The group will host a monthly journal club (usually at noon on the third Tuesday of the month), where we provide an opportunity for faculty to present mini lectures, for trainees to present works in progress, or just to discuss new literature. If you are interested in presenting, please email dlag@fredhutch.org.

Upcoming Schedule

• Tuesday, May 20, 2025
  12:10-1:00 p.m.
  Hybrid: Arnold Building M3-A805 and Zoom
  Presenter/Discussion lead: TBD
  Topic: TBD

• Tuesday, September 16, 2025
  12:10-1:00 p.m.
  Hybrid: Arnold Building M3-A805 and Zoom
  Presenter/Discussion lead: TBD
  Topic: TBD

• Tuesday, October 21, 2025
  12:10-1:00 p.m.
  Hybrid: Arnold Building M3-A805 and Zoom
  Presenter/Discussion lead: TBD
  Topic: TBD

• Tuesday, November 18, 2025
  12:10-1:00 p.m.
  Hybrid: Arnold Building M3-A805 and Zoom
  Presenter/Discussion lead: TBD
  Topic: TBD

• Tuesday, December 16, 2025
  12:10-1:00 p.m.
  Hybrid: Arnold Building M3-A805 and Zoom
  Presenter/Discussion lead: TBD
  Topic: TBD

Past Meetings

• March 18, 2025
  Recording Available
  Presenter/Discussion lead: Ty Lambert (DaSL); Robert McDermott (IT)
  Topic: DeepSeek Discussion, continued:
  Ty Lambert, AI and Research Data Protections Program Manager, Office of Chief Data Officer
  Title: Artificial Intelligence Governance at Fred Hutch
  Robert McDermott, Director, Solutions, Engineering & Architecture
  Title: Inside the Private Thoughts of AI: How DeepSeek’s Inner Monologue Redefines What We Expect from Language Models

• February 18, 2025
  Recording Available; Slide Deck
  Presenter/Discussion lead: Eardi Lila and Youyi Fong
  Topic: DeepSeek R1 and V3 Tech Reports
  Intro, Scaling Laws,  - Youyi Fong (VIDD & PHS)
  

  Multitoken Latent Attention, MoE - Eardi Lila (UW Biostat)

  
  
  
• November 19, 2024
  Recording Available
  Presenter/Discussion lead: Lucas Liu
  In this meeting, we will discuss the latest findings from the American Medical Informatics Association (AMIA) Annual Symposium hosted last week. AMIA Annual Symposium is the world's premier meeting for the research and practice of biomedical and health informatic. We will focus on the expert opinion about the current state and future of AI in medical informatics, with special emphasis on the role of NLP/LLMs. In addition, we will discuss a couple of example papers.
  Xie Q, Chen Q, Chen A, et al. Me LLaMA: Foundation Large Language Models for Medical Applications, https://arxiv.org/abs/2402.12749

• October  15, 2024
  Recording Available
  Presenter/Discussion lead: Wei Sun
  Topic: Exploit Spatially Resolved Transcriptomic Data to Infer Cellular Features from Pathology Imaging Data
  In the September journal club, we talked about medical imaging, focusing on CT or MRI images. This month, we will talk about pathology images. Particularly, we will discuss an on-going work that exploits spatial transcriptomic data to annotate pathology images and uses such annotations to train deep learning models to characterize whole slide H&E stained images.
  Zhining Sui, Ziyi Li, Wei Sun, bioRxiv 2024.08.05.606654; doi: https://doi.org/10.1101/2024.08.05.606654

• September 17, 2024
  Recording Available
  Presenter/Discussion lead: Saishi Cui
  Transformers in Medical Imaging: A New Era of AI for Diagnosis. This presentation explores the emerging role of transformer models in medical imaging, challenging the dominance of convolutional neural networks (CNNs). By capturing global context, transformers have demonstrated superior performance in a variety of tasks such as image segmentation, classification, and more. This review will cover architectural advancements, their application in diagnostic accuracy, and how transformers are reshaping the future of AI in healthcare. Shamshad, F., Khan, S., Zamir, S. W., Khan, M. H., Hayat, M., Khan, F. S., & Fu, H. (2023). Transformers in medical imaging: A survey. Medical Image Analysis, 88, 102802. 

• June 4, 2024
  Recording Available
  Presenter/Discussion lead: Si Liu
  One Version of Attention in Graph Neural Network
  Brody et al. 2022, How attentive are graph attention networks? ICLR, 2022 , which is an updated version of the graph attention as in a previous paper: Velickovic et al. Graph attention networks. ICLR, 2018

• May 28, 2024
  Recording Available
  Presenter/discussion lead: Wei Sun
  A continuation our discussion of the Transformer model that powers chatGPT. We will discuss transformer and related methods for omics data.
  TULIP—a Transformer based Unsupervised Language model for Interacting Peptides and T-cell receptors that generalizes to unseen epitopes.
  scGPT: toward building a foundation model for single-cell multi-omics using generative AI.

• April 02, 2024
  Recording Available  Slide Deck
  Discussion Lead: Lucas Liu
  A continuation of our discussion of the Transformer model that powers GPT and walth through the math behind it step-by-step. Additionally, we will introduce Vision Transformers — an essential breackthrough in image analysis.
  Attention is All You Need
  An Image is worth 16x16 Words: Transformers for image recognition at scale

• March 19, 2024
  Recording Available 
  Lucas Liu and John Kang: Transformer Model
  Discussion of the Transformer model that powers GPT and walth through the math behind it step-by-step. Additionally, we will introduce Vision Transformers — an essential breackthrough in image analysis.
  Attention is All You Need
  An Image is worth 16x16 Words: Transformers for image recognition at scale

• February 20, 2024
  Recording Available
  Robert McDermott: Multimodal (language and vision) model
  Visual Instruction Tuning
  Improved Baselines with Visual Instruction Tuning
  Experimentation results and real example

• January 23, 2024
  Recording Available
  Youyi Fong
  Segment Anything Model
  A Foundation model for Cell Segmentation
  Fast SAM

• November 21, 2023
  Recording Available
  Lucas Liu: Survival analysis and deep learning
  https://bmcmedresmethodol.biomedcentral.com/articles/10.1186/s12874-018-0482-1
  https://arxiv.org/abs/1907.00825
  https://ojs.aaai.org/index.php/AAAI/article/view/11842
  https://www.mdpi.com/2072-6694/14/23/5807
  John Kang:  Scaling Clinical Trial Matching
  https://arxiv.org/abs/2308.02180

• October 17, 2023
  Recording Available
  Wei Sun
  A brief introduction biologically informed neural network and generalizability of neural network methods
  John Kang and Youyi Fong
  Sybil: A Validated Deep Learning Model to Predict Future Lung Cancer Risk From a Single Low-Dose Chest Computed Tomography