in calix[4]arenes显示文摘 | Bernadette S Creaven Mary Deasy John F Gallagher John McGinley Brian A Murray | 2001 | Tetrahedron2001，，42： | 1 |
| 5 | Can peak systolic velocities be used for prediction of stroke in sickle cell anemia?显示文摘 | Anne Jones Suzanne Granger Don Brambilla Dianne Gallagher Elliott Vichinsky Gerald Woods Brian Berman Steve Roach Fenwick Nichols Robert J. Adams | 2005 | Pediatric Radiology2005，，1： | 1 |
| 6 | A Randomized Controlled Trial of Multi-Slice Coronary Computed Tomography for Evaluation of Acute Chest Pain显示文摘 | James A. Goldstein Michael J. Gallagher William W. O’Neill Michael A. Ross Brian J. O’Neil Gilbert L. Raff | 2007 | Journal of the American College of Cardiology2007，，8： | 1 |
| 7 | The economics of producing biodiesel from algae显示文摘 | Brian J Gallagher | | 0，，01： | 1 |
| 8 | Brief report: Hospitalized patients’ attitudes about and participation in error prevention显示文摘 | Dr. Amy D. Waterman PhD Thomas H. Gallagher MD Jane Garbutt MB ChB Brian M. Waterman MPH Victoria Fraser MD Thomas E. Burroughs PhD | 2006 | Journal of General Internal Medicine2006，，4： | 1 |
| 9 | A study of real-world micrograph data quality and machine learning model robustness显示文摘Machine-learning(ML)techniques hold the potential of enabling efficient quantitative micrograph analysis,but the robustness of ML models with respect to real-world micrograph quality variations has not been carefully evaluated.We collected thousands of scanning electron microscopy(SEM)micrographs for molecular solid materials,in which image pixel intensities vary due to both the microstructure content and microscope instrument conditions.We then built ML models to predict the ultimate compressive strength(UCS)of consolidated molecular solids,by encoding micrographs with different image feature descriptors and training a random forest regressor,and by training an end-to-end deep-learning(DL)model.Results show that instrument-induced pixel intensity signals can affect ML model predictions in a consistently negative way.As a remedy,we explored intensity normalization techniques.It is seen that intensity normalization helps to improve micrograph data quality and ML model robustness,but microscope-induced intensity variations can be difficult to eliminate. | Xiaoting Zhong Brian Gallagher Keenan Eves Emily Robertson T.Nathan Mundhenk T.Yong-Jin Han | 2021 | npj Computational Materials2021，，1： | 0 |
| 10 | 灰姑娘显示文摘拍给“90后”、“00后”看的灰姑娘来了!
这部灵感源自经典童话故事的电影阵容强大:去年在奥斯卡斩获影后的凯特·布兰切特饰演恶毒的继母,老戏骨海伦娜·伯翰·卡特饰演仙女教母(海伦娜竟然是演仙女教母) | Brian Gallagher 俞力 | 2015 | 疯狂英语(初中天地)2015，0，3： | 0 |
| 11 | Explainable machine learning in materials science显示文摘Machine learning models are increasingly used in materials studies because of their exceptional accuracy.However,the most accurate machine learning models are usually difficult to explain.Remedies to this problem lie in explainable artificial intelligence(XAI),an emerging research field that addresses the explainability of complicated machine learning models like deep neural networks(DNNs).This article attempts to provide an entry point to XAI for materials scientists.Concepts are defined to clarify what explain means in the context of materials science.Example works are reviewed to show how XAI helps materials science research.Challenges and opportunities are also discussed. | Xiaoting Zhong Brian Gallagher Shusen Liu Bhavya Kailkhura Anna Hiszpanski T.Yong-Jin Han | 2022 | npj Computational Materials2022，，1： | 0 |
