David Bani-Harouni

I am a PhD Student at the Computer-Aided Medical Procedures (CAMP) group under the supervision of Prof. Nassir Navab. My research focuses on enabling calibrated confidence expressions of language and vision-language models through reinforcement learning. I also investigate how agents can be trained for efficient clinical decision-making, by iteratively refining diagnostic hypotheses, selecting informative clinical tests, and optimizing the trade-off between diagnostic accuracy and resource use.

Before my PhD, I completed my master’s and bachelor’s degrees in computer science at TUM with a semester abroad at the National University of Singapore.

News

Jul 2026	One main track paper, ConRad, and one workshop paper, DeToxR, have been accepted to MICCAI 2026.
Apr 2026	I was part of the delegation from the Munich Center of Machine Learning to visit the UK and strengthen collaboration with their universities.
Jan 2026	Two papers, Rewarding Doubt and LA-CDM have been accepted to ICLR 2026 in Rio de Janeiro.
Sep 2024	MAGDA was published at MedAGI@MICCAI 2024.
Nov 2023	U-GAT, a graph-based method for COVID-19 outcome predictions was published in Nature Scientific Reports.
Oct 2023	Our paper Gradient Self-alignment in Private Deep Learning, based on my master’s thesis with Prof. Daniel Rückert was published at CareAI@MICCAI 2023.
Jun 2023	I started my PhD at the Technical University of Munich under the supervision of Prof. Nassir Navab.

Selected Publications

Calibrated Confidence Expression for Radiology Report Generation

David Bani-Harouni^*, Chantal Pellegrini^*, Julian Lüers, Su Hwan Kim, Markus Baalmann, Benedikt Wiestler, Rickmer Braren, Nassir Navab, and Matthias Keicher

MICCAI, 2026

arXiv Bib Code

@article{baniharouni2026calibrated,
  title = {Calibrated Confidence Expression for Radiology Report Generation},
  author = {Bani-Harouni, David and Pellegrini, Chantal and Lüers, Julian and Kim, Su Hwan and Baalmann, Markus and Wiestler, Benedikt and Braren, Rickmer and Navab, Nassir and Keicher, Matthias},
  journal = {MICCAI},
  year = {2026},
}

Language Agents for Hypothesis-Driven Clinical Decision Making with Reinforcement Learning

David Bani-Harouni, Chantal Pellegrini, Ege Özsoy, Matthias Keicher, and Nassir Navab

ICLR, 2026

arXiv Bib Code

@article{baniharouni2026language,
  title = {Language Agents for Hypothesis-Driven Clinical Decision Making with Reinforcement Learning},
  author = {Bani-Harouni, David and Pellegrini, Chantal and Özsoy, Ege and Keicher, Matthias and Navab, Nassir},
  journal = {ICLR},
  year = {2026},
}

Rewarding Doubt: A Reinforcement Learning Approach to Calibrated Confidence Expression of Large Language Models

David Bani-Harouni^*, Chantal Pellegrini^*, Paul Stangel, Ege Özsoy, Kamilia Zaripova, Matthias Keicher, and Nassir Navab

ICLR, 2026

arXiv Bib Code

@article{baniharouni2026rewarding,
  title = {Rewarding Doubt: A Reinforcement Learning Approach to Calibrated Confidence Expression of Large Language Models},
  author = {Bani-Harouni, David and Pellegrini, Chantal and Stangel, Paul and Özsoy, Ege and Zaripova, Kamilia and Keicher, Matthias and Navab, Nassir},
  journal = {ICLR},
  year = {2026},
}

MAGDA: Multi-Agent Guideline-Driven Diagnostic Assistance

David Bani-Harouni, Nassir Navab, and Matthias Keicher

MICCAI MedAGI, 2024

arXiv Bib

@article{baniharouni2024magda,
  title = {{MAGDA}: Multi-Agent Guideline-Driven Diagnostic Assistance},
  author = {Bani-Harouni, David and Navab, Nassir and Keicher, Matthias},
  journal = {MICCAI MedAGI},
  year = {2024},
}