Researchers from the planned Cluster of Excellence “Reasonable Artificial Intelligence” (RAI) and hessian.AI took part in the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), held in Seattle (USA) from June 17 to 21. As the premier computer vision event, CVPR attracts more than 10,000 participants annually.

RAI research papers received awards

Several workshop papers presented by researchers from the RAI project received special recognition:

The publication “Is Synthetic Data All We Need? Benchmarking the Robustness of Models Trained with Synthetic Images”, won a Best Paper Award at the SyntaGen – Harnessing Generative Models for SyntheticVisual Datasets workshop. The paper, which examines the robustness of models trained with real images compared to models trained with synthetic images, was presented by Krishnakant Singh, PhD candidate at the Visual Inference Lab (VISINF) at TU Darmstadt. Singh was accompanied by Independent Research Group Leader Simone Schaub-Meyer and Professor Stefan Roth.

A Best Student Paper Award was awarded to Christoph Reich, who presented the paper “A Perspective on Deep Vision Performance with Standard Image and Video Codecs” at the AIS: Vision, Graphics and AI for Streaming Workshop.

At the GenAI Media Generation Challenge Workshop, the paper “LEDITS++: Limitless Image Editing using Text-to-Image Models” achieved an outstanding second place. LEDITS++ presents an efficient yet versatile and precise textual image manipulation technique. The LEDITS++ research team, represented by Manuel Brack, Felix Friedrich, and Patrick Schramowski, was rewarded with a Meta Quest 3 headset.

Scientists from the RAI cluster project shared their expertise

Professor Justus Thies delivered two keynote speeches: “Digital Humans: Recent Advances in Capturing and Controlling Digital Humans – from 2D DeepFakes to volumetric avatars that can be used in immersive telepresence,” was presented at the Workshop on Media Forensics. The Populating Empty Cities – Virtual Humans for Robotics and Autonomous Driving workshop featured his keynote “Capturing and Controlling Digital Humans: Text-based appearance and motion control of volumetric human avatars using diffusion-based generative models.”

Professor Stefan Roth gave an Invited Talk on accelerating and evaluating explanation methods in computer vision during the SAIAD 2024 – Safe Artificial Intelligence for All Domains workshop.

Professor Kristian Kersting joined the conference remotely to present on Reasonable Artificial Intelligence (RAI) during the 3rd Explainable AI for Computer Vision (XAI4CV) Workshop. Kersting’s talk emphasized the need for both explainable and interpretable AI, illustrated by several examples.

Manuel Brack, researcher at the German Research Center for Artificial Intelligence (DFKI) and PhD candidate at the Artificial Intelligence & Machine Learning Lab at TU Darmstadt, presented an Invited Talk on “Enhancing Visual Content Safety: Multimodal Approaches for Dataset Curation and Model Safeguarding” as part of the Workshop on Multimodal Content Moderation (MMCM).

With six conference papers and five workshop papers, the researchers from the RAI project and hessian.AI presented a total of 11 papers at CVPR 2024. An outstanding achievement, considering that from 11,532 paper submissions, only 2,719 (23.6%) were accepted. Topics ranged from human face and body modeling and motion synthesis to text-driven image editing. “Our various contributions to CVPR 2024 show that we are part of the rapid development in computer vision. The enormous potential of this technology has been demonstrated by the diversity of the research presented,” remarks Simone Schaub-Meyer, Independent Research Group Leader, TU Darmstadt.

