Weronika Hryniewska-Guzik

I am a PhD candidate at the Faculty of Mathematics and Information Science at Warsaw University of Technology in Poland, where I specialize in responsible, explainable Deep Learning on lung images. My PhD supervisor is Przemysław Biecek.

I have an MSc in Computer Science from Łódź University of Technology. I have a BSc in Mechatronics from Warsaw University of Technology.

Email  /  LinkedIn  /  GitHub  /  Google Scholar  /  ORCID  /  CV

Research

I'm interested in deep learning, computer vision, and multi-modality.

CNN-based explanation ensembling for dataset, representation and explanations evaluation

This research investigates the potential of combining explanations generated by deep learning classification models using convolutional models to create a more coherent and reliable understanding of the model’s behavior. Using introducted metrics, the paper addresses how these explanations aid in assessing datasets, representations, and explanations.

NormEnsembleXAI: Unveiling the Strengths and Weaknesses of XAI Ensemble Techniques

This research analyzes methods for ensembling explainable AI (XAI) techniques. We introduce new method NormEnsembleXAI, analyze the strengths and weaknesses of XAI ensembles, and provide a library facilitating the practical implementation of XAI ensembling.

A comparative analysis of deep learning models for lung segmentation on X-ray images

This study evaluates deep learning solutions for lung segmentation in X-rays, ranking existing methods and analyzing their performance under diverse image modifications. Out of 61 analyzed papers, only nine offered implementation or pre-trained models, enabling assessment of three prominent methods: Lung VAE, TransResUNet, and CE-Net.

Multi-task learning for classification, segmentation, reconstruction, and detection on chest CT scans

A novel multi-task framework for classification, segmentation, reconstruction, and detection. To the best of our knowledge, we are the first ones who added detection to the multi-task solution. Additionally, we checked the possibility of using two different backbones and different loss functions in the segmentation task.

Challenges facing the explainability of age prediction models: case study for two modalities

We investigate the use of Explainable Artificial Intelligence (XAI) for age prediction focusing on two specific modalities, EEG signal and lung X-rays. We share predictive models for age to facilitate further research on new techniques to explain models for these modalities.

LIMEcraft: Handcrafted superpixel selection and inspection for Visual eXplanations

We propose an approach called LIMEcraft that allows a user to interactively select semantically consistent areas and thoroughly examine the prediction for the image instance in case of many image features.

Prevention is better than cure: a case study of the abnormalities detection in the chest

Faults in the data acquisition and annotation process in a dataset for detection of abnormalities in X-ray lung images.

Checklist for responsible deep learning modeling of medical images based on COVID-19 detection studies

The review study of approaches proposed in the numerous Machine Learning articles and indication of typical errors emerging from the lack of deep understanding of the radiography domain. We present the perspective of both: experts in the field - radiologists and deep learning engineers dealing with model explanations. The final result is a proposed checklist with the minimum conditions to be met by a reliable COVID-19 diagnostic model.

Teaching

Teaching assistance in courses.

Ensemble explanations in image classification

Case Study - What should you know before writing your thesis

Case Study - Preparing and training models on tabular, text and image data

Case Study - Deep Learning on medical image data

Text Mining

This course focuses on basic theory, advanced models and methods for information retrieval in large textual datasets.