no code implementations • 27 Sep 2024 • Soroosh Tayebi Arasteh, Mahshad Lotfinia, Paula Andrea Perez-Toro, Tomas Arias-Vergara, Mahtab Ranji, Juan Rafael Orozco-Arroyave, Maria Schuster, Andreas Maier, Seung Hee Yang
To generalize our findings across languages and disorders, we validated our approach on a dataset of Spanish-speaking Parkinson's disease patients, leveraging pretrained models from healthy English-speaking datasets, and demonstrated that careful pretraining on large-scale task-specific datasets can maintain favorable accuracy under DP constraints.
no code implementations • 22 Jul 2024 • Soroosh Tayebi Arasteh, Mahshad Lotfinia, Keno Bressem, Robert Siepmann, Lisa Adams, Dyke Ferber, Christiane Kuhl, Jakob Nikolas Kather, Sven Nebelung, Daniel Truhn
We evaluate the diagnostic accuracy of various LLMs when answering radiology-specific questions with and without access to additional online information via RAG.
1 code implementation • 11 Apr 2024 • Soroosh Tayebi Arasteh, Tomas Arias-Vergara, Paula Andrea Perez-Toro, Tobias Weise, Kai Packhaeuser, Maria Schuster, Elmar Noeth, Andreas Maier, Seung Hee Yang
This study investigates anonymization's impact on pathological speech across over 2, 700 speakers from multiple German institutions, focusing on privacy, pathological utility, and demographic fairness.
1 code implementation • 1 Oct 2023 • Soroosh Tayebi Arasteh, Christiane Kuhl, Marwin-Jonathan Saehn, Peter Isfort, Daniel Truhn, Sven Nebelung
So far, the impact of training strategy, i. e., local versus collaborative, on the diagnostic on-domain and off-domain performance of AI models interpreting chest radiographs has not been assessed.
1 code implementation • 27 Aug 2023 • Soroosh Tayebi Arasteh, Tianyu Han, Mahshad Lotfinia, Christiane Kuhl, Jakob Nikolas Kather, Daniel Truhn, Sven Nebelung
A knowledge gap persists between machine learning (ML) developers (e. g., data scientists) and practitioners (e. g., clinicians), hampering the full utilization of ML for clinical data analysis.
2 code implementations • 15 Aug 2023 • Soroosh Tayebi Arasteh, Leo Misera, Jakob Nikolas Kather, Daniel Truhn, Sven Nebelung
In this study, we explored if SSL for pre-training on non-medical images can be applied to chest radiographs and how it compares to supervised pre-training on non-medical images and on medical images.
1 code implementation • 10 Jun 2023 • Soroosh Tayebi Arasteh, Mahshad Lotfinia, Teresa Nolte, Marwin Saehn, Peter Isfort, Christiane Kuhl, Sven Nebelung, Georgios Kaissis, Daniel Truhn
We specifically investigate the performance of models trained with DP as compared to models trained without DP on data from institutions that the model had not seen during its training (i. e., external validation) - the situation that is reflective of the clinical use of AI models.
no code implementations • 18 May 2023 • Soroosh Tayebi Arasteh, Cristian David Rios-Urrego, Elmar Noeth, Andreas Maier, Seung Hee Yang, Jan Rusz, Juan Rafael Orozco-Arroyave
Parkinson's disease (PD) is a neurological disorder impacting a person's speech.
1 code implementation • 18 Apr 2023 • Gustav Müller-Franzes, Fritz Müller-Franzes, Luisa Huck, Vanessa Raaff, Eva Kemmer, Firas Khader, Soroosh Tayebi Arasteh, Teresa Nolte, Jakob Nikolas Kather, Sven Nebelung, Christiane Kuhl, Daniel Truhn
Accurate and automatic segmentation of fibroglandular tissue in breast MRI screening is essential for the quantification of breast density and background parenchymal enhancement.
