The EU Project MASTER started in January 2023. We met for a kick-off meeting on January 17 at the University of Patras. More information on the MASTER project can be found in our project profile or on the official MASTER website at www.master-xr.eu.
On June 16, the research department IML attended the congress of economics and innovation (Wirtschafts- und Innovationskongress) with its focus area future and innovation (Zukunfts- und Innovationsmeile) in Hanover as an exhibitor. In the Convention Center, Omair Bhatti presented the demo “Visual Attention Detection” as an approach to eyetracking for industrial AI. In this demo, mobile eyetrackers and modern deep learning models are used to identify objects that are fixated by the user. Real-time signals are processed with a multisensor pipeline, i.e., a flexible framework for the development of multimodal multisensor interfaces. The video signal of the front camera, the gaze signal and the attention events are visualized with a real-time monitoring tool.
At the congress, various topic areas addressed the latest developments in information technology, energy, nutrition, mobility, etc. Among the guests were Lower Saxony’s Prime Minister Stephan Weil and numerous others.
As part of the Innosys dialog event “Künstliche Intelligenz im Mittelstand” (AI in small and medium-sized enterprises) on May 23, Dr. Marc Herrlich of the research department IML gave a talk with the title “Human-centered Interactive Machine Learning”. Therein, he addressed the special approach of Interactive Machine Learning (IML) and the connection to Explainable AI (XAI). Particular attention was paid to interface design and the consideration of “Human Factors” to actually realize the potential of Active Learning and IML. The rather traditional approach (longer iteration cycles) was compared to the IML approach, and present and future example topics from the IML Oldenburg location were presented. Subsequently, a vivid and productive discussion with the audience emerged. More than 100 representatives of the local science and economic sector attended the well-frequented event. Two live VR demos that were supervised by Bengt Lüers and Dr. Marc Herrlich were enthusiastically received as well.
Oldenburger Press reports about our Applied Artificial Intelligene (AAI) Endowed Chair and the first DFKI IML demonstrator in the context of image processing, deep learning, automatic captioning, and the connection to the market potential of semantic image processing. See the article here or here.
CEWE Chairman Dr. Christian Friege, stresses: “We have to sustain the ability in Europe to create innovation in the area of Artificial Intelligence ourselves – also, in order to ensure a handling of it that corresponds with our values. As the regional economy, we want to contribute to the further establishing of this research cluster.”
The endowed professorship AAI is sponsored by CEWE, EWE AG, BÜFA, the OLB foundation, Dr. Caroline Neumüller, Alexander Neumüller, the economic association DER KLEINE KREIS e.V. and the Foundation of the Metal Industry in the North West for five years.
We successfully presented our demonstrator at DATEV Software Campus meeting. Visitors could see a real-time mobile eye tracking demo that highlighted the currently read text. This demo will be used to conduct our user studies in the SciBot Project.
The goal of our project SciBot is to identify behavioral patterns in human-computer interaction that can help in improving this interaction. In this project, we are mainly focusing on multimodal-multisensor interfaces, especially using eye tracking technology with text-based interfaces.
SciBot is a SoftwareCampus project (funded by BMBF). We collaborate with DATEV eG.
We achieved a new milestone in our EIT project ERICS: The chatbot Eike, which is based on our multimodal multisensor platform, was launched at handbookgermany.de.
Interakt: Daniel Sonntag becomes new editor-in-chief of the German Journal of Artificial Intelligence.
ERICS: German Telecom launches the Capiche product based on DFKI’s ERICS server (in German):
ERICS provides cloud-based translation and search services to improve communication with refugees in Europe when they are looking for online information about their integration.
DFKI supports GPU mobility concept via 5G.
DFKI provides open-source software tools to be downloaded from the Software section. For example, Fine-tuning deep CNN models on specific MS COCO categories (PDF). It covers our freely available software py-faster-rcnn-ft which allows straightforward fine tuning of deep convolutional neural networks for image classification.
