Imagine: Can machines compose? A team of AI experts from hessian.AI and Aleph Alpha as well as musicians from the Singakademie Dresden have dared the exp­eriment: on 20 November 2022, the piece “Meistersinger Reloaded”, composed by artificial intelligence, was premiered.

Roughly a year ago, the conductor Michael Käppler from Dresden approached technical universities in Germany with an unusual request: He had in mind an AI that composes music. Kristian Kersting, the founding co-director of the Hessian Research Centre for AI (hessian.AI) and professor of computer science at TU Darmstadt, was enthusiastic about the idea. He put Käppler in touch with researchers and students from his field, who started to work together with the musician. With their project, they have now reached a wider audience in Dresden and beyond.

“Meistersinger reloaded” was performed on 20 November 2022 in the Lukaskirche Dresden by the Elbland Philharmonie and the Great Choir of the Singakademie under the direction of choirmaster Käppler with a total of around 130 musicians. The AI composition was embedded in a programme of pieces by Franz Schubert, Felix Draeseke, Richard Wagner and Alexander von Zemlinsky. Participants were surprised at how harmoniously and pleasingly the composition blended into the concert as a whole. Since there was no signal like an acoustic beep and the order of the announced pieces changed shortly before the performance, only a few managed to solve the acoustic riddle. None of the pieces were out of the ordinary or sounded surprisingly different, which was perhaps the most surprising thing about this evening that brought the project to its artistic and scientific climax.

Cooperation with Aleph Alpha and the Dresden Singakademie

Over the course of a year, a team of computer scientists from the AI R&D company Aleph Alpha at Heidelberg, from the TU Darmstadt and musicians from the Singakademie Dresden created an AI composition about Wagner’s music: In the AI data center of the Heidelberg AI researchers, MIDI files on piano music of the German Romantic period were used to train a large AI transformer model, which independently learned basic rules of composition on the large data set and then received special training on Wagner’s works to be able to imitate his style. Koen Oostermijer and the research team from Aleph Alpha, Michael Lang and Wolfgang Stammer from the TU Darmstadt, as well as artistic director Käppler were in close exchange over many experiments, jointly evaluating different experiments, approaches and iterations until their Wagner AI seemed perfect. It was helpful that large language models, such as Aleph Alpha is developing, have already internalised a similar pattern recognition: Music has a temporal structure like language, but some details of the modelling had to be adapted. After successful training, the AI now continues to write notes for any piece, using the learned structure and style of music in general and Wagner in particular.

“The first piece from the fifth generation was special,” Käppler revealed. It had taken hold of him. The researchers continued to work with this piece, which had something like a motif and most closely resembled a composition, until in the end a piece emerged that Käppler arranged for the orchestra and he adapted the text for the choir to it. Choral music requires longer notes to carry the language: subtleties that the researchers had to take into account when providing input to the AI. For the AI researchers, the project was an exciting balancing act between influencing the system to produce a desired output and leaving it free to produce something creative, playable and worthwhile to listen to that would also please human ears. The three-minute piece contains over 3,000 hours of music that were used for training, as well as many hours of human listening samples. Käppler has not altered the musical substance of the AI composition, but has taken it over for the orchestra as faithfully as possible.

A Feast for the Ears: Can Machines be Creative?

A discussion panel before the concert provided information about the background of the project and its practical foundations. Many concert guests took the opportunity to get in the mood and came to the community centre to learn about AI and art creation. Cultural curiosity united the audience, some guests had a technical background. Michael Käppler, Kristian Kersting, the mathematician and emeritus music professor Christfried Brödel and the Darmstadt AI master’s student and classical musician Matthias Lang discussed with the musicologist Miriam Akkermann (TU Dresden). 

The starting point for Kersting and the machine learning researchers from Aleph Alpha and the TU Darmstadt was the question of how music can be formalised in order to produce it by machine. Composition with artistic pretensions is the creation of something new according to existing rules – recognisable, but not too close to the existing, i.e. not mere “parroting”, as Käppler explained. Can that work without emotion? After all, machines presumably don’t have feelings. Ultimately, art is “the encounter of people, made for people”, as Christfried Brödel put it. People remain the recipients of machine-made art, and it should speak to them. 

