Klavdiia Naumova

Hello I am an AI Research Engineer at the Yale-EPFL Laboratory for Intelligent Global Health Technologies. I work on developing interpretable and trustworthy deep learning-based approaches to support healthcare in resource-limited settings.

I am soon starting my PhD studies at the Intelligent Medical Systems Lab (IMSY) at the German Cancer Research Center (DKFZ).

Education

École Polytechnique Fédérale de Lausanne (EPFL)
MSc in Life Sciences Engineering
2021-2023

Master Thesis

inDISCO: INterpretable DIStributed COllaborative learning for biomedical images (iGH/MLO)

Supervisers: Prof. Mary-Anne Hartley (Yale/EPFL) and Prof. Martin Jaggi (EPFL)
External advisor: Dr. Sai Praneeth Karimireddy (UC Berkeley)

You can find a full text of my thesis at this link or enjoy a short description here.

Other Projects

Semester projects:

• Optical Elastography Pipeline for Shear Modulus Calculation (MicroBioRobotic Systems Laboratory)
• xDISCO: eXplainable DIStributed COllaborative learning for images (iGH/MLO)

Course projects:

• Detection of traffic cones coordinates using neural networks
• Image denoising with Noise2Noise neural network
• Understanding vegetarianism and veganism through the media using sentiment analysis (read)
• Developing language models for digital educational assisting

Moscow State University (MSU)
Specialist in Fundamental and Applied Chemistry
2015-2021

Final qualifying paper

Drug-templated synthesis as method of obtaining high-capacity silica containers with controlled structure (Surface Phenomena in Polymer Systems Lab, Russian Academy of Science (IPCE RAS))

Superviser: Dr. Olga V. Dement'eva (IPCE RAS)
External advisor: Dr. Maria V. Poteshnova (MSU)

Other Projects

Semester project:

• Synthesis of a 2-thiohydantoin derivative for cancer treatment (Biologically Active Organic Compounds Lab, MSU)

Research

My research interests lie at the intersection of deep learning and medicine with the main focus on model reliability. Currently, I am developing an approach based on prototypical part learning network to allow visually-interpretable and privacy-preserving bias identification in federated learning for biomedical images. I encourage you to check our latest preprint for more details.

Additionally, I review submissions for JAMA journals and care about sustainable AI initiatives

Publications

1. MyThisYourThat for Interpretable Identification of Systematic Bias in Federated Learning for Biomedical Images.
   Klavdiia Naumova, Arnout Devos, Sai Praneeth Karimireddy, Martin Jaggi, Mary-Anne Hartley
   npj Digital Medicine, 7, 238 (2024)
   

   [paper]

2. inDISCO: INterpretable DIStributed COllaborative learning for images.
   Klavdiia Naumova, Mary-Anne Hartley, Arnout Devos, Sai Praneeth Karimireddy, Martin Jaggi
   Submitted to the ICML 2023 Workshop on Federated Learning
   

   [paper]

Earlier publications can be found at my Google scholar page.

Talks

Distributed Learning: enabling AI in low resource environments
at the Workshop SMART-AI: Leveraging AI for the design and implementation of Health Decision Support Systems
EPFL, April 29-30, 2024

[LinkedIn post]

Posters

EPFL Engineering Industry Day
March 8, 2023

[poster]

Internships

Quantitative biology Intern
Nanolive SA
July-September 2022

Teaching

• Student Assistant at the CS-433 Machine learning course (EPFL, Fall semester 2022)
• Student Assistant at the MATH-205 Analysis IV course (EPFL, Spring semester 2022)

Personal

I enjoy jogging and dancing. I travel by train where possible and strive to acquire other eco-friendly practices in my daily life. Also, I am a huge fan of books on popular science and feminism. If you are passionate about them too, I'd be delighted to share a list of my favorite pieces to provide you with some captivating reading ideas