Rochester, Minnesota




Hamid Tizhoosh, Ph.D., explores the applications of artificial intelligence (AI) in medicine, particularly in medical image analysis and cross relations to other patient data such as molecular, laboratory and textual data. His research is currently focused on search and matching in archives of patient data.

Given the data of a new patient, one can infer invaluable diagnostic and prognostic knowledge about the patient in a precise and timely manner by identifying similar patients evidently diagnosed and treated. The ultimate goal is to automatically generate consensus reports for the expert physician to reduce intra- and inter-observer variability.

Focus areas

  • Unsupervised and self-supervised learning. Although supervised AI is intensively investigated, Dr. Tizhoosh emphasizes the role of nonsupervised topologies in artificial intelligence. He and his team investigate the design and performance of unsupervised and self-supervised paradigms such as search, matching, clustering and visualization.
  • Generative models in histopathology. Learning a histology to generate synthetic histopathology data can be extremely valuable for model development and education. Dr. Tizhoosh and his team investigate the use of many generative adversarial networks in histopathology by looking at clinical data.
  • Federated learning. The patient privacy and computational and storage challenges of curating large medical archives across multiple hospitals can be addressed by decentralized federated learning to share solutions without sharing patient data. Dr. Tizhoosh and his team work on design and testing of different federated schemes to test across Mayo Clinic campuses and beyond.
  • Bias, shortcuts and generalization. Dr. Tizhoosh, his teams and collaborators are paying special attention to the generalization capabilities of AI solutions, particularly deep models, to facilitate widespread and reliable deployment of computerized techniques in medicine. Detecting biases and shortcuts in the form of redundant, irrelevant information in data is a major component of most experiments.
  • Multimodal and cross-modal patient representation for search. Dr. Tizhoosh's research culminates in search and matching technologies to reduce observer variability. Design and validation of deep topologies that simultaneously learn different patient data and their internal correlations is based on his research in all aforementioned AI subfields.

Significance to patient care

Delivering computational consensus to physicians through search and matching in multimodal data of previous patients can increase the accuracy of personalized medicine for both diagnosis and treatment. Patients can be diagnosed and treated with higher levels of confidence, resulting in fewer side effects and shorter treatment times. As well, the waiting times for patients to receive diagnostic decisions from physicians can be shortened, resulting in — among other advantages — reduction or elimination of biopsy stress.

Professional highlights

  • Ontario Research Fund — Research Excellence (ORF-RE) grant on AI in digital pathology, 2018-2023
  • Founder and director, Laboratory of Knowledge Inference in Medical Image Analysis, University of Waterloo, 2013-present
  • Recognized in the Power List of pathologists who achieved big breakthroughs, The Pathologist, 2020


Administrative Appointment

  1. Senior Associate Consultant II-Research, Department of Artificial Intelligence and Informatics

Mayo Clinic Footer