List of papers published at the CVPR 2024 main conference:

Visual Inference Lab (VISINF), led by Prof. Stefan Roth

3D Graphics & Vision Group (GraVis), led by Prof. Justus Thies

List of papers presented at CVPR 2024 workshops:

Artificial Intelligence and Machine Learning Lab, led by Prof. Kristian Kersting


Visual Inference Lab (VISINF), led by Prof. Stefan Roth

About the CVPR

Sponsored by the IEEE Computer Society (CS) and the Computer Vision Foundation (CVF), the Conference on Computer Vision and Pattern Recognition (CVPR) is the preeminent annual computer vision conference for students, academics, and industry researchers. It showcases significant advancements in all areas of computer vision and pattern recognition, as well as their impact on different fields and industries. This year’s conference attracted more than 12,000 scientists and engineers from 76 countries and regions. Featuring the main conference, an expo, numerous co-located workshops, and a variety of short courses, CVPR provides ample opportunity for networking, recruiting, inspiration, and motivation. Learn more

During a visit, Mario Brandenburg (FDP), Member of the German Parliament and Parliamentary State Secretary at the Federal Ministry of Education and Research (BMBF), got to know about the work of the Hessian Center for Artificial Intelligence (hessian.AI). Key topics included the ongoing research projects, the projects in the federal government’s Excellence Strategy competition and the AI innovation ecosystem with a wide range of spin-offs from hessian.AI.

Brandenburg met with the President of TU Darmstadt, Professor Tanja Brühl, and the co-directors of hessian.AI, Professor Mira Mezini and Professor Kristian Kersting, as well as scientists who conduct research at hessian.AI and provided insights into their work.

Kristian Kersting and Marcus Rohrbach (Professor of Multimodal Reliable Artificial Intelligence) presented “Reasonable Artificial Intelligence (RAI)” and “The Adaptive Mind (TAM)”, two of the three projects with which TU Darmstadt is represented in the competition of the Excellence Strategy of the German federal and state governments. The planned Cluster of Excellence RAI is dedicated to the development of a new generation of AI systems based on the rational use of resources, data protection and continuous improvement. This next generation of AI systems is intended to address some of the weaknesses of current AI systems. “The Adaptive Mind (TAM)” investigates human behavior under changing external conditions. TAM is developing new approaches to understanding and computationally modeling human perception, thinking, decision-making, action and learning.

Occiglot is intended to create a coherent language modeling system that takes into account all 24 official languages of the European Union as well as other unofficial and regional languages. The first version will initially focus on the five largest European languages: English, German, French, Spanish and Italian. With LEDITS++, images can be modified by entering text. Editing specific image regions ensures that the visual and contextual coherence of the images is maintained. During the visit, Brandenburg also gained practical insights into the work of the Robotics Institute Germany (RIG), an association of ten universities and four non-university research institutions from all over Germany, with which the expertise of robotics research is synergistically bundled and strategically strengthened. hessian.AI is also represented in this network.

Energy Robotics and Birdmapper, two startups at different stages of development from the ecosystem, completed the series of stations. Energy Robotics manufactures autonomous robots for the inspection of industrial plants and thus simplifies predictive maintenance, among other things. The aim of Birdmapper is to autonomously measure bird populations as a contribution to preserving biodiversity.

The Konrad Zuse Graduate School “Excellence in Learning and Intelligent Systems” (ELIZA), presented by Stefan Roth (Professor of Visual Inference), is a key component of the AI ecosystem for qualifying the AI experts of tomorrow.

The infrastructure provided in hessian.AI with the fortytwo supercomputer and the AI innovation laboratory and service center support AI research and application in Hesse in a unique way, as outlined by the CTO of hessian.AI, Dr. Wolfgang Stille.

“hessian.AI is the supporting pillar of a lively, dense and constantly growing AI innovation ecosystem around TU Darmstadt, in Hesse and with great appeal far beyond. hessian.AI strengthens and promotes excellent AI research and teaching at Hessian universities. In hessian.AI, innovative applications with transformative potential are created from disruptive ideas. The diverse insights gained during the visit show impressively that partners in science, business and society benefit from this strong network.”

Tanja Brühl, President of TU Darmstadt
hessianAI-Besuch-BrandenburgDownload

What problem would you want to solve with #AI? Under the motto “Change the world with your #AIdea,” the Open Idea Competition by hessian.AI challenges participants to propose solutions using AI.
The competition addresses people, who want to explore the possibilities of AI further to tackle challenges of a social or economic nature.

Key highlights:

Choice: Participants have the freedom to present their ideas to the public OR seek feedback from experienced mentors and to quitly develop their ideas further.