AI is an Emotionally Charged Field

The project had moved in the field of tension between technology and human emotions. The Dresden concert audience reacted with curiosity and openness: obviously no one found anything strange about the performance, and at the end there was applause and flowers for the computer scientists as well. The audience treated them like part-composers. Perhaps that is precisely what is original about it and an answer to some of the open questions: Wagner had considered himself an unsurpassable genius, his musical legacy is overshadowed by anti-Semitism and the closeness of some of his descendants to the Nazi regime as well as the reception of his music during that time. AI “knows” nothing of this and can “concentrate” on the formal aspects of Wagner’s music. 

AI is an emotionally charged field: People react to machine learning with fears, hopes and fascination. Fears, because people feel redundant when machines can do something better. Hopes, because practical assistance systems already accompany us in everyday life. The big problems of humanity will probably only be solvable with superhuman intelligence. But can machines also create art, can creativity be attributed to machine production? And if so, what does that mean for us? Some fears, insofar as they existed, were certainly dissolved by the evening and transformed into informed curiosity. The concert was well received, the church was crowded.

Art and Computer Science Appeal – Together 

On the technical side, the joint project created a complex AI system optimized for musical art – similar to the one at the base of Aleph Alpha’s multimodal models. On the human side, it brought and continues to bring art and research together for a sensual form of interaction between man and machine. Wagner’s music was not disenchanted at the concert, but renovated for today. In the Meistersingers, an unshakeable belief in art as the basis for meaning and cohesion is expressed, says Käppler. We can all do with some of this spirit (purged of its nationalistic overtones). AI makes a contribution here to contemporary Wagner reception.

hessian.AI – the Hessian Center for Artificial Intelligence together with the CAISA Lab of Philipps-University Marburg hosted the IT Summer School Women4Women from August 22 to September 3. Lucie Flek, hessian.AI member, launched the initiative to get young women excited about the field of artificial intelligence and to show them future career paths.

The Women4Women IT Summer School was aimed at high school girls who already have an initial interest in STEM subjects. The field of computer science was introduced to the female students in a confidential and relaxed atmosphere at Philipps-University Marburg. The aim was to provide initial insights into career prospects in computer science, both in research and in technology and commercial companies in the Marburg area and throughout Hesse. The primary goal is the practical teaching of basic elements of programming, working together in a team and especially having fun while working on the versatile tasks.

With five participants, the event took place in a personal and confidential atmosphere. The students were able to talk with like-minded people about difficulties regarding their professional future: They were alone in their circles of friends with their interest in computer science, furthermore it was known that sexism/discrimination in the field of IT and gaming was a current problem and women were still “underestimated in these fields, role models were missing, to name just a few of the concerns of the female students”.

Hosting the Summer School serves to try to address these concerns: Early exposure to new technologies is intended to foster young women’s interest and create understanding of the areas of application. The hands-on sessions at the event helped to remove initial inhibitions.

The schoolgirls were accommodated in Marburg for the entire event. They were introduced to programming in a playful manner and by creating concrete applications. Already on the second day, they learned to develop a Pacman game. The highlight was the programming of the humanoid robot “Pepper”. This robot is programmed to analyze people and their facial expressions and gestures and to react accordingly to their emotional states.

Participants developing a Pacman game

In addition to the content of the program, the students got to know Marburg better. Together with their tutors, the participants went pedal boating on the Lahn River, took part in a city tour entitled “Famous Women in Marburg” and visited the Botanical Garden.

The percentage of women in technical professions, including those in the AI environment, has increased in recent years, but the gender ratio is still unbalanced. Only just over 20 percent of people working in AI-related fields are women, according to business magazine Forbes. A recent FAZ article highlights misalignments in scientific institutions that make it difficult for women to pursue careers in the system.

That makes it all the more important to show young women the way to take away their fears – it’s worth it:

It was a nice experience to know that there are more girls who share the same interests after all. It encouraged me to continue on the science path, as I had my reservations about it, too.

Feedback from a participant

To encourage the growth of women in the AI industry, hessian.AI aims to promote female talent in the field, make career direction more accessible to them and raise awareness of the need for gender balance.

