The course will follow the training logic of the major technology players in the field (e.g. NVIDIA) and is aimed at business/academia interested in learning about efficient prompting through real use cases in addition to a general overview. The course will demonstrate through sample implementations how to easily and efficiently exploit large language models based on prompting in areas such as document analysis, writing marketing materials, developing custom AI personas and custom chatbots, customer assistance, generating program code/tests, creating lyrics and music or detecting objects in images. At the conceptual level, the course will cover: precise prompts, multi-step prompting, response control, sentiment analysis, prompt templates, prompt fine-tuning, one-shot/few-shot learning, putting the system in context, random sampling for non-deterministic responses, adjusting temperature to control the degree of randomness, preserving conversation history, limits on text length. A basic knowledge of the Python language and communication with large language models is a useful pre-requisite, but the sample implementations are sufficiently illustrated and automated to make practical training easy.
The training will take place on Friday 07 June 2024 from 08.00 to 12.00 (CET), in a hybrid format; in person at the University of Limassol Cyprus campus or simultaneously online.
Students and staff of the University of Debrecen are entitled to an 80% discount on the participation fee:
https://eec.uol.ac.cy/programs/promptengineering/