YouTube video by どれこちゃんねる

Join Dr. Seraphine Maerz, Senior Lecturer at the University of Melbourne, for a practical, research-focused online seminar on fine-tuning large language models for text analysis. Designed for PhD students and professional researchers across disciplines, this hands-on workshop explores when to fine-tune and when to use alternatives such as prompt engineering, preparing and evaluating data, applying LoRA, and building reproducible workflows. Guided Jupyter/Colab notebooks make each step approachable, and no prior Python or R experience is required—just a willingness to work hands-on with code. Attend live on Zoom or learn asynchronously: seminar recordings and materials are available for 30 days, with an expert-monitored Q&A forum and a certificate of completion included. disciplines #HealthSciences #PoliticalScience #SocialSciences #InformationScience #Statistics #ComputerScience #DataScience #PublicHealth #Linguistics #GoogleColab #Jupyter #Python #R #RStudio #Research #ResearchTraining #Instats

How can you use large language models to produce reliable, reproducible insights from specialized text? Join Dr. Seraphine Maerz, Senior Lecturer at the University of Melbourne, for a practical, one-day online seminar on fine-tuning LLMs for research. Designed for PhD students and professional researchers across disciplines, the workshop combines clear conceptual guidance with hands-on Jupyter/Colab notebooks—no prior Python or R experience required. Explore when to fine-tune and when to use prompt engineering, prepare and evaluate data responsibly, fine-tune BERT with LoRA, and build reproducible workflows for text classification and analysis. Attend live via Zoom or learn asynchronously: recordings and materials are available for 30 days, with an expert-monitored Q&A forum and a certificate of completion. disciplines #HealthSciences #PoliticalScience #SocialSciences #InformationScience #Statistics #ComputerScience #DataScience #Ethics #GoogleColab #Jupyter #Python #R #RStudio #Research #ResearchTraining #Instats

What if you could analyze your research data using plain English instead of writing code? Join Peter Gruber for Colloquial Data Science, a three-day online Instats workshop designed for PhD students, academics, and professional researchers. Now in its sixth iteration, the seminar explores AI-powered statistical workflows across ChatGPT, Google Gemini, Claude, and Microsoft Copilot—from data cleaning and visualization to regression, time-series analysis, and publication-ready tables and figures. Through practical exercises and interactive challenges, you’ll learn not only how to prompt AI tools effectively, but also how to verify their outputs, recognize their limitations, and protect sensitive data while maintaining academic rigor. Delivered via Zoom, the workshop includes access to recordings and materials for 30 days after the seminar begins, instructor-supported discussions, office hours for participants in challenging time zones, and an Instats certificate of completion. Discover a more efficient approach to empirical research—without sacrificing critical judgment. #Psychology #Education #Management #PublicHealth #Statistics #DataScience #InformationScience #ComputerScience #ArtificialIntelligence #PhilosophyofScience #SoftwareEngineering #Mathematics #ChatGPT #GoogleColab #Gemini #Claude #Copilot #Research #ResearchTraining #Instats

What if you could explore your research data using plain English instead of writing code? Join Professor Peter Gruber for “Using ChatGPT for Advanced Data Analysis 5.0,” a three-day online workshop designed for PhD students, academics, and professional researchers. Through practical examples, exercises, and interactive challenges, you’ll learn how to use ChatGPT for descriptive statistics, regression, data visualization, interpreting results, and drafting research reports—while developing the critical skills to assess AI-generated outputs and understand their limitations. Updated content explores the latest AI tools, privacy and data security, and emerging AI agents. Delivered via Zoom with a schedule optimized for American time zones, the workshop includes recordings and materials available for 30 days after it begins, instructor-supported discussions, and an Instats certificate of completion. Explore a more intuitive approach to statistical analysis—find out more and register at instats.org. #Psychology #Education #Management #PublicHealth #Statistics #DataScience #InformationScience #ComputerScience #ArtificialIntelligence #ChatGPT #GoogleColab #Research #ResearchTraining #Instats

Fine-tuning large language models is becoming an essential research skill for extracting reliable, reproducible insights from specialized text data. Join our one-day online workshop, Fine-Tuning Large Language Models for Text Analysis, taught by Dr. Seraphine Maerz, Senior Lecturer at the University of Melbourne, and designed for PhD students, professors, and professional researchers in health sciences, political science, the social sciences, and related fields. This practical seminar combines clear conceptual guidance with hands-on Jupyter/Colab notebooks in Python and R, covering when to fine-tune versus use prompt engineering, data preparation and annotation, BERT fine-tuning for classification, LoRA for parameter-efficient fine-tuning, evaluation metrics, error analysis, reproducibility, and interpretation of results. No prior Python or R experience is assumed, but participants should be ready to work hands-on with code. The seminar is delivered via Zoom, with recordings, materials, and a monitored Q&A forum available for 30 days after the session, plus an official Instats certificate of completion. disciplines #HealthSciences #PoliticalScience #SocialSciences #InformationScience #Statistics #ComputerScience #DataScience #PublicHealth #Linguistics #GoogleColab #Jupyter #Python #R #RStudio #Research #ResearchTraining #Instats

Political and social scientists are working with more text than ever, from parliamentary speeches and treaty preambles to government reports, news coverage, social media posts, and policy documents. This 3-day intensive online seminar shows PhD students and professional researchers how to use large language models to turn complex text into defensible, analysis-ready data without requiring prior machine learning or Python experience. Through live Google Colab demonstrations, participants will learn when to use encoder models such as BERT and DeBERTa, when generative models such as GPT or Claude are more appropriate, how to apply zero-shot classification, build and validate gold-standard datasets, evaluate performance with metrics such as accuracy, F1 scores, Cohen’s kappa, and confusion matrices, and create end-to-end pipelines for extracting and cleaning data from PDFs, websites, and other messy sources. The seminar is designed to help researchers strengthen methodological credibility, accelerate data collection and coding, and expand the range of research questions that can be tackled with text. All sessions are livestreamed via Zoom, with recordings, slides, code, data, and a monitored Q&A forum available for 30 days after the seminar, plus an official Instats certificate of completion. #ArtificialIntelligence #ComputationalSocialScience #PoliticalScience #ComputerScience #Statistics #DataScience #InformationScience #SocialTheory #Sociology #Claude #GoogleColab #HuggingFace #Jupyter #Python #Research #ResearchTraining #Instats

Google AI subscribers now get premium Colab benefits! Unlock faster accelerators, Premium GPUs, and uninterrupted background execution for long training runs.

Ready to move from “I’d like to use text-as-data” to actually building and running your own pipeline? This three-day hands-on online seminar is designed for PhD students and professional researchers in the social sciences who want to apply text-as-data methods without getting stuck on the programming. Starting from no coding background, you’ll learn the Python you actually need, link messy datasets using spelling and semantic similarity, model themes with LDA and the Structural Topic Model, and work with BERT-based encoder models to classify political text at scale. By the end, you’ll have fine-tuned, validated, saved, and reused your own language model in free browser-based notebooks, with every step scaffolded and supported. The seminar is delivered live via Zoom in three 2.5-hour sessions, with recordings, materials, and a monitored Q&A forum available for 30 days afterward, plus an official Instats certificate of completion. #ComputationalSocialScience #PoliticalScience #ComputerScience #DataScience #Statistics #GoogleColab #Jupyter #Python #Research #ResearchTraining #Instats