DLR Institute of Software Technology’s Post

🌴 Ferienzeit! 🌴 Wer am Strand 🏖️, im Paddelboot 🚣🏽♀️ oder beim Bergwandern ⛰️ in ein paar aktuelle Forschungsthemen reinhören möchte - hier sind einige 🎙️ #Podcast-Empfehlungen mit unseren ExpertInnen Sophie Jentzsch und Benedikt Fauseweh. Um #Humor und #Moral in der #KI geht es in den Gesprächen mit unserer Large Language Model-Forscherin Sophie Jentzsch. 🎧 Deutschlandfunk: „Wie gut kann KI Humor?“ 🔉 https://lnkd.in/evMpEVX9 🎧 Digital Sensemaker: „Wie wäre es, wenn ich meiner KI moralische Fragen stellen könnte?“ 🔉 https://lnkd.in/ebJqGQHd Als Experte für #Quantencomputing und #Festkörperphysik ist unser Kollege Benedikt Fauseweh in diesen beiden Podcasts zu Gast. 🎧 Von der Erde ins All: „Quanten-Supra-Computer-Leiter“ 🔉 https://lnkd.in/eEGUXu_6 🎧 Post Quantum World: „The Quantum Many-Body Problem” 🔉 https://lnkd.in/e9HQgZ6C   Viel Spa�� beim Anhören und allen eine erholsame Ferienzeit! ☀️☀️☀️

🔎  #QuickLookIntoResearch - 🤖🧠 #AI in #space 🚀 🌌 As part of our "Quick Look Into Research" series, we present projects in which our institute is working on AI in space research. In this project, we aim to make #AI #simulations more #robust by combining them with the #laws of #physics. ⚛️ 🌳 🍎 #Numerical #simulation has been a major leap forward in the development process of complex engineering systems such as space vehicles and avionics. 🚀✈️ The next step is creating more robust and #safer #simulations by introducing artificial intelligence in the form of #neural #networks 🧠 to them: this helps to reduce complexity and to speed up computational evaluation. 📈 The danger with neural networks, which are usually trained from data, is that some laws of nature, such as conservation of energy, mass and momentum, are only poorly or not at all taken into account. ⚠️ Our multifaceted approach makes the numerical simulation method safer and more robust, as shown in our "PISA" project ➡️ https://lnkd.in/e6nFhayx. The goal of the project is to develop #hybrid #predictive #models, methods and deep learning architectures that respect fundamental physical principles and integrate them into data-driven AI models. 💬 Our colleague Philipp Knechtges explained at the DLR Quarterly Meeting on Space research how this comes into play not only in the preparatory calculation of aerodynamics before the production of aircraft components, but also in real-time during future flights: 👉 "The #advantage of this #hybrid #approach is the generation of models that combine the accuracy of physics-based simulations with the computational speed of data-driven AI models. Hybrid predictive models are particularly useful when computational time or resources are limited. This is often the case in control systems engineering, where computational resources are limited and real-time capability is critical". We invite you to join us in our 🔎 #QuickLookIntoResearch series by following this hashtag!

👏 Congratulations to our colleague, Felix Terhag, on winning the #Best #Paper #Award at the 9th European Seminar on Computing (ESCO) 2024 in Pilsen! 🏆 Felix was awarded for presenting his work on #machine #learning algorithms for use in #cardiac #MRI 🏥. In cardiovascular medicine, techniques such as magnetic resonance imaging (MRI) are essential for accurate analysis of the cardiovascular system. Felix describes the aim of his and his colleagues' research: “Real-world machine learning applications involve data with varying difficulty. Transferring information from easier examples to complex ones can be very beneficial. Another challenge for machine learning algorithms is assigning uncertainty to their predictions. Not knowing the prediction uncertainty is especially dangerous in safety-critical applications like rocket-control or health. In our approach we tackle both problems at once. In the application of estimating the ventricle volume from cardiac-MRIs, machine learning algorithms are capable of estimating the volume in the middle of the heart with high precision. The apical and basal slices are a lot harder to predict. Our approach automatically identifies the main contributing frequencies (e.g. heart rate and respiratory rate) from the well predicted slices. With an expert in the loop we are able to obtain a handful of manually labelled images. We can use these to accurately predict the volumes over time at the difficult slices. By choosing a Bayesian approach, we also obtain a posterior distribution given the labelled images. This distribution helps us to obtain an uncertainty measure for these slices.” ➡️ More on our research on realtime MRI: https://lnkd.in/ey8yk4jU