2 code implementations • 3 Feb 2023 • Soroosh Tayebi Arasteh, Alexander Ziller, Christiane Kuhl, Marcus Makowski, Sven Nebelung, Rickmer Braren, Daniel Rueckert, Daniel Truhn, Georgios Kaissis
In this work, we evaluated the effect of privacy-preserving training of AI models regarding accuracy and fairness compared to non-private training.
1 code implementation • 18 Dec 2022 • Firas Khader, Gustav Mueller-Franzes, Tianci Wang, Tianyu Han, Soroosh Tayebi Arasteh, Christoph Haarburger, Johannes Stegmaier, Keno Bressem, Christiane Kuhl, Sven Nebelung, Jakob Nikolas Kather, Daniel Truhn
Multimodal deep learning has been used to predict clinical endpoints and diagnoses from clinical routine data.
1 code implementation • 14 Dec 2022 • Gustav Müller-Franzes, Jan Moritz Niehues, Firas Khader, Soroosh Tayebi Arasteh, Christoph Haarburger, Christiane Kuhl, Tianci Wang, Tianyu Han, Sven Nebelung, Jakob Nikolas Kather, Daniel Truhn
The success of Deep Learning applications critically depends on the quality and scale of the underlying training data.
1 code implementation • 24 Nov 2022 • Soroosh Tayebi Arasteh, Peter Isfort, Marwin Saehn, Gustav Mueller-Franzes, Firas Khader, Jakob Nikolas Kather, Christiane Kuhl, Sven Nebelung, Daniel Truhn
Due to the rapid advancements in recent years, medical image analysis is largely dominated by deep learning (DL).
1 code implementation • 7 Nov 2022 • Firas Khader, Gustav Mueller-Franzes, Soroosh Tayebi Arasteh, Tianyu Han, Christoph Haarburger, Maximilian Schulze-Hagen, Philipp Schad, Sandy Engelhardt, Bettina Baessler, Sebastian Foersch, Johannes Stegmaier, Christiane Kuhl, Sven Nebelung, Jakob Nikolas Kather, Daniel Truhn
Furthermore, we demonstrate that synthetic images can be used in a self-supervised pre-training and improve the performance of breast segmentation models when data is scarce (dice score 0. 91 vs. 0. 95 without vs. with synthetic data).
1 code implementation • 13 Apr 2022 • Soroosh Tayebi Arasteh, Tobias Weise, Maria Schuster, Elmar Noeth, Andreas Maier, Seung Hee Yang
Navigating the challenges of data-driven speech processing, one of the primary hurdles is accessing reliable pathological speech data.
1 code implementation • 21 Apr 2021 • Soroosh Tayebi Arasteh, Mehrpad Monajem, Vincent Christlein, Philipp Heinrich, Anguelos Nicolaou, Hamidreza Naderi Boldaji, Mahshad Lotfinia, Stefan Evert
As a strong baseline, we propose a two-stage DL-based method: first, we create automatically labeled training data by applying a standard sentiment classifier to tweet replies and aggregating its predictions for each original tweet; our rationale is that individual errors made by the classifier are likely to cancel out in the aggregation step.
Ranked #1 on
Tweet-Reply Sentiment Analysis
on RETWEET
1 code implementation • 4 Feb 2021 • Mahshad Lotfinia, Soroosh Tayebi Arasteh
Moreover, instead of using a constant value of flow stress during the multi-pass rolling process, we use a Finite Difference (FD) formulation of the equilibrium equation in order to account for the dynamic behavior of the flow stress, which leads to the estimation of the mean pressure, which the strip enforces to the rolls during deformation.
1 code implementation • 16 Nov 2020 • Soroosh Tayebi Arasteh, Adam Kalisz
Splines are one of the main methods of mathematically representing complicated shapes, which have become the primary technique in the fields of Computer Graphics (CG) and Computer-Aided Geometric Design (CAGD) for modeling complex surfaces.
Graphics Computational Geometry Algebraic Geometry
no code implementations • 10 Nov 2020 • Soroosh Tayebi Arasteh
While many researchers in the speaker recognition area have started to replace the former classical state-of-the-art methods with deep learning techniques, some of the traditional i-vector-based methods are still state-of-the-art in the context of text-independent speaker verification.