斯坦福大学-机器学习与地球科学

研究计划：深度学习与地球科学

课题：理解深度学习算法的基础知识并将其应用于不同的地球科学场景

摘要：

深度学习在包括地球科学在内的各种工业和科学领域中变得越来越重要。例如，它可以用于基于遥感数据的景观识别和作物监测、基于地球物理数据的地下断层和结构解释、气候和天气预测、地质绘图、地下水探测、地球物理数据反演等。通过本课程，学生将学习关于深度学习的基础知识和技术，使用Python和典型的深度学习框架进行编码入门，了解不同的地球科学领域及深度学习算法在这些领域的应用。考虑到学生可能缺乏相关背景知识，讲座将跳过相关主题的数学和物理部分，而更多地侧重于算法背后的思想和地球科学应用。学生将在讲师的帮助下完成一到两个基于深度学习的地球科学应用编码任务。本课程不要求学生具备数学、物理或地球科学背景，讲座将以直观的方式进行，便于理解复杂的算法。

Deep learning and Geosciences Research Schedule

Tutoring Subject: Understanding basics of deep learning algorithms and applying them into different geoscientific scenarios

Abstract:

Deep learning becomes increasingly important in various industrial and scientific fields including geosciences. Some examples can include landscape identification and crop monitoring based on remote sensing data, subsurface fault and structure interpretation based on geophysical data, climate and weather prediction, geological mapping, ground water exploration, geophysical data inversion, etc. Through this program, students would learn basic knowledge and techniques about deep learning, get started with coding using python and a typical deep learning framework, know different geoscientific areas and how deep learning algorithms can be used in these areas. Considering the possible weak backgrounds of students, the lecture will skip mathematical and physical parts for related topics and focus more on the ideas behind algorithms and geoscientific applications. Students are expected to finish one or two coding tasks for deep learning-based geoscientific applications with the help of lecturer. No mathematical, physical, or geoscientific background is required for this lecture. Lectures will be given intuitively for better understanding of complex algorithms.