Course Overview
Course in brief
This course will teach you the principles of large language models. The focus is on forming an understanding of the underlying idea of contemporary artificial intelligence and large language models, i.e. learning patterns and structures from data.
We look into the history of artificial intelligence and the developments that have led to contemporary artificial intelligence, including generative artificial intelligence. We look into the history of language models, discussing the concept of probability and probability distribution. We explore approaches for building language models that led to the current large language models, and discuss current large language models. We also discuss some of the challenges related to artificial intelligence and using large language models, including the role of artificial intelligence developers in making decisions on what sorts of viewpoints the models are exposed to and consequently expose to others. We also look into the challenges in using large language models, including hallucination, biases, and learn basics of prompt engineering.
This course is a pre-requisite course for the course Software Engineering with Large Language Models.
Learning objectives
The high-level learning objectives of the course are as follows.
-
You understand the role of data in contemporary artificial intelligence.
-
You understand what language models are.
-
You have a high-level understanding of how large language models work.
-
You know of the history and the key milestones leading to the present large language models.
-
You know techniques for prompting large language models and understand the basics of prompt engineering.
-
You understand some of the challenges in using large language models, including hallucination and potential biases.
Specific learning objectives are outlined at the beginning of each chapter.
Prerequisites and tools
The course materials have been written in English, so a basic understanding of the English language is needed. Beyond the language, the course has no prerequisites. The ability to search for additional information when needed is expected.
The course can be taken fully online. A modern browser with JavaScript enabled is needed to access the course materials. The course does not require any installations on your computer.
Workload
The estimated workload of the course is between 10 and 20 study hours. The concrete workload varies between individuals due to e.g. differences in experience.