| CODE | GSC3202 | ||||||||
| TITLE | Machine Learning and Computer Vision for the Environmental Sciences | ||||||||
| UM LEVEL | 03 - Years 2, 3, 4 in Modular Undergraduate Course | ||||||||
| MQF LEVEL | 6 | ||||||||
| ECTS CREDITS | 4 | ||||||||
| DEPARTMENT | Geosciences | ||||||||
| DESCRIPTION | It is only natural for humans to observe the physical world and build mental models to try and predict the environment's future state. Over time, rules that helped to forecast the state of the sky and ocean, were formed. Eventually, mathematics and partial differential equations allowed for more formal definitions to be defined. The field of Artificial Intelligence and Machine Learning have enabled machines to be programmed and act according to the surrounding conditions. Algorithms that compliment the biological processes such as the human brain are implemented to solve a particular problem. It was only natural that Machine Learning methods find their way to the Geosciences. In this unit, students will initially be introduced to the theories behind Decision Tree Learning, Genetic Algorithms, Neural Networks, Fuzzy Logic and Component Analysis. Such methods will then be applied to real problems. For instance, Fuzzy Logic will be used to analyse Doppler radar data and to predict atmospheric turbulence for the aviation industry. Neural Networks will be used to model non-linear phenomena such as the El Nino oscillation. Intelligent methods to assimilate satellite data products will also be demonstrated. The advantages gained in image processing techniques through the use of Machine Learning methods will also be presented. Study-unit Aims: The aim of this unit is not to cover the statistical techniques used in data analysis but to concentrate on Machine Learning methods are their applicability to the Geosciences fields of study. The algorithms and their usefulness will be initially described and demonstrated on real problems. The main focus here will be on methodology and applications. Artificial Intelligence techniques can be applied to a lot of challenging problems in the natural sciences. Students will be introduced to these new concepts and encouraged to construct models based on supervised or unsupervised learning methods. Another aim of this unit is to introduce students to digital image processing. Methods used to enhance spatial information and perform clustering will be demonstrated. Feature extraction from two dimensional data will also be put forward. Learning Outcomes: 1. Knowledge & Understanding By the end of the study-unit the student will be able to: - recall the theory behind Decision Tree Learning; - discuss the principles of Fuzzy Logic Control; - explain how Neural Networks operate; - apply component Analysis; - recognise the limitations of traditional algorithms; - recall the principles behind Image Processing techniques. 2. Skills By the end of the study-unit the student will be able to: - adopt Machine Learning methods for optimised solutions; - building training and testing datasets; - develop models through supervised and unsupervised learning methods; - solve forward and inverse problems in the atmospheric and ocean sciences; - use AI techniques for short term forecasting or missing data imputation; - perform image processing techniques for enhancement, clustering and feature extraction. Main Text/s and any supplementary readings: - Haupt, Sue Ellen and Pasini, Antonello and Marzban, Caren. (2009). Artificial Intelligence Methods in the Environmental Sciences. Springer. - Hsieh, William W. (2009). Machine Learning Methods in the Environmental Sciences. Cambridge University Press. - Trauth, Martin H (2007). Matlab Recipes for Earth Sciences. Springer. - Mitchell, Tom (1997). Machine Learning. McGraw Hill. |
||||||||
| ADDITIONAL NOTES | Pre-Requisite Study-unit: GSC2400 | ||||||||
| STUDY-UNIT TYPE | Lecture | ||||||||
| METHOD OF ASSESSMENT |
|
||||||||
| LECTURER/S | Adam Gauci |
||||||||
|
The University makes every effort to ensure that the published Courses Plans, Programmes of Study and Study-Unit information are complete and up-to-date at the time of publication. The University reserves the right to make changes in case errors are detected after publication.
The availability of optional units may be subject to timetabling constraints. Units not attracting a sufficient number of registrations may be withdrawn without notice. It should be noted that all the information in the description above applies to study-units available during the academic year 2023/4. It may be subject to change in subsequent years. |
|||||||||