The complete tutorials with code and slides are posted on Github.

hessian.AI funds four exciting projects

hessian.AI annually awards the Connectom Networking Fund to projects in which hessian.AI members are significantly involved.

The Connectom Networking Fund provides seed funding for research collaborations between hessian.AI members and other colleagues at and between participating universities. In a competitive process, time-limited projects from across the spectrum of research, teaching, training or application are funded with a maximum of EUR 40,000 per project.

In the third round of calls, the Hessian Center for Artificial Intelligence has approved four great and innovative projects.

Congratulations to Dr. Ivan Habernal and Prof. Christoph Burchard with Enabling Multimodal and Multilingual Argument Mining in Court Hearings!

We also congratulate Charles Welch, PhD. and Dr. Ivan Habernal to the project Reliable Anonymization Methods for Enabling Language Understanding Models to Assist with Air Traffic Controller Communication.

On the KIDER project, congratulations go to Prof. Michael Guckert and Prof. Gerd Manthei.

For the continuation of the AI4Birds project with AI4BirdsDemo we congratulate Dr. Markus Mühling, Prof. Nina Farwig and Prof. Bernd Freisleben.

Read more about the Connectom Networking Fund and the already approved and ongoing projects here.

Anyone who owns a smartphone probably uses mobile apps. These everyday helpers track, push and accompany us from morning to night. In order to adapt to users and run their services, they need access to extensive private information. Criminals can access this data via malicious apps and process it to the detriment of users who may not even be aware of it. Consequently, very sensitive information, such as access data or data on financial transactions, can be tapped with deceptive or missing protection.

With its platform, the start-up project Queryella enables the in-depth inspection of mobile apps for security vulnerabilities and data protection breaches in order to prevent this from happening.

Queryella uses machine learning to analyze apps for security vulnerabilities and unauthorized, or undetected, data flows. The security vulnerabilities or threats are often contained in “hidden code” that is obfuscated. The platform automatically detects whether such obfuscated code is contained based on ML.

The analysis of apps is just the beginning. Further development will focus on business software, such as CRM systems, and thus promises broad application possibilities.

The originators of the ideas, Dr. Leonid Glanz, Dr. Lars Baumgärtner, Patrick Müller and Florian Breitfelder, come from the research team of Prof. Mira Mezini, hessian.AI Co-Director and PI at the Department of Computer Science at TU Darmstadt.

About the European Cybersecurity Month

Since 2012, European Cybersecurity Month (ECSM) has been a pan-European format to promote cybersecurity. Every Thursday in October 2022, hessian.AI, the Hessian Center for Artificial Intelligence, and the Department of Computer Science at TU Darmstadt provide insights into how artificial intelligence and cybersecurity successfully interact and benefit from each other.

On October 11, 2022, this year’s annual conference of efl will take place at the Westend Campus of Goethe University. We hereby cordially invite you to attend. The efl Annual Conference 2022 is a special event, as it is also the 20th-anniversary celebration of the efl. 

The focus of the annual conference will be on the area of “Democratizing of Data Science and AI” – that is, making Data Science and AI easy to use. The invited talks cover a wide range of topics, including automating typical machine learning (ML) tasks, facilitating access to ML, or modern visualization techniques that are important for the user experience and transparency of data science applications. In addition, ethical aspects of AI will be discussed. More details about the program can be found here: https://www.eflab.de/annual-conference-2022 .

The event is planned as an “in-person” event. If you would like to attend the event, please register (free of charge) by 23.09.2022 via the following link the latest: https://www.eflab.de/annual-conference-2022. We look forward to seeing you. 

The Hessian Center for Artificial Intelligence, hessian.AI, and the organization “KI macht Schule” conducted a series of courses on artificial intelligence at Leibnizschule Offenbach from July 18 to 20, 2022. The series of events was very well received by teachers and students.

As part of a project week at Leibnizschule Offenbach, hessian.AI and KI macht Schule held three courses in artificial intelligence and machine learning. The courses focused on how Artificial Intelligence can be applied in the subject areas of medicine, ethics and art.

The courses focused on the basic functioning of Artificial Intelligence, its impact on different fields of application, as well as ethical and social components of AI.