🔎 #QuickLookIntoResearch - 🤖🧠 #AI in #space 🚀 🌌 As part of our "Quick Look Into Research" series, we present some of our projects where our institute is working on AI in space exploration. In an Extended Reality project, we are examining #AI for use in #digital #twins of aircraft or space. ✈️ 🚀 🥽 #Extended #Reality (XR) comes into play when working with digital twins of spacecraft or aircraft. The #Airbrake use case, which we recently presented at #HannoverMesse together with the DLR - Institut für Systemleichtbau, is a study in which digital twins coupled with AR applications enable #remote #collaboration on aircraft #maintenance tasks 💻⚙️🤝. In the future, the damage analysis, including photos or ultrasound scans, can be transmitted in real time to the digital twin in a data cloud. This allows experts around the world to participate in the assessment and support the technicians on site. Our institute is responsible for developing the AR-based #user #interface. Luljeta Sinani, who works in our Extended Reality group, sums up our goals: "We focus on creating an optimal user experience, generating a detailed resolution of the virtual object and flawless integration of data in the digital twin. The use of AI can support real-time data analysis 👩🏽💻, predictive maintenance 🛠️, and operational efficiency 📈, thereby helping spacecraft and aircraft inspectors in better and faster decision making ✅. For instance, an AI assistant could guide you through inspection procedures and assist in decision-making based on previous anomaly findings and report submissions.”   We invite you to join us in our 🔎 #QuickLookIntoResearch series by following this hashtag!

🔎  #QuickLookIntoResearch - 🤖🧠 AI in space 🚀 🌌   As part of our #QuickLookIntoResearch series, we present some of our projects where our institute is working on AI in space research. In the first project, we use AI to check the quality of #Earth #observation #images 🌍 🛰️ 📷 📷 Earth observation satellites produce an enormous number of images with their on-board cameras during their lifetime. But not all images are good enough to give us valuable information about the Earth's surface, clouds, climate or weather. 🌐🗺️🌧️☀️ As part of the project #ScOSA Flight Experiment (ScOSA stands for Scalable On-Board Computing for Space Avionics), our institute developed an #AI-assisted #system that automatically recognises which images taken on a satellite are good enough to download and which can be discarded, significantly reducing the data traffic between the satellite and ground stations. 🛰️ This image selection application we tested successfully on board an European Space Agency - ESA research satellite (https://lnkd.in/dxYkhAq8). The miniature satellite OPS-SAT acted as a test lab in orbit, experimenting with new mission control software and techniques. With the AI-based image quality verification application, we were able to demonstrate the reliability and flexibility of ScOSA, where our main focus is the development of a reliable on-board computer for space avionics. At a space research meeting at Deutsches Zentrum für Luft-und Raumfahrt e.V., our colleague Arnau Prat presented the successful results of the AI software application. He also gave a sneak peek of what is to come on the #next #flight in #2026. He showed a 3D printed model of the #hardware #structure where the ScOSA on-board computer will be integrated on a DLR CubeSat satellite. Arnau explains: "We have been working on a stable computer system and testing for possible disturbances, such as processors failing due to space radiation. Our software is able to identify faulty nodes in the processors and keep the system running. Seeing the ScOSA software successfully running on the ESA OPS-SAT satellite was an exciting moment! And now we are already looking forward to the next highlight with the ScOSA Flight Experiment flying on the DLR CubeSat in 2026 with other AI applications.” The DLR Institute of Software Technology leads the project #ScOSA, an innovative #distributed #computing #platform for spacecraft that allows applications to be dynamically moved to other computing nodes, for example, in case of failure, to provide high computing performance with #high #reliability for #space #missions. For more information visit the project page ScOSA Flight experiment 👉 https://lnkd.in/dhP6p_6E We invite you to join us in our 🔎 #QuickLookIntoResearch by following this hashtag!

