Selected work

Projects & publications

Research in AI and machine learning, from methods and benchmarks to applied systems. Explore the work, read the papers, and find code and citations in one place.

2026

SpaRRTa: A Synthetic Benchmark for Evaluating Spatial Intelligence in Visual Foundation Models

Turhan Can Kargin, Wojciech Jasiński, Adam Pardyl, Bartosz Zieliński, Marcin Przewięźlikowski

IJCV · Accepted for publication

A controlled benchmark for testing whether visual foundation models recognise relative positions between objects.

BibTeX citation
@article{kargin2026sparrta,
  title = {SpaRRTa: A Synthetic Benchmark for Evaluating Spatial Intelligence in Visual Foundation Models},
  author = {Kargin, Turhan Can and Jasiński, Wojciech and Pardyl, Adam and Zieliński, Bartosz and Przewięźlikowski, Marcin},
  journal = {International Journal of Computer Vision},
  year = {2026},
  note = {Accepted for publication},
  url = {https://arxiv.org/abs/2601.11729}
}
2025

FlySearch: Exploring how vision-language models explore

Adam Pardyl, Dominik Matuszek, Mateusz Przebieracz, Marek Cygan, Bartosz Zieliński, Maciej Wołczyk

NeurIPS

A photorealistic outdoor benchmark for testing goal-driven exploration by vision-language models.

BibTeX citation
@inproceedings{pardyl2025flysearch,
  title = {FlySearch: Exploring how vision-language models explore},
  author = {Pardyl, Adam and Matuszek, Dominik and Przebieracz, Mateusz and Cygan, Marek and Zieliński, Bartosz and Wołczyk, Maciej},
  booktitle = {Advances in Neural Information Processing Systems, Datasets and Benchmarks Track},
  year = {2025},
  doi = {10.52202/085713-5594}
}
2025

Beyond Grids: Exploring Elastic Input Sampling for Vision Transformers

Adam Pardyl, Grzegorz Kurzejamski, Jan Olszewski, Tomasz Trzciński, Bartosz Zieliński

IEEE/CVF Winter Conference on Applications of Computer Vision

Defines input elasticity and studies how vision transformers can process patches beyond fixed grids.

BibTeX citation
@inproceedings{pardyl2025beyondgrids,
  title = {Beyond Grids: Exploring Elastic Input Sampling for Vision Transformers},
  author = {Pardyl, Adam and Kurzejamski, Grzegorz and Olszewski, Jan and Trzciński, Tomasz and Zieliński, Bartosz},
  booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision},
  pages = {8536--8545},
  year = {2025},
  doi = {10.1109/WACV61041.2025.00827}
}
2025

AI-Driven Rapid Identification of Bacterial and Fungal Pathogens in Blood Smears of Septic Patients

Agnieszka Sroka-Oleksiak, Adam Pardyl, Dawid Rymarczyk, Aldona Olechowska-Jarząb, Katarzyna Biegun-Drożdż, Dorota Ochońska, Michał Wronka, Adriana Borowa, Tomasz Gosiewski, Miłosz Adamczyk, Henryk Telega, Bartosz Zieliński, Monika Brzychczy-Włoch

Computers in Biology and Medicine

Combines microscopic image segmentation with multiple-instance learning to identify pathogens in Gram-stained blood samples.

BibTeX citation
@article{srokaoleksiak2025aidriven,
  title = {AI-Driven Rapid Identification of Bacterial and Fungal Pathogens in Blood Smears of Septic Patients},
  author = {Sroka-Oleksiak, Agnieszka and Pardyl, Adam and Rymarczyk, Dawid and Olechowska-Jarząb, Aldona and Biegun-Drożdż, Katarzyna and Ochońska, Dorota and Wronka, Michał and Borowa, Adriana and Gosiewski, Tomasz and Adamczyk, Miłosz and Telega, Henryk and Zieliński, Bartosz and Brzychczy-Włoch, Monika},
  journal = {Computers in Biology and Medicine},
  volume = {199},
  pages = {111328},
  year = {2025},
  doi = {10.1016/j.compbiomed.2025.111328}
}
2024

AdaGlimpse: Active Visual Exploration with Arbitrary Glimpse Position and Scale

Adam Pardyl, Michał Wronka, Maciej Wołczyk, Kamil Adamczewski, Tomasz Trzciński, Bartosz Zieliński

European Conference on Computer Vision

Learns where to look and at what scale during sequential visual exploration.