In cooperation with the Hessian Center for Artificial Intelligence, we were able to present our three state-of-the-art course modules AI & Medicine, AI & Art, and AI & Mobility to the students of Leibnizschule Offenbach. By collaborating with we come a big step closer to the goal of KI macht Schule to provide all students with a sound basic education in the field of AI and its diverse application areas.

Marius Süßmilch, Head of the Heilbronn office of KI macht Schule

Ten participants from each grade 10 were able to train a neural network themselves by making training data available that was classified by the AI and thus made initial contact with the concept of machine learning. The interested students took pictures of themselves with and without masks, for example. The application classified the data and could then judge whether a person was wearing a medical mask correctly or not. In addition, the students playfully devised their own conditions with which they trained the neural network.

AI systems like the ones the participants were able to test are already part of our everyday lives and are already an integral part of many industrial processes. To ensure that even schoolchildren* understand the impact of this technology, but can also participate in shaping it, hessian.AI has set itself the goal of firmly integrating the acquisition of basic skills in the field of artificial intelligence into the curriculum of Hessian schools.

It is great that through the cooperation with KI macht Schule we have the chance to bring the topic of artificial intelligence to Hessian schools with experts. In this way, hessian.AI wants to make a contribution to spreading the understanding of this technology and to start where it will reach the target groups of the future: in the field of education.

Elena Stahl, Press & Media Relations Officer, hessian.AI

In order to create acceptance and understanding for new technologies and also their already current use, an examination of these areas must be fundamentally guaranteed. This should include technical-scientific as well as ethical and social aspects in order to clarify facts, chances and also existing deficits of artificial intelligence. Only through “participation” in the topic can the willingness to engage with AI innovations be promoted.

On the one hand, the school-based discussion and the first taster of the practical handling of artificial intelligence should awaken the students’ interest and understanding of what this technology is, what it can do and where it is used. On the other hand, the first practical contact is also intended to ensure that this area represents an attractive (?) career perspective for the students and to impart initial skills even before their studies, as well as to remove the inhibitions of girls in particular to pursue a technically oriented career path. It is also intended to counteract the shortage of skilled workers.

About the cooperation

Since the beginning of the year, hessian.AI and KI macht Schule have been working hand in hand to bring knowledge about artificial intelligence to Hessen’s schools. Six courses have already been held successfully, and further courses are being planned.

hessian.AI – Hessian Center for Artificial Intelligence

The Hessian Center for Artificial Intelligence, which is in the process of being founded, pursues the goal of conducting excellent basic research with concrete practical relevance and also of promoting transfer to industry and society. The center, in which 13 Hessian universities are involved, bundles the expertise of 22 AI scientists and expands it through 22 new professorships.
Based on cutting-edge research as conducted at the TU Darmstadt and other Hessian universities, significant contributions are made to the research and development of novel AI systems with human-like thinking and communication capabilities, which are attributed to the so-called “Third Wave of AI”.

KI macht Schule

KI macht Schule gUG (haftungsbeschränkt) brings AI education to German schools with over 70 PhD students, undergraduates and young professionals in eight local groups – both on-site in the classroom and online. In addition to courses directly for students, further training is also provided for teachers.

Die Software AG hat heute ein neues Mitglied ihres wissenschaftlichen Beirats vorgestellt: Professor Dr.-Ing. Dr. h.c. Mira Mezini gehört dem Gremium mit sofortiger Wirkung an und berät das Unternehmen rund um das Thema Softwaretechnologien.

Prof. Mezini ist ordentliche Professorin für Informatik an der Technischen Universität (TU) Darmstadt und leitet die Gruppe Softwaretechnologie am Fachbereich Informatik. Nach ihrer Promotion in Informatik an der Universität Siegen verbrachte sie zwei Jahre als Visiting Assistant Professor an der Northeastern University in Boston (USA), bevor sie im Jahr 2000 an die TU Darmstadt wechselte. Sie hatte beziehungsweise hat Gastprofessuren an der Lancaster University (UK) von 2013 bis 2016 und an der University of Lugano (Schweiz) im Jahr 2022 inne. Von 2013 bis 2014 arbeitete sie als Dekanin des Fachbereichs Informatik, von 2014 bis 2016 war sie Vizepräsidentin für Wissens- und Technologietransfer und von 2017 bis 2019 Vizepräsidentin für Forschung und Innovation an der TU Darmstadt.