| 12 | Effect of genetic sources on anatomical, morphological,and mechanical properties of 14-year-old genetically improved loblolly pine families from two sites in the southern United States显示文摘Tree improvement programs on loblolly pine(Pinus taeda) in the southeastern USA has focused primarily on improving growth, form, and disease tolerance.However, due to the recent reduction of design values for visually graded southern yellow pine lumber(including loblolly pine), attention has been drawn to the material quality of genetically improved loblolly pine. In this study,we used the time-of-flight(TOF) acoustic tool to assess the effect of genetic families on diameter, slenderness, fiber length, microfibril angle(MFA), velocity and dynamic stiffness estimated using green density(DMOEG) and basic density(DMOEB) of 14-year-old loblolly pine stands selected from two sites. All the 184 and 204 trees of the selected eight half-sib genetic families on sites 1 and 2 respectively were tested using TOF acoustic tool, and two 5 mm core samples taken at breast height level(1.3 m)used to for the anatomical and physical properties analysis.The results indicated a significant positive linear relationship between dynamic MOEs(DMOEGand DMOEB)versus tree diameter, slenderness, and fiber length while dynamic MOEs negatively but nonsignificant correlated with MFA. While there was no significant difference in DMOEBbetween sites; velocity 2 for site 1 was significantly higher than site 2 but DMOEGwas higher for site 2 than site 1. Again, the mean DMOEGand DMOEBreported in the present study presents a snapshot of the expected static MOE for green and 12% moisture conditions respectively for loblolly pine. Furthermore, there were significant differences between families for most of the traits measured and this suggests that forest managers have the opportunity to select families that exhibit the desired fiber morphology for final product performance. Lastly,since the dynamic MOE based on green density(DMOEG),basic density(DMOEB) and velocity 2 present difference conclusions, practitioners of this type of acoustic technique should take care when extrapolating results across the sites. | Charles Essien Brian K.Via Gifty Acquah Thomas Gallagher Timothy McDonald Lori Eckhardt | 2018 | Journal of Forestry Research2018，29，6： | 0 |
| 13 | 儿科日间手术中超重(或肥胖症)与胃液特性的关系:在禁食指南和误吸风险评估方面的意义显示文摘背景超重或肥胖的日间手术儿科患者手术前2小时禁水的安全性尚未得到证实。健康的儿童和肥胖的成人在手术前禁水2小时后残留胃液量(GFVs)并不多,因此认为误吸风险并不高。因此我们将测算日间手术患者中超重症或肥胖症的患病率,并假设不管是体重指数还是禁食时间均不会对GFV或胃液pH值产生显著的影响。所有受试儿童均于手术前禁清水2小时,并假设超重或肥胖患儿的GFV不会多于偏瘦或正常体重患儿,而呕吐或误吸的风险也很低。方法本研究中连续纳入了1000例2-12岁行日间手术的患儿,记录人口统计学资料、病史、身高和体重。另外纳入1000例需全麻插管的日间手术患儿(2。12岁)进行研究。气管插管后,经口插入一根14—18F的胃管,抽空胃内容物。用药情况、禁食时间、GFV、pH值和呕吐事件均被详细记录。采用疾病预防及控制中心的生长图表(2000)评定理想体重(IBW=第50个百分位数),并对患儿进行分层:偏瘦或正常体重(BMI在第25—75个百分位数)、超重(第95个百分位数〈BMI≥第85个百分位数)、肥胖(BMI≥第95个百分位数)。结果在所有的日间手术患儿中有14%属于超重,13.3%属于肥胖。肥胖儿按公斤体重算出的GFV较低(P〈0.001)。当我们用IBW进行校正后,所有BMI组中GFV(mw校正后)的容量都相等(均值为0.96ml/kg,SD为0.71;中位数为0.86ml/kg,IQR为0.96)。手术前使用对乙酰氨基酚和咪达唑仑均导致GFV(1BW校正后)增加(P=0.025和P=0.001)。而ASA11I级(P=0.024)、男性(P=0.012)、胃食管反流性疾病(P=0.049)和使用质子泵抑制剂(P=0.018)患儿的GFV(mw校正后)较低。GFV(IBw校正后)与禁食时间和年龄无关。胃液酸度较低与年龄较小(P=0.005)、BMI较高(P=0.036)和美国非洲裔(P=0.033)患儿等因素相关。有8例患儿在全麻诱导时发生了呕吐(其中50%有肥胖症,P=0.052,75%患有阻塞性睡眠呼吸暂停症,P=0.061)。呕吐与ASA分级较高有关,但却与禁食时间的长短无关。研究中没有出现误吸事件。结论27%的儿科日闻手术患儿属超重或肥胖儿童。无论进食时间长短或BMI如何,GFV(IBw校正后)均为1ml/kg,因此这些患儿均可以在手术前2小时饮清水。在此研究群体中,罕见的呕吐事件与禁食时间的缩短并无相关性。 | Scott D. Cook-Sather Paul R. Gallagher Lydia E. Kruge Jonathan M. Beus Brian P. Ciampa Kevin Conor Welch Sina Shah-Hosseini Jieun S. Choi Reshma Pachikara Kim Minger Ronald S. Litman Mark S. Schreiner 周洁(译) 李士通(校) | 2010 | 麻醉与镇痛2010，，6： | 0 |
您所在的机构暂未开通此部分的功能，请联系管理员