research information

Have you ever wondered how much energy your deep models consume during training and inference? Have you ever attempted cutting down energy usage? Do you have what it takes to develop state-of-the-art AI models without investing hundreds of GPU hours?

If so – then the 2nd AI-HERO hackathon is for you! Build an accurate and energy-efficient model to solve exciting use cases from either health or energy. You will be competing against other teams from Helmholtz.

The hackathon is organized jointly by Helmholtz Imaging, Helmholtz AI, HMC and KIT, and will take on 19-21 June 2023. You can register now! Registration is on a first-come, first-served basis.

Target group: AI-researchers with experience in model development, training and optimization
Teams: Participants will be divided into teams of three to tackle model development for one of two possible use-cases (one in research field health & one in energy). Premade teams are of course allowed!
Challenge: Build and optimize a model while using as little energy as possible

More information & registration

Institution

• Universitäts-Gesellschaft Hamburg e.V., Edmund-Siemers-Allee 1, 20146 Hamburg

Um Methoden aus dem Bereich der Künstlichen Intelligenz adäquat in der Forschung anwenden zu können, ist es erforderlich, sie in ihren Grundzügen zu verstehen und idealerweise aktiv (spielerisch) zu erproben.

• Welche Verfahren gibt es überhaupt und wie funktionieren sie technisch?
• Wo kommen sie im Akademienprogramm bereits zum Einsatz, wie können Forschende voneinander lernen und Synergien nutzen?
• Profitieren die Geisteswissenschaften nur von den neueren Entwicklungen im Bereich Generativer KI oder leisten sie auch einen Beitrag zur KI-Entwicklung?

Die Themenkonferenz, die von der Akademie der Wissenschaften in Hamburg und der Niedersächsische Akademie der Wissenschaften zu Göttingen ausgerichtet wird, möchte sich diesem Spannungsfeld widmen. Teilnehmende haben die Möglichkeit, die Potenziale neuer Methoden für die Analyse und Aufbereitung digitaler Ressourcen, speziell im Akademienprogramm, gemeinsam zu erörtern und zu erproben.

Um einen Überblick zur (auch geplanten) Verwendung von KI-Methoden zu geben, laden wir Angehörige der Unionsakademien und Langzeitvorhaben im Akademienprogramm herzlich ein, einen Impuls-/Kurzvortrag oder einen Vorschlag für eine Hands-On-Session oder Systemdemonstration beizusteuern (s. vorläufiges Programm).

Vorschläge (i.d.R. nicht mehr als zwei pro Akademie) können bis zum 30. Juni 2024 formlos an ki@akademie-der-wissenschaften-in-hamburg.de gesendet werden. Es ist vorgesehen, die Foliensätze der Beiträge auf Zenodo zu veröffentlichen.

Vorläufiges Programm hier

On September, 25th to 27th, 2023 the third MLE Days on machine learning in engineering will take place on the campus of the Hamburg University of Technology (TUHH). This year, it combines a summer school with a one-day startup-challenge. Participation is free of charge.

The Summer School teaches insights into the world of machine learning with a focus on engineering. It provides sessions about fundamentals of machine learning, concrete application examples, as well as hands-on sessions to try out and consolidate lessons learned. Keynote talks complete the program. Three parallel tracks are provided to allow participants to choose contributions adapted to their interest and machine learning experience. The use cases range from sensor and image processing, to electrical engineering and materials science, to aviation and maritime logistics. A poster session and an elevator pitch event allow participants to present their own work on machine learning topics. The best posters and pitches will be selected by a jury and awarded prizes. A networking event allows attendees to establish contacts with selected corporate partners and sponsors from start-ups and medium-sized businesses to large corporations. In the startup-challenge attendees learn how to turn machine learning ideas into a business.

The MLE Days are organized by the Machine Learning in Engineering research initiative of the TUHH (MLE@TUHH) in collaboration with the Helmholtz Center Hereon, DASHH, and the Career Center of TUHH, the AI.Startup.Hub, and AI.HAMBURG. The MLE initiative joins the competencies in the field of machine learning at the Hamburg University of Technology with the goal of transferring knowledge towards business and industry. Students, PhD students, postdocs, and professors from all disciplines of the TUHH are engaged together with colleagues from the Helmholtz Center Hereon to make methods and applications of machine learning known, to network, and to foster scientific exchange.

Join us to discover the new applications of machine learning in engineering practice!

Institutions

• Hamburg University of Technology (TUHH)