Slides Framework

IEEE VIS 2022 Workshop:
Visualization in Biomedical AI

Virtual, Monday, Oct 17 9:00 am-12:00 pm (UTC-5)
IEEE VIS 2022, Oklahoma City, USA (Hybrid), Oct 16-21 ๐Ÿ”—

Artificial Intelligence (AI) is advancing biomedical science in many ways, from improving image-based diagnostics to identifying new drugs. Despite its extraordinary performance, AI in biomedical applications should be treated cautiously due to the high stakes involved. Effective visualization can significantly improve the information communication between AI systems and human users and is playing an increasingly important role in biomedical AI.

The aim of this workshop is to explore the challenges and opportunities in this highly interdisciplinary research by bringing together researchers from data visualization, biomedical informatics, machine learning, and computational biologists.

Call for Participation

In the PCS system, please select society: VGTC, Conference: VIS 2022, Track: Visualization in Biomedical AI

Submission: Aug 17, 2022

Notification: Sep 10, 2022

Camera-Ready: Sep 22, 2022


We invite submissions from both 1) visualization researchers who conduct application studies on biomedical AI and 2) biomedical AI researchers who address their domain challenges with the help of visualizations. Apart from visualization tools, case studies, evaluations, and design guidelines, We also highly encourage white papers that discuss scenarios where visualizations are most needed and the obstacles in applying visualizations, regardless of whether the biomedical researchers have developed visualization tools or not.

This workshop will accept two types of submissions:

  • A 2-4 page paper (without references) of unpublished work using the IEEE VGTC Conference Style Template. Accepted submissions in this category will be published similar to posters: available on the workshop's website, not available at a central digital library, included in the downloadable proceedings, not considered archival, but possible to be cited. Re-use of the content in a follow-up publication or conference paper will be fully allowed.
  • An up-to 400-word short abstract related to a recent study that has been published at a non-Visualization venue (e.g., Oxford Bioinformatics, NeurIPS, Nature Methods). No template is required. The previously published paper must be attached as supplementary material along with the submission.

Topics include (but are not limited to):

  • Visual exploration and analytics that support the application of AI to biomedical and healthcare problems, including medical image diagnosis, clinical decision making, biomedical discovery, computational omics, etc
  • Visualization techniques and tools that are designed for a certain group of users of biomedical AI, such as wet lab biologists, physicians, histopathologists, drug developers, patients
  • Visual AI explanations for solving biomedical and healthcare challenges
  • Design guidelines for visualization in biomedical AI
  • The role of and the needs for visualization in biomedical AI
  • Evaluation of visual explanations, interactive visualization techniques, and visualization tools for biomedical AI
  • Broader challenges and research opportunities

Submissions are not anonymous and should include all author names, affiliations, and contact information. At least one author of each accepted submission needs to register for the IEEE VIS conference and attend the workshop.

Schedule

(Oct 17, 2022, UTC-5)

Time Session
9:00-9:05 am
(5 min)
Workshop Introduction
9:05-9:55 am
(50 min)
Presentation: Human-AI Collaboration in Medicine
Carrie Cai (Google Research)
9:55-10:15 am
(20 min)
Paper Session 1

Tightening the Loop in Mixed-Initiative ML Engineering and Domain Annotation
using Active Learning and Visual Analytics
๐Ÿ“Ž PDF
Mert Erkul, Piriyakorn Piriyatamwong, Batuhan Tomekce, Manuel Morales Wyden, William Baumgartner,
Elizabeth White, Michael Bada, Lawrence Hunter, Mennatallah El-Assady


Opening Access to Visual Exploration of Audiovisual Digital Biomarkers:
an OpenDBM Analytics Tool
๐Ÿ“Ž PDF
Carla Gabriela, Jacob Epifano, Stephanie Caamano, Sarah Kark,
Rich Christie, Aaron Masino, Andre Paredes
30 min Break
10:45-11:05 am
(20 min)
Paper Session 2

An Interactive Interpretability System for Breast Cancer Screening with Deep Learning ๐Ÿ“Ž PDF
Yuzhe Lu, Adam Perer