Support: Workshops and events during the application phase provide opportunities to refine project ideas and deepen understanding.

Opportunity: Pre-selection for video pitches and potential advancement to the finalist stage offer avenues for recognition and collaboration.

Key dates:

Application Phase

April 25 – July 2, 2024

Competition Phase Begins

July 3, 2024

Jury Meeting

Early September 2024

Award Ceremony

End of September 2024

The competition aims to drive positive impact and shape the future of AI. Ready to embark on a transformative AI journey? Join the Open Idea Competition and be part of a community striving to change the world with their #AIdeas. Just start and find out where your #AIdeas will take you.

Apply now

The shortage of skilled workers, the increasing flood of information due to digitalisation and new needs of citizens make it necessary for the state to explore the potential of AI. On the face of it, public administration has particularly good prerequisites for the use of AI: Many administrative processes are highly structured and the execution steps are usually clearly regulated. Last but not least, the public sector has a large amount of data at its disposal. This blog post shows where public administration is already using AI, what role start-ups play in this and where the challenges lie.

AI use cases in the administration

The application of AI in the public sector today goes beyond test trials and pilot projects. A wide variety of administrative areas are already using AI in operations. Systems based on machine learning support the verification of documents, such as study certificates, which may be necessary for applying for child benefit. At customs, image recognition is used to identify counterfeit products. In the internal IT infrastructure of public authorities, AI is used to predict the overloading of server landscapes or recognise anomalies in data streams. However, AI systems are also increasingly penetrating the ‘front office’ of the administration, where the state and citizens meet directly. In an international context, for example, AI is occasionally used to improve counselling for the unemployed. There are also many other applications of AI in the public sector; the EU alone had collected over 680 use cases in its member states by the end of 2021.

Startups help to realise the potential of AI in administration

AI can be used to automate processes and relieve employees of repetitive tasks. One example of a repetitive task is the processing of housing benefit applications in local authorities. A cognitive AI application from a company in Darmstadt can provide support here, enabling administrative staff to automate certain process steps independently, even without programming knowledge. However, a large number of complex decisions are also made in administration every day. For example, in statutory accident insurance: for every single accident at work or on the way to work, the responsible accident insurance provider has to check whether it is worth taking recourse against a third party liable to pay compensation. In collaboration with a data scientist startup, an AI-based system has been developed that identifies cases with a high probability of successful recourse proceedings.

With generative AI, the latest development in the field of artificial intelligence is also finding its way into public administration. The Baden-Württemberg state administration is already using a chatbot internally that supports employees in their text work using an AI-based language model. There are various government funding programs for the development of AI applications that are geared towards the common good.

Special framework conditions for the use of AI

Artificial intelligence offers many advantages for the administration. On the one hand, recurring activities that can be processed more efficiently with the help of AI can benefit. On the other hand, complex decision-making processes can be placed on a broader empirical basis with AI. Positive effects are expected both for administrative staff, who will be relieved of some of their workload, and for citizens, who will benefit from better government services. The administration is therefore already using AI in selected processes. However, the public sector is subject to special framework conditions in the fulfilment of its diverse tasks. High security and data protection requirements are reasons why it is likely to take some time before the potential of AI reaches the wider public sector. The European Commission’s AI Act will provide legal certainty in this context, and many are looking forward to its transposition into national law in the coming years.

Mareike Sirman-Winkler is a political scientist and researches artificial intelligence in public administration at the Weizenbaum Institute for the Networked Society. Her focus is on the normative conflicts, organisational challenges and behavioural consequences of AI-based innovations in social administration. She is doing her doctorate at the Institute of Social Sciences at the University of Oldenburg. Before that, she was a consultant in a federal ministry. The blog post reflects her personal opinion.

We’re on the brink of an exhilarating transformation in the AI landscape in Hesse and beyond, as the renowned AI Startup Rising Accelerator Program, “RISING,” unveils its latest cohort.
This year, RISING is set to catapult AI-centric startups into a new dimension of growth and innovation.

