Barbara Klaudel
Co-founder and co-leader of TheLion.AI
A bit about me
I lead TheLion.AI, an interdisciplinary research group devoted to creating AI-based solutions for healthcare. TheLion.AI comprises 30+ researchers, software developers, healthcare professionals, and students. We build open-source solutions for healthcare and bring together healthcare professionals. We also provide consulting services and develop commercial solutions.
TheLion.AI is an organization I started from scratch with Aleksander Obuchowski. When we were freshman students, there was no research group devoted to AI in healthcare, so we had to create one ourselves! We were the supervisors of all our research projects, got several grants from the EU, and co-founded a startup.
For my work, I was awarded the title of the best student at the Gdańsk University of Technology. I received a congratulatory letter from the rector that they had never had a student who would achieve so much in so little time.
I was listed at Forbes 25 under 25 in New Technology.
I am currently looking for PhD opportunities.
Research Experience
My work is focused on developing applications to tackle a specific healthcare problem and democratizing the process of developing such applications.
My accumulated impact factor is 28.254. So far, I published four journal articles and four conference proceedings. I presented my work at four international and 20+ national conferences.
In all of the following projects, either I or my co-founder acted as a supervisor.
Foundation models
UMIE datasets: A framework for creating a large-scale medical imaging dataset of 1 million images from 20 opensource datasets, downloadable within a click, with a unified set of annotations for 20 masks and 46 distinct labels and a pipeline with easy-to-use drag-and-drop steps for appending additional medical datasets.
UMIE: A Universal Medical Image Encoder pre-trained on UMIE dataset of a million X-ray, CT and MRI images for tasks, such as classification, segmentation and reconstruction. Open-source release in 2024!
Cross-adaptation: A novel method for domain generalization particularly suited for developing computer-aided diagnosis models based on patient cohorts from different countries, accepted to AAAI-24.
MIDAS: A tool for the creation and management of medical datasets, either for deep learning or any form of statistical research. This project turned into a startup of which I am the CEO. The startup won the final prize at the EIT Digital Venture competition, and my business plan won second place at the EIT Digital Venture.
Parametrization of radiology reports: A tool for creating structured radiology reports from free-text reports in Polish. Paper published at EACL.
AI applications in healthcare
TITAN: a computer-aided diagnostic tool for kidney tumor malignancy classification based on CT scans. Implemented in 3 hospitals and included in the top 10 AI solutions already changing the Polish hospitals -- "AI is not Sci-Fi" report and presented at the European Congress of Radiology.
Morpheus.ai: A screening tool for identifying patients with SARS-CoV-2 and forecasting the disease prognosis based on routine blood tests.
CICAAD-RC: A catheter-induced coronary and aortic dissection risk calculator. We created the largest in the world database of dissections and were able to identify its predictors for the first time.
Machine learning for X-ray diffraction: A XAI pipeline for filtering X-ray diffraction patterns from my internship at the Deutsches Elektronen-Synchrotron DESY. This is the only project where I was supervised.
Teaching Experience and Community Service
I co-authored the new curriculum for the AI major at the Polish Japanese Academy of Information Technology in Gdańsk (the other author leads TheLion.AI with me). I have taught five self-prepared courses (lectures with laboratory classes):
-
Future of Deep Learning - LLMs, Foundation Models, Ethics
-
Computer Vision - CNNs, Vision Transformers, Segmentation, Detection, Classification
-
AI in Automatic Control
-
Process Diagnostics
-
Heuristic Methods.
I also teach at the Gdańsk University of Technology (1 BE course and 1 MA - even though I do not have a master's degree yet).
Some of my lectures are going to be available at YouTube soon.
I also created an AI in medicine course targeted at healthcare managers.
I co-host a popular science podcast about AI, "Cybernetyczny pomyślunek" (en. "Cyber-savvy").
I co-created several opensource apps:
-
An app for allocating volunteers over several refugee aid centers based on their unique skills. The app was put into production a week after the Russian invasion of Ukraine. The app is used by the Gdańsk Pomaga volunteer organization and received special thanks from the mayor of Gdańsk.
-
SMS Shopping for Seniors - an app connecting seniors with volunteers during the pandemic. Used by scouts and awarded 2nd place in the Hack the Crisis competition organized by the Polish Prime minister.
-
All of TheLion.AI projects are released as open-source upon completion. We also assist the hospitals with the implementation. For instance, check Chattum.
I am a mentor of a student from Ukraine in the Intel New Technologies for Women from Ukraine program.