2019 Imaging Informatics Summit
Add to Calendar
Start with the fundamentals. Learn key AI concepts through these five videos. You’ll hear how AI is developed and how AI tools can be applied to imaging data and improve radiology workflow. We’ll explore the areas of AI with the most tools available, look at bias and fairness in AI models, and explain what radiologists should look for in evaluating an AI algorithm.
What AI Is (and Is Not): 10 Thoughts
By: Katherine Andriole, PhD
Radiologists need to be able to describe what artificial intelligence (AI), machine learning (ML) and deep learning (DL) are — and how they work at a high level — to better understand how they can contribute to medical imaging. This video explores the processes and infrastructure required to develop and implement machine learning model applications and the importance of data cohort design in developing a model.
- To describe at a high level what AI is and what it is not.
- The processes and infrastructure required for AI.
- The importance of data cohort design to AI.
- Examples of AI applied to imaging data and to radiology workflow.
Workflow-Based AI: Beyond Image Interpretation
By: Paras Lakhani, MD
Workflow-based AI applications can be powerful tools and may achieve results more quickly than image-based tools because they can be rapidly implemented in practice. Just a few of the AI tools in this category include study protocoling/decision support, hanging protocol facilitation, improving image quality, decreasing scanner time, and optimizing patient and staff scheduling — all valuable to radiology practice.
What workflow-based AI includes.
Examples of workflow-based AI tools.
How workflow-based AI can result in performance gains for radiology practices.
Applications of Pixel-Based AI
By: Judy Wawira Gichoya, MBchB, MS
There are potentially thousands of possibilities for pixel-based AI to help radiologists in image classification, detection and segmentation. In this video you’ll learn which three areas of AI have the most AI tools available today and hear about how they are developed.
The current state of pixel-based AI systems.
The huge potential for pixel based AI products.
Bias and Fairness in AI Models
By: Monica J. Wood, MD
Unwanted bias may be incorporated unwittingly into AI models at points throughout an algorithm’s lifecycle — including the creation of training datasets, selection of model architecture and refinement of the algorithm post-deployment. And just as radiologists perform quality assurance on all our imaging modalities, we should aim to incorporate bias evaluation in routine assessment of AI algorithms.
How biases make their way into AI algorithms.
Examples of bias affecting the training data.
How to engage in meaningful conversations on bias.
Brittleness of AI Models
By: Woojin Kim, MD
“Brittleness” is the term for an AI model which works well where it was trained but fails when taken outside that area and subjected to new data. There are no guarantees that an algorithm developed elsewhere will function well at your institution — even if it has FDA-clearance. The video explores issues relating to brittleness including concept drift, where the goal of the algorithm changes over time, and data drift, which involves changes created by using new data sources to generate medical imaging data.
The limitations of AI and what can be done to mitigate them.
How validation and training issues contribute to brittleness.
What radiologists should look for in evaluating an AI algorithm.