Meet the Game-Changers:

  1. Proteineer: Venturing into the depths of DNA sequencing with advanced deep-learned protein language models, they’re set to revolutionize protein discovery and accelerate research.
  2. Cyreen GmbH: Cyreen sets a new standard in in-store retail media by offering precise measurement and targeting opportunities to brick-and-mortar stores – basically a GDPR-compliant, offline web cookie.
  3. Persival GmbH: Specializing in perception sensor simulation software, their work is pivotal for AI-driven automated driving systems, expanding into various robotic and infrastructure applications.
  4. ​Veli GmbH: With a heart for elder care, they are revolutionizing safety at home based on the energy consumption data to enable a safe and worry-free life into old age.
  5. ANTICIPATE GmbH: Transforming manufacturing processes with an AI-based system to detect and reduce errors in manual assembly, enhancing efficiency and accuracy.

What’s in Store?

From November 2023 to April 2024, these startups will undergo an intense 5-month journey, equipped with a €10,000 prize, mentorship from AI experts, and exclusive, tailored workshops. They’ll gain invaluable insights in sales, VC-readiness, leadership, and storytelling, preparing them to navigate the challenges of scaling up and securing investments.

Connect and Grow:

This program isn’t just about acceleration; it’s a melting pot of innovation and collaboration. As these startups take on this transformative journey, they’re not just rising – they’re reshaping the future of AI.

Hessen Ideen Wettbewerb sucht zum achten Mal die besten Ideen aus den Hochschulen

Jetzt abstimmen: Online-Voting für hessischen Hochschulgründungswettbewerb gestartet
 
Im landesweiten Gründungswettbewerb von Hessen Ideen wetteifern in diesem Jahr 34 innovative Ideen um den Titel „Beste Gründungsidee aus einer hessischen Hochschule“. Im jetzt gestarteten öffentlichen Online-Voting können sich die drei Teams mit den meisten Stimmen einen Platz im Finale des Wettbewerbs sichern. Das Online-Voting läuft bis zum 19. Oktober. Zehn weitere Finalteams werden durch ein Gutachtenverfahren ermittelt.

Der Wettbewerb richtet sich an gründungsaffine Hochschulangehörige und Alumni aus hessischen Hochschulen, die sich noch in einem frühen Stadium der Ideenfindung für eine Gründung befinden. Ziel des Hessen Ideen Wettbewerbs ist es, die Gründungskultur an den hessischen Hochschulen nachhaltig zu stärken.

Alle hessischen Hochschulen waren aufgerufen, bis zu drei ihrer besten Gründungsideen für den Wettbewerb zu nominieren. 15 Hochschulen haben in diesem Jahr eine oder mehrere Ideen ins Rennen geschickt. Auch das Hessische Zentrum für Künstliche Intelligenz – Hessian.AI – durfte in diesem Jahr wieder Gründungsvorhaben nominieren.
 

Teilnehmende Ideen aus dem AI Startup Rising Netzwerk

Aus dem AI Startup Rising Netzwerk sind folgende Gründungsprojekte im Wettbewerb vertreten:

PHONT
ARCHIPLAN
MySympto
TecLex
I3DEnergy


Der Hessen Ideen Wettbewerb steht unter der Schirmherrschaft der Hessischen Ministerin für Wissenschaft und Kunst, Angela Dorn. Erstmals ausgeschrieben im Jahr 2016, wurden bislang 293 Gründungsideen für den Hessen Ideen Wettbewerb nominiert. 

Öffentliches Online-Voting ist gestartet

Auf der Webseite von Hessen Ideen werden alle Ideen vorgestellt, die bis zum 19. Oktober Stimmen sammeln können, um in das Finale einzuziehen: Hier geht es zum Online-Voting. Die bis zu 13 Teams, die durch das Gutachtenverfahren und das Online-Voting in das Finale einziehen, pitchen vor einer Jury, die durch ihre Bewertung die Sieger*innenteams ermittelt.

