Deep Learning Competence Center

German Research Center for Artificial Intelligence (DFKI)


The Competence Center for Deep Learning of the DFKI focuses on:

Our research is based on deep learning and machine learning algorithms and ranges from basic research to industrial knowledge transfer. In this context we employ deep Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) approaches to advance machine perception, work with deep learning frameworks such as CAFFE and Torch, collaborate with academic institutes in this area, and teach these approaches to build a new generation of students embracing machine learning. Our work is funded through public national and European research grants and direct industrial contracts.

Our expertise is centered around Deep Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), Bi-directional Recurrent Neural Networks, Convolutional Deep Belief Networks, Deep Boltzmann Machines, Autoencoders, Deep Reinforcement Learning, Support Vector Machines (SVM), Vector Quantization (VQ), Energy Functional Minimization, RANSAC - based optimization, Kernel Methods, k-Nearest Neighbor (kNN), Gaussian Mixture Models (GMM), Conditional Random Fields (CRF), Max Entropy, Hidden Markov Models (HMM), Case-base Reasoning (CBR).

research projects


Tasty food, sweet animals, breathtaking landscapes - to characterize something briefly, an adjective together with a noun is being used. These adjective noun pairs (ANP) describe the visual content of an image together with the feelings that it triggers at the viewer. If they occur with great frequency, they can be used for the machine description of images that go beyond a textual reproduction of the visual content. Capttitude (Captions with Attitude) is a system able to produce affective captions with an emotional component. Two different methods are used, which as a result provide two variants of a text caption with emotional content; a Convolutional Neural Network (CNN) together with a long-short-term-memory (LSTM) network, and in the second approach a graph-based concept and syntax transition (CAST) network.


The project DeepEye faces the challenge of recognizing natural catastrophes in satellite images and to enrich the information with multi-media content gained from the social media. With an image analysis, the different spectral bands of satellite data are combined and the geographic areas affected by a natural catastrophe will be extract. In a multimodal analysis, relevant information from text, image and meta data is extracted using various methods of machine learning such as Convolutional Neural Networks. The focus of the analysis is the extraction of contextual aspects in order to obtain a comprehensive and complete view of a specific event. These prepared data can be used in the context of crisis management, for example for the coordination of the rescue forces in situ. With the combination of satellite data and multi-media content from social media, DeepEye aims for the next big step in crisis management: detailed depictions of natural catastrophes by the fusion of different information channels.

What do Deep Networks Like To See

A novel way to measure and understand convolutional neural networks is by quantifying the amount of input signal they let in. To do this, an autoencoder (AE) was fine-tuned on gradients from a pre-trained classifier with fixed parameters. Comparing the reconstructed samples from AEs that were fine-tuned on a set of image classifiers (AlexNet, VGG16, ResNet-50, and Inception v3), substantial differences were found. The AE learns which aspects of the input space to preserve and which ones to ignore, based on the information encoded in the backpropagated gradients. Measuring the changes in accuracy when the signal of one classifier is used by a second one, a relation of total order emerges. This order depends directly on each classifier’s input signal but it does not correlate with classification accuracy or network size. Further evidence of this phenomenon is provided by measuring the normalized mutual information between original images and auto-encoded reconstructions from different fine-tuned AEs. These findings break new ground in the area of neural network understanding, opening a new way to reason, debug, and interpret their results.


Prof. Dr. Prof. h.c. Andreas Dengel


Dr. Boris

Senior Researcher

Dr-Ing. Saqib Bukhari

Senior Researcher

Dr. Christian Dugast

Senior Researcher

Dr. Ahmed

Senior Researcher

Dr.-Ing. Michael Feld

Senior Researcher

Dr. Peter

Senior Researcher

Dr. Alexis

Senior Researcher

Dr. Daniel

Senior Researcher

Dr.-Ing. Yohannes Kassahun

Senior Researcher

Dr. Su-Kyoung

Senior Researcher

Dr. Elsa

Senior Researcher

Dr. Mario Michael Krell

Senior Researcher

Dr. Christian Schulze

Senior Researcher

Dr.-Ing. Ahmed Sheraz

Senior Researcher

Dr.-Ing. Daniel Sonntag

Principal Researcher

Dr. Jon

Senior Researcher

Contact Us

Prof. Andreas Dengel
German Research Center for Artificial Intelligence
Trippstadter Straße 122
67663 Kaiserslautern

Phone: (+49 631 20575 1000)