Application development with LLM: Education on large language models

The development of applications based on large language models (LLM) represents a revolutionary paradigm in the field of AI. This education, presented by the esteemed professor dr. sc. Sande Martinčić-Ipšić from the Faculty of Informatics and Digital Technology Rijeka, aims to provide participants with a deep understanding of the principles of work, implementation and ethical issues that accompany the use of LLMs.

Content of the Education:

1. Introduction to LLM:
   • Overview of LLM applications in business contexts and tasks.
   • Analysis of the principles of operation of transformer architectures that lie at the basis of LLM.
   • Understanding the principles of learning large language models for text generation.
2. Phases of the Application Development Project:
   • Detailed overview of phases and activities in the process of developing applications based on LLM.
   • Consideration of techniques such as prompt engineering, fine tuning, castastrophic forgetting.
3. Techniques for Developing Applications with LLM:
   • Overview of techniques such as Parameter Efficient Fine-Tuning (PEFT) and Low-Rank Adaptation of Large Language Models (LoRA).
   • Learning from human feedback and applying Reinforcement Learning with Human Feedback (RLHF).
   • Access to Retrieval Augmented Generation (RAG) and vector databases to improve application performance.
4. LangChain - Building LLM Applications:
   • Presentation of LangChain, a project that illustrates the practical implementation of LLM in application building.
   • Consideration of hardware requirements, quantization of network parameters and challenges in building LLM applications.
5. Risks and Limitations Generating AI:
   • Analysis of biases and hallucinations that may arise from the use of generative language models.
   • Difference between learned and fine-tuned models and implications on the quality of results.
6. Responsible Use of AI and Impact on Society:
   • Consideration of ethical issues and responsible use of generative AI.
   • Identifying security aspects in the use of LLM and thinking about social impact.
7. Date, Location and Duration:
   • The education will be held at the Faculty of Informatics and Digital Technologies Rijeka, both in Hall 357, Radmile Matejčić 2, Rijeka and online.
   • The dates are 28.02.2024 and 20.03.2024, with a duration from 2 p.m. to 7 p.m.

This training provides an extraordinary opportunity for participants to venture into the world of LLMs, master key application development techniques and create the basis for responsible use of generative AI in future projects. Professor dr. sc. Sanda Martinčić-Ipšić, with her professional knowledge and experience, ensures a high-quality learning experience. Don't miss the opportunity to participate in this training that promises to transform your approach to app development.

More information about education is available at the link.

Submit your application via the published the Public Call You will fill in the required documentation defined in the public call and send it by post to the project coordinator.

If you need help filling out documentation and responding to a public call, we are here for you. Contact us at info@edihadria.eu Together, we will prepare everything you need to start the digital transformation process.