College classes are becoming increasingly large.A critical component in scaling class size is the collaboration and interactions among instructors,teaching assistants,and students.We develop a prototype of an intellig...College classes are becoming increasingly large.A critical component in scaling class size is the collaboration and interactions among instructors,teaching assistants,and students.We develop a prototype of an intelligent voice instructorassistant system for supporting large classes,in which Amazon Web Services,Alexa Voice Services,and self-developed services are used.It uses a scraping service for reading the questions and answers from the past and current course discussion boards,organizes the questions in JavaScript object notation format,and stores them in the database,which can be accessed by Amazon web services Alexa skills.When a voice question from a student comes,Alexa is used for translating the voice sentence into texts.Then,Siamese deep long short-term memory model is introduced to calculate the similarity between the question asked and the questions in the database to find the best-matched answer.Questions with no match will be sent to the instructor,and instructor’s answer will be added into the database.Experiments show that the implemented model achieves promising results that can lead to a practical system.Intelligent voice instructor-assistant system starts with a small set of questions.It can grow through learning and improving when more and more questions are asked and answered.展开更多
Ground penetrating radar（GPR） can be used to image fractures and monitor fluid flow in the subsurface. Conventional GPR imaging uses single-polarization, co-polarized acquisition. We examine the use of cross-polariz...Ground penetrating radar（GPR） can be used to image fractures and monitor fluid flow in the subsurface. Conventional GPR imaging uses single-polarization, co-polarized acquisition. We examine the use of cross-polarized GPR signals for imaging flow channeling in a discrete horizontal fracture. Numerical modeling（FDTD） demonstrates that when the fracture channel is oriented at an oblique angle to the survey line, depolarization of the GPR signal results in scattered energy along the cross-polarized components. When the channel is oriented parallel or orthogonal to the survey line, all scattered energy is captured by the co-polarized components and no signal is present in the cross-polarized orientation. Multipolarization, time-lapse 3D GPR field data were acquired at the Altona Flat Rock test site in New York State. The GPR surveys were conducted during background fresh fracture water conditions and during a natural gradient saline tracer test which was used to highlight flow channels along a sub-horizontal fracture. Amplitude analysis of the cross-polarized data reveals flow channeling that is in agreement with the co-polarized GPR images and with independent hydraulic tests. This investigation demonstrates that cross-polarized components of GPR signals can be used to enhance imaging of flow channels in fractured media.展开更多
基金The authors wish to thank their colleagues and students who were involved in this study and provided valuable implementation and technical support.The research is partly supported by general funding at IoT and Robotics Education Lab and FURI program at Arizona State University and is partly supported by China Scholarship Council,Guangdong Science and Technology Department,under Grant Number 2016A010101020,2016A010101021,and 2016A010101022Guangzhou Science and Information Bureau under Grant Number 201802010033.
文摘College classes are becoming increasingly large.A critical component in scaling class size is the collaboration and interactions among instructors,teaching assistants,and students.We develop a prototype of an intelligent voice instructorassistant system for supporting large classes,in which Amazon Web Services,Alexa Voice Services,and self-developed services are used.It uses a scraping service for reading the questions and answers from the past and current course discussion boards,organizes the questions in JavaScript object notation format,and stores them in the database,which can be accessed by Amazon web services Alexa skills.When a voice question from a student comes,Alexa is used for translating the voice sentence into texts.Then,Siamese deep long short-term memory model is introduced to calculate the similarity between the question asked and the questions in the database to find the best-matched answer.Questions with no match will be sent to the instructor,and instructor’s answer will be added into the database.Experiments show that the implemented model achieves promising results that can lead to a practical system.Intelligent voice instructor-assistant system starts with a small set of questions.It can grow through learning and improving when more and more questions are asked and answered.
基金supported by the US Department of Energy Environmental Remediation Sciences Program (No. 07-358509)the US Department of Energy Geothermal Technologies Program (No. DE-EE0002767)GPR instrumentation was funded by the National Science Foundation (No. EAR/IF-0345445)
文摘Ground penetrating radar（GPR） can be used to image fractures and monitor fluid flow in the subsurface. Conventional GPR imaging uses single-polarization, co-polarized acquisition. We examine the use of cross-polarized GPR signals for imaging flow channeling in a discrete horizontal fracture. Numerical modeling（FDTD） demonstrates that when the fracture channel is oriented at an oblique angle to the survey line, depolarization of the GPR signal results in scattered energy along the cross-polarized components. When the channel is oriented parallel or orthogonal to the survey line, all scattered energy is captured by the co-polarized components and no signal is present in the cross-polarized orientation. Multipolarization, time-lapse 3D GPR field data were acquired at the Altona Flat Rock test site in New York State. The GPR surveys were conducted during background fresh fracture water conditions and during a natural gradient saline tracer test which was used to highlight flow channels along a sub-horizontal fracture. Amplitude analysis of the cross-polarized data reveals flow channeling that is in agreement with the co-polarized GPR images and with independent hydraulic tests. This investigation demonstrates that cross-polarized components of GPR signals can be used to enhance imaging of flow channels in fractured media.