BibTeX citation
@inproceedings{pardyl2024adaglimpse,
  title = {AdaGlimpse: Active Visual Exploration with Arbitrary Glimpse Position and Scale},
  author = {Pardyl, Adam and Wronka, Michał and Wołczyk, Maciej and Adamczewski, Kamil and Trzciński, Tomasz and Zieliński, Bartosz},
  booktitle = {Computer Vision -- ECCV 2024},
  pages = {112--129},
  year = {2024},
  doi = {10.1007/978-3-031-72664-4_7}
}
2023

Active Visual Exploration Based on Attention-Map Entropy

Adam Pardyl, Grzegorz Rypeść, Grzegorz Kurzejamski, Bartosz Zieliński, Tomasz Trzciński

International Joint Conference on Artificial Intelligence

Uses attention-map entropy to choose informative observations for active visual exploration.

BibTeX citation
@inproceedings{pardyl2023ame,
  title = {Active Visual Exploration Based on Attention-Map Entropy},
  author = {Pardyl, Adam and Rypeść, Grzegorz and Kurzejamski, Grzegorz and Zieliński, Bartosz and Trzciński, Tomasz},
  booktitle = {Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence},
  pages = {1303--1311},
  year = {2023},
  doi = {10.24963/ijcai.2023/145}
}
2023

CompLung: Comprehensive Computer-Aided Diagnosis of Lung Cancer

Adam Pardyl, Dawid Rymarczyk, Joanna Jaworek-Korjakowska, Dariusz Kucharski, Andrzej Brodzicki, Julia Lasek, Zofia Schneider, Iwona Kucybała, Andrzej Urbanik, Rafał Obuchowicz, Zbisław Tabor, Bartosz Zieliński

European Conference on Artificial Intelligence

A computer-aided CT pipeline for lung nodule analysis and patient-level malignancy prediction.

BibTeX citation
@inproceedings{pardyl2023complung,
  title = {CompLung: Comprehensive Computer-Aided Diagnosis of Lung Cancer},
  author = {Pardyl, Adam and Rymarczyk, Dawid and Jaworek-Korjakowska, Joanna and Kucharski, Dariusz and Brodzicki, Andrzej and Lasek, Julia and Schneider, Zofia and Kucybała, Iwona and Urbanik, Andrzej and Obuchowicz, Rafał and Tabor, Zbisław and Zieliński, Bartosz},
  booktitle = {Proceedings of the European Conference on Artificial Intelligence},
  pages = {1835--1842},
  year = {2023},
  doi = {10.3233/FAIA230471}
}
A whole-slide tissue image with highlighted regions, important image patches and matching positive and negative visual prototypes.
Figure 1 from ProtoMIL: explaining whole-slide classification through image patches and visual prototypes. Figure: ProtoMIL authors.

ECML PKDD · 2022

ProtoMIL: Multiple Instance Learning with Prototypical Parts for Whole-Slide Image Classification

Dawid Rymarczyk, Adam Pardyl, Jarosław Kraus, Aneta Kaczyńska, Marek Skomorowski, Bartosz Zieliński

ECML PKDD

An interpretable multiple-instance learning approach for whole-slide image classification using visual prototypes.

BibTeX citation
@inproceedings{rymarczyk2022protomil,
  title = {ProtoMIL: Multiple Instance Learning with Prototypical Parts for Whole-Slide Image Classification},
  author = {Rymarczyk, Dawid and Pardyl, Adam and Kraus, Jarosław and Kaczyńska, Aneta and Skomorowski, Marek and Zieliński, Bartosz},
  booktitle = {Machine Learning and Knowledge Discovery in Databases},
  pages = {421--436},
  year = {2022},
  doi = {10.1007/978-3-031-26387-3_26}
}