BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:Fine Tuning Large Language Models (LLMs) with Domain Specific 
 Datasets
X-WR-TIMEZONE:Pacific Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260607T053246Z
UID:tag:localist.com\,2008:EventInstance_51745542260313
DTSTART:20260414T180000Z
DTEND:20260414T190000Z
DESCRIPTION:Large language models (LLMs) are trained on massive\, publicly 
 available text datasets comprising trillions of tokens\, enabling them to 
 excel at general language tasks like next-token prediction. However\, LLMs
  often struggle with domain-specific prompts\, exhibiting reduced accuracy
  or generating inaccurate information (hallucinations). This is because th
 ey lack sufficient subject matter expertise. Two primary approaches exist 
 to address this limitation for augmenting LLMs knowledge: Retrieval-Augmen
 ted Generation (RAG) and fine-tuning. This presentation focuses on fine-tu
 ning smaller LLMs with domain-specific instruct datasets using the LoRA (L
 ow-Rank Adaptation) technique on Gaudi hardware. We will leverage publicly
  available LLMs and datasets from the Hugging Face Hub for this demonstrat
 ion. Though it is possible to fine tune LLMs with plain text data – sour
 ced from documents\, articles\, and other materials.\n\nInstructor Biograp
 hy:\nMadhusudan Gujral is currently a bioinformatics lead at SDSC. His bac
 kground is in structural biology\, but he transitioned to the field of inf
 ormatics over 20 years ago. He began by developing a client-based laborato
 ry information system (LIM) for a distributed biological project with user
 s across the US. This was followed by a large project creating complex pip
 elines for metagenomics research. He then spent a decade processing and an
 alyzing whole genome sequencing (WGS) data from thousands of samples colle
 cted from patients with psychiatric disorders. For the past two years\, he
  has focused on learning and benchmarking fine-tuning large language model
 s (LLMs) on Gaudi hardware.
LOCATION:
SUMMARY:Fine Tuning Large Language Models (LLMs) with Domain Specific Datas
 ets
URL;VALUE=URI:https://calendar.ucsd.edu/event/fine-tuning-large-language-mo
 dels-llms-with-domain-specific-datasets
CATEGORIES:Science and Technology
END:VEVENT
END:VCALENDAR