ANNOUNCING The Handbook of Multimodal-Multisensor Interfaces: Foundations, User Modeling, and Common Modality Combinations. Volume 1
EDITORS: Oviatt, Sharon; Schuller, Bjorn; Cohen, Philip R; Sonntag, Daniel; Potamianos, Gerasimos; Krueger, Antonio
PUBLISHER: Morgan and Claypool/ACM Press.
This is a THREE volume series that presents the definitive state of the art and future directions of the field of Multimodal and Multi-Sensor interfaces.
The EIT Digital NewsRoom entry published ERICS Roadmap in Digital Cities.
See also the EIT digital news entry.
DFKI releases video about anomaly detection use case within distributed digital manufacturing architecture.
Kick-off ERICS in EIT Digital Infrastructure; we will adapt the digital manufacturing architecture to the ERICS usecases of real-time multimodal interaction and interactive machine learning with crowd-sourcing options.
DFKI deploys distributed digital manufacturing architecture at Siemens corporate technologies.
Springer Professional features our publication “Overview of the CPS for Smart Factories Project: Deep Learning, Knowledge Acquisition, Anomaly Detection and Intelligent User Interfaces”
DFKI conducted a user study regarding the evaluation of the handwriting and gesture recognition systems used in Smart Factories. The study was carried out in cooperation with MyScript and attracted a total of 120 participants.
Our publication gaze-guided object classification using deep neural networks for attention-based computing is featured by Pupil Labs. We built our software on top of their open source eye tracking platform.
Patrick van der Smagt joins Volkswagen Data Lab and leads an new AI division on Robotic Enterprise.
ELTE and DFKI present two technical papers about deep learning for gaze estimation and hand gestures at KI 2016.
DFKI presents 4 smart factories demos at Ubicomp 2016.
Michael Barz participates in The International School on Human-Centred Computing (HCC 2016) and demonstrates our gaze-based techniques towards spatial reasoning in smart factories.
Joint press release of Volkswagen AG and DFKI: Volkswagen participates in DFKI and becomes a partner (shareholder). In a joint carrier project for cps for smart factories, Volkswagen and DFKI have started to develop a software framework to allow for new collaborative HRI (human-robot interaction) scenarios.
Daniel Sonntag will present our activity at ESWC 2016 for EU H2020 networking.
Mehdi Moniri presents DFKI’s smart factories Baxter demonstrator Eyebots to Joachim Gauck, the President of the Federal Republic of Germany. See the future scenario of this human-robot interaction demo here.
DFKI presents with EyeBots a new form of Human-Robot-Collaboration (HRC) at CeBIT 2016, contact Mohammad_Mehdi.Moniri@dfki.de.
We thank our sponsors BMBF, MADMACS, EIT DIGITAL, and SMI.
New Carrier programme: One of five Federal competence centers of Industry 4.0 in Germany will be opened in Kaiserslautern at DFKI and smartFactory.
New Carrier programme: DFKI and ZeMA signed a cooperation agreement for “Power4Production” (P4P); smart production technologies in the automobile sector are in the focus.
We will exhibit at AUTOMATICA 2016, the leading platform for optimizing your production processes.
Siemens. Facts and Forecasts. Boom for Learning Systems: Autonomous Robots and Smart Machine Market.
Markus Weber (DFKI, CPS for Smart Factories) wins one of the CeBIT Innovation Awards 2015.
CPS for Smart Factories integrates the EIT cPAS activity and focuses on combined use cases in the steel and robot domain. We welcome our new
Industry 4.0 partners from Fraunhofer (transfer of DFKI’s digital product memory technology).
Datagarten is a spin-off company of Fortiss GmbH. It is our most recent EIC-ICT Labs Innovation spinoff within the CPS for Smart Factories activity.
For solving distributed parallel computations in a scalable manner that can still undergo optimisation, ELTE proposes the Task Oriented Programming (TOP) paradigm.
The international programmer workshop will take place in Budapest on December the 1st, 2014.
DFKI unveils hand-writing and gesture based knowledge acquisition tool and the new smart factories knowledge portal at Siemens Headquarters, Munich. See architecture.
First Siemens patent filed!
DFKI – Siemens Partner Contract signed.
Kick-Off at EIT in München!