🔎 #QuickLookIntoResearch 🤖🧠 AI in space 🚀 🌌 #Space #exploration has become indispensable in the 21st century. It provides important infrastructure for modern life on Earth, such as communications, navigation, Earth and climate monitoring, and robotic applications. 📡🛰️🌍 As in many areas of research, #AI will provide advanced ways of collecting and analysing huge amounts of data, using large amounts of computing power, and opening up more and more areas of application. As a software research institute, we are at the heart of the Deutsches Zentrum für Luft-und Raumfahrt e.V.'s development of artificial intelligence for aerospace applications, as well as future human-AI collaborations for space missions. 🚀👩🏽🚀🪐 Over the next few weeks, we will give you an #insight into some of our #projects involving AI in space research. The projects range from imaging in Earth observation, to working with digital twins, to specifying AI algorithms by integrating the laws of physics. We invite you to join us in our 🔎 #QuickLookIntoResearch by following this hashtag!

🎧📢 Podcast Alert - "The Post-Quantum World" with our quantum computing expert Benedikt Fauseweh Benedikt Fauseweh was invited to Protiviti's #podcast "The Post-Quantum World" to discuss the progress and future implications of (post-)quantum technologies with host Konstantinos Karagiannis: “When Richard Feynman proposed the idea of a quantum simulator or computer in 1981, he was frustrated by the limitations of classical systems. He logically suggested that if we live in a quantum world, we need a quantum device to simulate all the interactions of particles that make up reality. An excellent example of such a transistor-choking calculation is the quantum many-body problem. Have quantum computers finally cracked it and reached the level of Feynman’s original idea? Can working on such approaches lead to better quantum computers and help solve real business use cases soon? Join Host Konstantinos Karagiannis for a chat with Benedikt Fauseweh from TU Dortmund University and the German Aerospace Center (DLR) as they explore these topics and whether this work has anything to do with The Three-Body Problem novel and Netflix show.” 🎧 Listen now: https://lnkd.in/e97cZdAa

Unser Kollege und #Quantencomputing-Experte Benedikt Fauseweh ist ein gefragter Autor und Interviewpartner, wenn es um das komplexe Thema supraleitender Materialien und Quantencomputer geht. 👏 📢 📰 Sein gemeinsam mit Andreas Alexander Buchheit und Dr. Torsten Keßler verfasster #Artikel über langreichweitige Wechselwirkungen in Festkörpern ist im aktuellen #Physik #Journal der Deutsche Physikalische Gesellschaft e. V. (DPG) erschienen: https://lnkd.in/ebbjB86E 📢 🎧 Eine weitere Empfehlung ist das Interview mit Benedikt Fauseweh im #Podcast des Deutsches Zentrum für Luft-und Raumfahrt e.V. „Von der Erde ins All“. Hier berichtet er über das Henne-Ei-Problem bei der Entwicklung solcher neuen Supraleiter, die für leistungsfähige Quantencomputer eingesetzt werden könnten. Hier ist die Folge zu hören:  https://lnkd.in/eSbu6Qxq

We are excited to be part of making the first ion trap computer at DLR accessible!   It's a particular pleasure to collaborate closely with QSea I on connecting the first ion trap computer at Deutsches Zentrum für Luft-und Raumfahrt e.V. We see this as a unique opportunity – in the future we will be able to address very different technologies via a standardised interface. The use cases may range from battery material simulation to climate models. 🔋🌍   It is exciting to see how the pieces of the puzzle fit together and how more and more prototype quantum computers of various technologies are gradually arriving. 🧩   As experts on hardware-software-codesign our institute executes the integration of DLR quantum hardware by providing user access on further projects such as #ALQU (https://lnkd.in/dWB7iGjR) and #CLIQUE (https://lnkd.in/d3bn8Fgg).