Do you have skills in applying AI methods for research? Are you interested in sharing your knowledge to help others advance their work? Applications are closed but we are looking for event helpers with technical expertise.
Who should apply?
Prerequisites:
Registration Fees:
Key Dates:
The tutorials will provide a broad overview of modern AI methods and tools for research, and will include hands-on examples in Python. We offer an optional AI Foundations Day during which participants will learn how neural networks are trained and evaluated, how vision and sequence models work, and how transfer learning lets researchers build on pretrained models rather than starting from scratch. Follow-up tutorials will illustrate best practices for applying these approaches across a full AI research lifecycle. Other topics will include foundation models, and the practicalities of scaling training and deployment. Dedicated sessions also address how to interpret AI results rigorously and make AI-based research reproducible.
Participants will be invited to develop their own ideas for project work, ideally building on an existing research project, exploring a new direction, or working directly with concepts from Day 1 to solidify learning. Throughout the week, participants will apply what they learn to their projects, with instructors and assistants available for guidance. We encourage participants working on similar problems to collaborate and learn from each other. On the final day, participants will share their progress and define a concrete plan for continuing the work after the event.
Training and evaluation workflow, loss function, stochastic gradient descent, fully connected networks
CNNs, computer vision tasks
RNNs, LSTMs, time series tasks
Freezing layers, embeddings, fine-tuning
Vision transformers, contrastive language-image pretraining
Loading pretrained models, fine-tuning
SkyPilot: scaling workflows on any infrastructure
Overview of available resources to UW researchers: Hyak, Tillicum, cloud credits