Die Forschungsschwerpunkte von Prof. Mezini sind unter anderem Programmierparadigmen und -sprachen für zuverlässige dezentralisierte Softwaresysteme, Code-Intelligenz und Entwicklungsmethoden für KI-Softwaresysteme. Für ihre Forschung erhielt sie verschiedene internationale Auszeichnungen, darunter zwei IBM Eclipse Innovation Awards, einen Google Research Award, den Deutschen IT-Sicherheitspreis und im Jahr 2012 einen hochangesehenen Advanced Grant des Europäischen Forschungsrates (ERC). Sie ist Mitglied der Akademie der Technikwissenschaften (acatech) seit 2016 und des Senats der Deutschen Forschungsgemeinschaft seit 2022.

Prof. Mezini wurde aufgrund ihrer exzellenten wissenschaftlichen Laufbahn und ihrer herausragenden Expertise in Softwaretechnologie, einem Kernthema für die Software AG, für den wissenschaftlichen Beirat vorgeschlagen.

Die Aufgabe des Gremiums ist es, der Software AG die wissenschaftliche Perspektive auf Technologietrends zu geben – und so externe Impulse aufzunehmen. Davon profitieren in erster Linie die Kunden des Technologiekonzerns: Der
wissenschaftliche Forschungsdiskurs komplementiert die strategische Entwicklungs- und Produktplanung der Software AG. Zu den Mitgliedern des Beirats zählen Vertreter aus Wissenschaft und Forschung. Der wissenschaftliche
Beirat nimmt eine beratende Funktion ein und handelt dabei nicht als gesellschaftsrechtliches Kontrollorgan.

 „Ich freue mich sehr über die Aufnahme in den wissenschaftlichen Beirat der Software AG und die dadurch entgegengebrachte Anerkennung für meine Forschungsleistungen und für meine Expertise in der Softwaretechnologie. Durch die Tätigkeit im Beirat erhoffe ich, Impulse aus der Forschung im Bereich der softwaretechnischen Methoden und Softwarewerkzeuge zur Entwicklungs- und Produktplanung der Software AG beizutragen. Umgekehrt erwarte
ich, spannende Einsichten in die industriellen Softwareentwicklung zu erhalten, die wichtige Impulse für meine Forschung liefern können“, sagte Frau Professor Mezini anlässlich ihrer Ernennung.

„Mit der Berufung von Kollegin Mezini verstärkt sich unser Beirat nicht nur in einem zentralen Themenfeld der Software AG, sondern gewinnt auch eine Expertin für den gezielten Transfer von neusten Forschungsergebnissen in die industrielle Praxis“, fügte Professor Wahlster als Beiratsvorsitzender hinzu.

Dr. Stefan Sigg, Chief Product Officer und Mitglied des Vorstands der Software AG, ergänzt: „Software Engineering ist das Kerngeschäft der Software AG. Ich freue mich sehr, dass Prof. Mezini sich bereit erklärt hat, unserem Beirat beizutreten und eine moderne ‚Kunst der Computerprogrammierung‘ und eine neue, ganzheitliche Definition von ‚großartig‘ für Softwareingenieure zu prägen.“

Der wissenschaftliche Beirat hat sich erstmals am 17. Juli 2017 konstituiert und tagt vier Mal im Jahr. Er unterstützt die Aktivitäten der Software AG in Forschung und Entwicklung, unter anderem mit den Schwerpunkten Artificial Intelligence, Security, Quantum Computing, Internet of Things, Software Engineering & Quality. Über die Schwerpunktsetzung entscheidet der Beirat mindestens einmal pro Jahr.