Preisverleihung in Frankfurt

Bei der feierlichen Preisverleihung am 23. November pitchen die Finalist:innen noch einmal vor dem Publikum, anschließend werden die Gewinner:innen bekannt gegeben. Insgesamt 10.000 Euro Preisgeld werden auf die ersten drei Plätze verteilt. Zusätzlich erhalten alle Finalist:innen ideelle Unterstützung durch ein Pitch-Training, professionelles Feedback sowie 500 Euro Roadshow-Budget für Messeauftritte oder Marketing. Interessierte, die gern bei der Preisverleihung in Frankfurt im Publikum dabei sein möchten, können sich hier bereits anmelden.

Die Initiative Hessen Ideen

Hessen Ideen ist eine Initiative des Landes Hessen, der hessischen Hochschulen und hessischer Unternehmen. Neben dem Hessen Ideen Wettbewerb werden unternehmerische Ideen mit den Säulen Hessen Ideen Stipendium, Hessen Ideen Crowdfunding und dem Hessen Ideen Hochschulnetzwerk entdeckt und gefördert. Die Gründungsförderungen der Hochschulen arbeiten dabei in einem Netzwerk Hand in Hand zusammen, um den Gründer:innen eine möglichst große Bandbreite an Unterstützungsangeboten bieten zu können. 
 
Die Initiative Hessen Ideen wird von UniKasselTransfer an der Universität Kassel in Kooperation mit HIGHEST von der Technischen Universität Darmstadt koordiniert.

On September 1, 2023, in the beautiful evening setting of Darmstadt at Woogs-Cafe, we successfully hosted our first MedAI Roundtable.

Our mission was simple but powerful: unite MedAI startups to identify shared needs and spark innovation. With no formal agenda, startups introduced themselves, sharing their focus and highlight their topics of interests.

The evening buzzed with electric discussions among startup guests. Delectable food, drinks, and ambiance led to inspiring connections.

We are thrilled to report that numerous connections were forged, and discussions about potential collaborations are already underway. In fact, two groups have already begun working on a joint project together, and they’re eager to dive into the details in the coming days!

Thanks to the 22 participants from 10 startups and 2 companies: Multilingo Medical AI, MySympto, Deep LS, PreScriptAI, EinMind, mediEM, KardioIQ, ViSPA, Green Convenience, UnderstandUrself, Eloquest, and Innoplexus. Your enthusiasm was truley inspiring.

And special thanks to our AI startup coach Phillip Travers for designing and organizing the event!

Stay tuned for more MedAI innovation and collaboration updates!

Last Friday, August 30, 2023, marked a major milestone as we successfully hosted an innovative Proof-of-Concept Workshop (PoC) for pioneering AI Startups. This initiative by AI Startup Rising | hessian.AI aimed to empower pre-incorporated AI Startups who participated in our Open Idea Competition or received our Lean AI Startup Funding. The workshop aimed to prepare these startups for the upcoming AI Startup Competition starting on October 12, 2023. Stay tuned for more information!

Our journey began during a previous AI startup competition, where we identified a gap in supporting startups targeting B2B and B2G markets. PoC Projects emerged as a way to break into these markets, but challenges persisted. We addressed key questions about navigating this path pre-incorporation, managing intellectual property, taxation, and gaining customer trust.

The workshop, featuring legal experts, tax consultants, and a successful startup, received positive feedback from 10 participating dynamic startups. Their involvement validated the importance of these topics. Kudos to all participating startups for their valuable insights: MySympto, ArchiPlan, CloudMent, mindwaves, copyrAight, UnderstandUrself, MediEM, Bird Mapper, FLIPoQ, PrescripAI and Phont.

Special thanks to legal expert Tim Steininger, tax consultant Jürgen Jaschke, and Marco Limm from cre[ai]tion Gmbh for sharing your expertise with us and Phillip Travers, our AI Startup Coach, for designing and organizing the workshop.

The workshop’s success inspires us to do more! We’re excited to announce that we’ll host this workshop again during the AI Startup Competition 2023/24.