METHODS OF INFORMATION TECHNOLOGY: Statistical Signal Processing and Machine Learning

We go on a journey from methods of signal processing (deterministic, statistical) to methods of machine learning. On our way we will apply the respective methods in various applications. Starting with an overview of the methods used, we look at
(1) methods for estimation and detection based on physics-based models. Here we particularly discuss the application to positioning using trilateration (GPS) and triangulation (using antenna arrays). The results are point estimates of the unknown parameters. Using statistical approaches (Bayesian statistics) we present the respective solutions, which not only yield the maximum likelihood estimate but also information on the reliability of the results. Next, we replace the physics-based models with machine learning models and look at
(2) methods for regression and classification. 

Having acquired a solid understanding of the basic approaches using statistical signal processing and machine learning, we also discuss the differences and similarities of the methods (1) and (2) from the mathematical and application point of view  Finally, based on our “lessons learned” we look at the use of large models (e.g. Large Language Models like ChatGPT) and how they work.