Dem Beirat gehören zwischen sechs und acht externe Mitglieder für eine Amtsdauer von mindestens drei Jahren an. Zu den aktuellen Mitgliedern zählen folgende Personen aus Wissenschaft und Forschung:

Von Seiten der Software AG:

Picture: Prof. Mira Mezini, TU Darmstadt, and Dr. Stefan Sigg, CPO Software AG

Prof. Jan Peters brings first Amazon Research Award to hessian.AI and TU Darmstadt

For a research project on “Learning Robot Manipulation from Tactile Feedback”, the Autonomous Intelligent Systems department of computer science professor and hessian.AI founding member Jan Peters has been awarded an Amazon Research Awards (ARA) 2021. As announced by the company on 18 July 2022, the team will receive funding of approximately $95,000, as well as access to selected Amazon research infrastructure. The award will support the work of one to two PhD students or postdocs for one year.

Manipulation of mechanical objects is essential for real-world robotic applications from industrial assembly to household robots: robots must, for example, be able to position components precisely or hand over objects carefully. To date, there is no robot that can match the manipulation capabilities of even an unskilled human.

To date, there is no robot that can match the manipulation capabilities of even an unskilled human.

Prof. Jan Peters

To advance the manipulations skills of robotics, the Darmstädter team plans to combine high-fidelity tactile sensing, machine learning and their robust modular grip controllers to learn dexterous manipulation policies. This combination enables advancing the state of the art of robot manipulation and get one step closer to automating complex assembly tasks. 

With the help of this Amazon Research Award, Jan Peters and his team aim to advance so-called grip stabilization controllers and augment them with a hierarchical reinforcement learning. The goal for the robot ist to pick up an object, re-orient it and insert the object into an opening. The obtained manipulation policies are going to be evaluated on different objects with different shape and weight.

Among the best

The Amazon Research Award funds research projects in a variety of research areas relevant to Amazon, including robotics, machine learning, security, sustainability and more. The 74 awardees from the 2021 autumn call, which have now been published, represent 51 universities in 17 countries.

Jan Peters, is the first ARA award winner from hessian.AI and TU Darmstadt. The fact that his cutting-edge research is internationally competitive is also shown by the fact that TU Darmstadt is the only German university to be represented among the current award winners alongside other globally renowned universities such as ETH Zurich, the Massachusetts Institute of Technology (MIT), Carnegie Mellon University or the University of California, Berkeley.

Picture © Katrin Binner

Zur Förderung von Postdoktorand*innen mit dem Karriereziel Professur hat hessian.AI drei DEPTH-Nachwuchsgruppen eingerichtet, die von jeweils einem Forschenden geleitet werden.

Mit Dr. Martin Mundt, Dr. Devendra Dhami und Dr. Charlie Welsh hat hessian.AI in einem kompetitiven Auswahlverfahren drei herausragende Forschende für die Leitungen von DEPTH-Nachwuchsgruppen bestellt. Das DEPTH-Programm dient der Förderung von Postdoktorand*innen mit dem Karriereziel Professur: Ausgezeichneten Nachwuchswissenschaftler*innen wird die Gelegenheit geboten, unabhängig zu forschen und zu lehren. DEPTH-Nachwuchsgruppenleiter*innen verfügen über eine herausragende Promotion sowie über Postdoc-Erfahrung von in der Regel bis zu 4 Jahren. Sie sind wissenschaftlich nach internationalen Qualitätsmaßstäben ausgewiesen und verfügen über die potentielle Fähigkeit zur Übernahme von Lehr- und Leitungsaufgaben.

Weitere Informationen zu den DEPTH-Nachwuchsgruppen und ihren Leitungen

Martin Mundt, DEPTH-Nachwuchsgruppenleitung “Open World Lifelong Learning (OWLL)”

Die OWLL-Nachwuchsgruppe zielt darauf ab, neue Methoden zu entwickeln und Teilbereiche miteinander zu verbinden, um die nächste Generation von KI-Systemen zu schaffen, die in einer „offenen Welt“ kontinuierlich lernen können. Solche Systeme können nicht nur im Laufe der Zeit aus Erfahrungen lernen, sondern auch erfolgreich neue Situationen erkennen und aktiv Daten zum Trainieren auswählen, während sie sich autonom auf robuste und interpretierbare Weise anpassen.

