LLM Summer School 2024
July 1-7, 2024 • Yerevan, Armenia
Key Topics
Intro & Pretraining
Presented by Armen Aghajanyan (Meta FAIR)
Alignment & LLM Evaluation
Presented by Jiao Sun (USC), Karen Hambardzumyan (UCL), Jonathan May (USC)
Vision LLMs
Presented by Alex Andonian (MIT), Armen Aghajanyan (Meta FAIR)
LLM Interpretability, Safety & Privacy
Presented by Narine Kokhlikyan (Meta FAIR), Eugene Bagdasaryan (Google/UMass Amherst)
RAG & KG-enhanced LLMs
Presented by Erik Arakelyan (UCPH), Hrayr Harutyunyan (Google Research)
Featured Speakers
Armen Aghajanyan
Meta FAIR
Alex Andonian
MIT
Erik Arakelyan
UCPH
Hrayr Harutyunyan
Google Research
Eugene Bagdasaryan
Google/UMass Amherst
Karen Hambardzumyan
UCL
Narine Kokhlikyan
Meta FAIR
Jonathan May
USC
Jiao Sun
USC
Highlights
The first LLM Summer School in Armenia brought together over 70 students and researchers for an intensive week of learning and collaboration. The program featured lectures, hands-on workshops, and practical sessions where participants applied their knowledge to innovative projects.
The curriculum covered a range of both core and specialized topics in LLMs, including Transformer architectures, pretraining, model alignment, parameter-efficient tuning, in-context learning and reasoning, multimodality and Vision LLMs, LLM interpretability, safety & privacy, knowledge graphs, and retrieval augmented generation (RAG).
The event was co-hosted by the American University of Armenia and YerevaNN, with the lectures taking place at the AUA campus. Nebius AI and Picsart provided generous support for the summer school.