Artificial Intelligence in Aerospace Engineering

课程简介

/ Indicative Syllabus Foundations of Mathematics and Programming for Aerospace AI: Introduction to the mathematical and coding essentials for aerospace AI. Topics include linear algebra, probability theory, optimization methods, hands-on Python and PyTorch exercises. Emphasis on translating theoretical concepts into functional models, fostering confidence in implementing algorithms for aerospace challenges. Classical Supervised Algorithms for Aerospace Data Analysis: Exploration of classical supervised learning techniques for aerospace data analysis. Includes concept learning, regression and classification algorithms (SVM, k-NN, decision trees, and naïve Bayes), with hands-on implementation. Developing proficiency in algorithm selection and performance analysis for aerospace data interpretation. -- 1 of 4 -- 2 Deep Spatial Learning in Aerospace Applications: Introduces deep learning architectures for extracting and interpreting spatial information in aerospace contexts. Covers multilayer perceptrons for feature embedding, convolutional neural networks for satellite imagery and grid-based analysis, and graph neural networks for unstructured data processing. Emphasis on leveraging spatial correlations to enhance pattern recognition and mapping fidelity. Temporal Sequence Modeling for Aerospace Systems: Exploration of advanced neural techniques for capturing temporal relationships in aerospace sensor data. Topics include recurrent neural networks, LSTM and GRU cells, and attention-based models, applied to control-signal forecasting and anomaly detection. Discussion on modeling time-dependent patterns to support accurate trajectory prediction, real-time monitoring, and adaptive control strategies. Data Analytics and Machine Learning Applications in Aerospace: Application of data analytics and machine learning in aerospace. Covers pattern discovery, topic modeling, genomics, and prediction in aerospace contexts. Discussion on scalability, interpretability, and legal/social/ethical considerations in aerospace data analytics.

目标

The objectives of this subject are to: 1. introduce the principles, concepts and models of most popular deep learning algorithms used in aerospace research and industry. 2. provide both theoretical and practical understanding of machine learning models such as deep neural networks. 3. develop proficiency in designing, training, and optimizing machine learning models using Python and PyTorch.

先修要求

/ Co-requisite/ Exclusion Nil

Teaching Pattern

Methodology The course facilitates learning through a combination of lectures, tutorials, lab sessions, and a group project, and does not include a reading week. Lectures provide comprehensive coverage of the course concepts, supplemented by examples and interactive question & answer sessions for clarity. Tutorials and lab sessions reinforce theoretical knowledge with practical exercises, emphasizing hands-on experience with machine learning tools and techniques. The group project engages students in collaborative work to apply theoretical concepts to real-world scenarios, enhancing their analytical and problem-solving skills.