Um diese Ziele zu erreichen, wird sich das OWLL auf Forschungsarbeiten konzentrieren, die eine ganzheitliche Perspektive auf sich entwickelnde Daten, adaptive Modelle und sich ändernde Bewertungsanforderungen von vornherein in den Entwurfsprozess einbeziehen.

Devendra Singh Dhami, DEPTH-Nachwuchsgruppenleitung “Causality And neUro-Symbolic artificial intelligence (CAUSE)”

Devendra Singh Dhami ist unabhängiger Nachwuchsgruppenleiter für “Causality And neUro-Symbolic artificial intElligence (CAUSE)” innerhalb des Hessischen Zentrums für Künstliche Intelligenz (hessian.AI) und der TU Darmstadt. Seine Forschungsinteressen sind vielschichtig und konzentrieren sich derzeit darauf, maschinelle Lernmodelle über Korrelationen hinaus in Richtung Kausalität zu bringen und dabei neue Methoden zu entwickeln und Teilbereiche miteinander zu verbinden, um Systeme zu schaffen, die nicht nur lernen, sondern auch erfolgreich Schlüsse mit Deep-Learning-Systemen ziehen können.

Charles F. Welsh, DEPTH-Nachwuchsgruppenleitung mit dem thematischen Bezug zu „Grenzen der anpassbaren Spracherzeugung und Wahrnehmungsmodelle“

Charles Forschung konzentriert sich auf die Grenzen der anpassbaren Spracherzeugung und Wahrnehmungsmodelle. Er interessiert sich dafür, wie Textattribute wie Toxizität und Empathie interagieren und zur gegenseitigen Verbesserung genutzt werden können, und wie diese Techniken zur Untersuchung von Bevölkerungsgruppen eingesetzt werden können. Durch eine bessere Modellierung sowohl von Menschen als auch von Spracheigenschaften möchte er die Sprachtechnologie nützlicher machen und ein besseres Verständnis dafür entwickeln, wie verschiedene Gruppen von Menschen die Welt sehen. Mit einem besseren Verständnis dafür, für wen wir Probleme lösen, können wir Modelle nachhaltiger trainieren, die Umweltkosten senken und eine integrativere Forschungsgemeinschaft unterstützen.

Einladung zur Eröffnung der Ausstellung am Donnerstag, 14. Juli ab 18 Uhr

“Die schwarzen Zonen in meiner Arbeit sind die Bereiche der Bilder, die von der KI nicht verstanden wurden, zu ihnen liegen keine Daten vor. Ich verstehe dies als Metapher, wie nämlich unser Bewusstsein und die Kenntnis von unserer Umgebung und der Welt funktioniert. Selbst wenn wir ein noch so gestochen scharfes, hochauflösendes Bild haben (oder erzeugen wollen), so fehlt uns immer ein Teil der Informationen, uns fehlen immer weitere Daten, und unsere Wahrnehmung der Realität ist darum auch immer voller schwarzer Zonen.”

Álvaro Rodriguez Badel

“Seit Menschengedenken neigen wir dazu, unser eigenes Bild von Sachverhalten zu kreieren, das aus Annahmen, Beobachtungen und Schlussfolgerungen besteht. Diese können beliebig unvollständig oder fehlerbehaftet sein, werden jedoch in der Regel zu einer in sich konsistenten und selbstbewussten Theorie zusammengeführt. Die dritte Welle von KI-Systemen wird selbständig erkennen, wenn sie nichts wissen, und somit falschen Entscheidungen entgegenwirken, sich anpassen und weiterentwickeln. Ziel einer neuen Generation von künstlicher Intelligenz sollte also sein, die blinden Flecke menschlicher Intelligenz auszuleuchten und als Partner der Menschen zu fungieren. Dies kann uns auch dabei helfen, Natur umfassender zu verstehen, um mehr im Einklang mit ihr zu leben.”

Dr. Wolfgang Stille,
CTO hessian.AI

Artist-in-Science-Residence

Zur Eröffnung der Ausstellung des Stipendiaten Álvaro Rodrigues Badel laden wir Sie herzlich in das Atelierhaus LEW1 auf der Rosenhöhe ein!

