Course overview
This course provides an in-depth exploration of large language models (LLMs), focusing on advanced principles underlying their architecture, development, and deployment. Students will engage in training, fine-tuning, and benchmarking LLMs, while critically analysing their capabilities and limitations across diverse downstream applications. Through extensive hands-on practicals, students will develop proficiency in building personalised AI agents with ethic consideration and adopt responsible approaches for their application in both industry and academia.
- Foundations of Large Language Models
- Training and Enhancing LLMs
- Downstream Applications of LLMs
Course learning outcomes
- Identify and explain the foundational principles of LLMs, including architecture, data sources, training pipelines, benchmarking, and ethical/societal considerations.
- Critically evaluate existing LLM architectures and training strategies, identifying their capabilities, limitations, and emerging trends.
- Apply practical techniques to train, fine-tune, and deploy LLMs, addressing performance limitations identified during evaluation.
- Design methods to enhance inference speed and reasoning capabilities, and extend LLM knowledge through techniques such as prompting, augmentation, and chaining.
- Develop and apply LLM-based solutions for real-world downstream applications across diverse domains and consider ethical/societal factors.