Kokiri: Random Forest-Based Cohort Comparison and Characterization ๐Ÿ“Ž PDF
Klaus Eckelt, Patrick Adelberger, Markus Johann Bauer, Thomas Zichner, Marc Streit
11:05-11:55 am
(50 min)
Panel Discussion
Adam Perer (Carnegie Mellon University), David Gotz (UNC-Chapel Hill),
Marc Streit (Johannes Kepler University Linz), and Marinka Zitnik (Harvard Medical School)
11:55-12:00 am
(5 min)
Workshop Summary


Keynote Presentation: Human-AI Collaboration in Medicine

Carrie Cai, Google Research

Carrie Cai is a Research Scientist at Google in PAIR (People+AI Research) and the Responsible AI division. Her research aims to make human-AI interactions more productive and enjoyable to end-users, ranging from tools to help doctors steer cancer-diagnostic systems in real time, to frameworks for effectively onboarding end-users to AI assistants. Her work has been published at top-tier human-computer interaction venues such as CHI, IUI, CSCW, and VL/HCC, receiving 7 best paper / honorable mention recognitions and profiled in the MIT Technology Review and TechCrunch. Carrie holds a PhD in Computer Science at MIT. Prior to MIT, she completed an MA in Education and a BA in Human Biology at Stanford University. Coming from an interdisciplinary background, she believes that the world's best AI innovations arise from a deep understanding of both machine learning technologies and human behavior.



Panel Discussion

Organizers

  • Qianwen Wang

    Harvard University, qianwen_wang@hms.harvard.edu

    Qianwen Wang is a Postdoctoral Fellow at Harvard University. Her research strives to facilitate the communication between end users and machine learning models through creating interactive visual analysis systems, with a special interest in their applications in solving biomedical challenges. She serves as a review for multiple visualization venues (IEEE VIS, IEEE TVCG, PacificVIS, EuroVIS, ACM CHI, ACM IUI), abstract chair for the ISMB BioVis COSI, and program committee for the visualization meets AI workshop at PacificVis. Her work has been published in the top visualization and human-conputer interaction venues (IEEE VIS, IEEE TVCG, ACM CHI) and received awards from both the BioVis community and the Machine Learning community.

  • Vicky Yao

    Rice University, vy@rice.edu

    Vicky Yao is an Assistant Professor in the Department of Computer Science at Rice University. Her research involves developing and applying statistical and machine learning methods to study complex disease with a particular focus on neurological disease and cancer. Accompanying these methods, she builds interactive visualizations that enable exploration of complex multi-dimensional data and predictions. Her work has been published in top-tier journals, including Nature Biotechnology, the New England Journal of Medicine, and Neuron. She is a CPRIT Scholar and received a Gordon Wu Fellowship from Princeton University. She was also a co-organizer for the ICML Workshop on Interpretable Machine Learning in Healthcare.

  • Bum Chul Kwon

    IBM Research, bumchul.kwon@us.ibm.com

    Bum Chul Kwon is a Research Staff Member at IBM Research, where he is a member of Center for Computational Health. His research goal is to enhance users' abilities to derive knowledge from data using interactive visualization systems. His work has been published at premier venues in visualization and human-computer interaction, such as IEEE VIS, IEEE TVCG, and ACM SIGCHI. He is an associate chair for the ACM CHI Paper Program Committee, a publicity chair for IEEE VIS, and a general chair for the Visual Analytics Health Care workshop. He also serves on the program committee for IEEE InfoVis, PacificVis, ACM IUI, and Visual Analytics in Healthcare Workshop.

  • Nils Gehlenborg

    Harvard University nils@hms.harvard.edu

    Nils Gehlenborg is an Associate Professor of Biomedical Informatics at Harvard Medical School. For the past 15 years, his work has focused on the interface between biomedical data and data visualization research. His publications have appeared in premier venues such as IEEE VIS, EuroVis, Nature, Science, and Cell. Nils is a co-founder of BioVis, the Symposium on Biological Data Visualization, and co-founder of VIZBI, the annual workshop on Visualizing Biological Data.

Feel free to contact the organizer committee through vis-biomed-ai-workshop@googlegroups.com if you have any questions.

Program Committee

  • Furui Cheng, the Hong Kong University of Science and Technology
  • Thomas Hรถllt, TU Delft
  • Robert Krรผger, Harvard University
  • Fritz Lekschas, Ozette Technologies
  • Heba Sailem, the University of Oxford
  • Marc Streit, Johannes Kepler University Linz