In Zusammenarbeit mit dem Hessischen Zentrum für künstliche Intelligenz (hessian.AI) setzt sich Álvaro Rodriguez Badel mit Bildmaterial auseinander, das er in den letzten Jahren im kolumbianischen Amazonasgebiet angefertigt hat. Aus unzähligen Fotos der Flora und der Topografie des Regenwaldes, die nach fotogrammetrischen Gesichtspunkten angefertigt wurden, entwickelt er mit Hilfe von KI-Software 3D-Modelle, von denen er Ausschnitte als großformatige Farbdrucke, Videos, Virtual Reality-Szenen oder 3D-Drucke präsentiert.

Unmittelbar vor seiner Residence in Darmstadt reiste Álvaro in den Chiribiquete Nationalpark in Kolumbien und fotografierte dort die zum Teil über 15.000 Jahre alten Felsmalereien, die Teil des dortigen UNESCO-Weltkultur- und Weltnaturerbes sind.

Auf Basis seiner Fotografien und Videoaufnahmen untersucht Alvaro neue Möglichkeiten der Rezeption und Interpretation dieses Ökosystems:

Er erzeugt 3D-Darstellungen des Amazonas sowie traumähnliche Bilder, indem er tiefe neuronale Netze für deren Generierung einsetzt. 

Mit Hilfe der Forschenden von hessian.Ai nutzt und verbessert Alvaro die neuronalen Netzwerke. Er versteht sie als ein Werkzeug zur Erweiterung von Kreativität und Bewusstsein, fokussiert dabei auf den Dialog zwischen indigener Kosmovision und Technologie. Die

 entstehenden Kunstwerke sind Proposition für eine neue Art der Naturwahrnehmung durch digitale Technologien. Ein neuer Raum wird angeboten, in dem wir unser Verständnis für die Beziehung zwischen digitaler und physischer Welt spielerisch auf die Probe stellen können.

Während seiner Residence im LEW1 beschäftigte sich Álvaro vor allem mit den Fragen, wie das Wissen einer KI in die künstlerischen Schaffungsprozesse einfließt und sichtbar wird, welche kulturelle Prägung KI-Modelle haben, woher sie stammen und welche Konsequenzen sie für eine KI-basierte Interpretation der Welt und der menschlichen Kultur haben. Álvaro versucht außerdem, offensichtliche Fehlinterpretation zu korrigieren – eine wichtige Aufgabe des “human in the loop”.

Álvaro Rodriguez Badel »A Breath At The Edge Of Future«

Eröffnung am Donnerstag, 14. Juli, 18 – 20 Uhr und 22 – 24 Uhr
im Atelierhaus LEW1, Ludwig-Engel-Weg 1, Darmstadt
sowie von 19.30 – 22.00 Uhr am Osthang, Mathildenhöhe.

Die Ausstellung im LEW1 ist vom 15. bis 18. Juli jeweils 11 bis 17 Uhr geöffnet.

The forest from the Yavarì river
©  Álvaro Rodriguez Badel

“Kunst kann unsere Beziehung zur Umwelt verändern, indem sie eine neue Perspektive auf die Welt um uns herum eröffnet. Sie kann uns die Schönheit der Natur vor Augen führen und uns zeigen, wie wichtig es ist, sich um unseren Planeten zu kümmern. Aber sie kann definitiv auch weitreichende Fragen und Bedenken aufwerfen. Meine Arbeit schlägt einen kontemplativen Weg vor, in dem neue Technologien und die Natur in Beziehung gesetzt werden und gleichzeitig versucht wird, den Prozess der bloßen menschlichen Wahrnehmung einzusetzen und zu erforschen.”

Álvaro Rodriguez Badel

“Generative KI-Modelle können heutzutage Texte, Bilder oder Klänge erzeugen, die analog zur Kunst bisher in dieser Form nicht auf diesem Planeten existierten. Die Rezeption dieser Inhalte hat Einfluß auf unsere Wahrnehmung und das menschliche Bewusstsein. Eine spannende Forschungsfrage ist, inwieweit wir diesen Mechanismus proaktiv nutzen können, um z.B. ein umfassenderes Verständnis jahrmillionen alter Ökosysteme zu erlangen, die unsere Lebensgrundlage bilden.”

Dr. Wolfgang Stille,
CTO hessian.AI