More than ever, the current century is characterized by the insatiable accumulation
and collection of data that are further used for reductionistic classifications and
categorizations. However, taking into account that data management, data processing and
data politics are not neutral domains, but instead, are skewed and biased by the provided
training data sets as well as underlying knowledge production systems and epistemes, it
becomes a crucial task to identify these biases and overcome them through integrative and
sensitive research practices. Also, the recent years were marked by an explosive growth of
biopolitical measurements and quantifications leading to ever more sophisticated AI
algorithms. Though, the exact mechanisms of data collection and data processing are often
unclear and not transparent leading to decision-making AI systems that produce either
skewed or biased results. In consequence, these results are not generalizable and work
properly only for certain population groups.
Therefore, the primary goal of our research group is to develop, train and evaluate
unbiased explainable AI systems. Additionally, our research group is trying to improve
fairness of data as well as the utilized AI systems. In order to achieve this, it is crucial to
consider postmodern, postcolonial, feminist as well as queer perspectives and critiques of
contemporary epistemic knowledge productions systems. Furthermore, our research group
is continuously evaluating processes of data collection, processing, interferencing and
reinterpretation to locate potential sources of bias. The mechanisms and interactions
between contemporary knowledge production systems and data are highly complex and may
lead to temporary solutions that may not be perfect and even bear the risk to reproduce
stigmatizing or marginalizing contents. However, our research group is actively engaging in
overcoming potentially stigmatizing and biased AI systems and contents. To illustrate our
approaches, some of our recent publications in the field of AI and fairness are linked below.
Besides, we are always open for discussion and suggestions for improvement.
In summary, our machines are learning and trying to become better and so are we.
So, if you notice contents, data or other materials that may be inappropriate, reproduce
contemporary marginalizing knowledge or be potentially harmful on any of our pages or
social media channels, you are highly welcomed to contact us anytime on the following
e-mail address: info@visir.org
List of Publications
Danner, M.; Huber, P.; Awais, M.; Feng, Z.; Kittler, J. and Raetsch, M. (2020).
Texture-based 3D Face Recognition using Deep Neural Networks for Unconstrained
Human-machine Interaction.In Proceedings of the 15th International Joint Conference on
Computer Vision, Imaging and Computer Graphics Theory and Applications – Volume 5:
VISAPP, ISBN 978-989-758-402-2, ISSN 2184-4321, pages 420-427. DOI:
10.5220/0008982504200427
Gerlach, T.; Danner, M.; Peng, L.; Kaminickas, A.; Fei, W. and Rätsch, M. (2020). Who
Loves Virtue as much as He Loves Beauty?: Deep Learning based Estimator for Aesthetics of
Portraits.In Proceedings of the 15th International Joint Conference on Computer Vision, Imaging
and Computer Graphics Theory and Applications – Volume 5: VISAPP, ISBN 978-989-758-402-2,
ISSN 2184-4321, pages 521-528. DOI: 10.5220/0009172905210528
J. Kittler, W. P. Koppen, P. Kopp, P. Huber, M. Rätsch. Conformal mapping of a 3D face
representation onto a 2D image for CNN based face recognition. 11th International Conference
on Biometrics, (ICB 2018, ISBN: 978-1-5386-4285-6), pp. 124-131, Australia, 2018.
X. Su, H. Zhou, V. P. Draghicic, M. Rätsch. Face naming in news images via Multiple
Instance Learning and Hybrid Recurrent Convolutional Neural Network. JOURNAL OF
ELECTRONIC IMAGING, 27(3), 033036), 2018.
P. Huber, W. Christmas, J. Kittler, P. Kopp, M. Rätsch. Real-time 3D Face Fitting and
Texture Fusion on In-the-wild Videos, IEEE Signal Processing Letters (Print ISSN: 1070-9908,
Online ISSN: 1558-2361), 2016.
P. Huber, G. Hu, R. Tena, P. Mortazavian, W. Koppen, W. Christmas, M. Rätsch , J.
Kittler. A Multiresolution 3D Morphable Face Model and Fitting Framework, International
Conference on Computer Vision Theory and Applications (VISAPP), vol. 4, pp. 79-86 (ISSN:
978-989-758-175-5), Rome, Italy. 2016.
S. Wittig, U. Kloos, M. Rätsch. Emotion Model Implementation for Parameterized Facial
Animation in Human-Robot-Interaction, International Conference on Computer Technology and
Development (ICCTD), Journal of Computers, vol. 11, no. 6, pp. 439-445 (JCP, ISSN:
1796-203X), Singapore. 2016. (journal award)
P. Huber, J. Kittler, M. Rätsch. Bottom-up and Top-down Face Analysis based on 3D
Face Models, Informatics Inside Conference for Human-Centered Computing, pp 138. 2014
(award)
An extensive list with all publications can be found here: https://www.visir.org/people/prof-dr-rer-nat-matthias-ratsch/