Yulia Sandamirskaya

📰 COREnext Scientific Publication Alert! 📰 👉 Check out our #research paper titled "#Software-Defined CPU Modes." Authors: Michael Roitzsch, @Till Miemietz, Christian von Elm, Nils Asmussen 💡 This paper delves into the innovative concept of defining CPU modes entirely through software. Our researchers demonstrate how this design could function, highlighting the advantages it offers. By unifying all existing modes under a common design umbrella, this approach enforces a cleaner structure and greater control over exposed functionality. Additionally, the flexibility of #software-defined modes opens up exciting new use cases. 👉 Discover the future of #CPU design and its potential impact on computing! #Research #CPUModes #SoftwareDefined #Innovation #TechResearch #FutureOfComputing 🔗

2

At the end of March, we were privileged with the visit of President of the Higher Regional Court Mag. Schwanda and our Rector Univ. Prof. Bischof. We talked about generative AI, embodied AI, Computer Vision, Robots and Virtual Reality at the Human Centered Computing labs, Graz University of Technology. #tugraz #hcctug #robots #ai #genai #aiact #metaverse #vr #humancenteredcomputing #styria

3

Professor Michael M. Resch, director of the HLRS - High-Performance Computing Center Stuttgart, gave an exciting talk on the potential of high-performance computing (#HPC) and artificial intelligence (#AI) at the TalTech – Tallinn University of Technology Academy of Architecture and Urban Studies, which we also streamed live due to the great international interest. Here you have the opportunity to watch the talk and the subsequent discussion - in which topics relating to artificial intelligence and quantum computing in particular were addressed. You might ask yourself: What does this have to do with architecture and urban studies? We live in an increasingly complex world with enormous challenges - from climate change and the resulting environmental and natural disasters to growing social tensions, global conflicts and pandemics. All of this also affects our settlements, homes, communities, health, wellbeing and the way we live. HPC, AI and future computing methods such as quantum computing can help us solve problems (even if they are not solutions per se, but only tools).

18

Michael Trott, Chief Scientist of Wolfram Alpha, delivered a talk titled "From Code to Creativity: Using LLMs for Problem-Solving" at the Physics Department of Vienna University. The University of Vienna has a rich history in advancing modern physics. We look forward to collaborating further, exploring how new technologies such as LLM can drive progress in physics research and education. A huge thank you to Markus Arndt for hosting our talk and fostering an engaging discussion. If you miss the talk, review the presentation material here: https://lnkd.in/gRhnPQMq 💡 Let’s keep the conversation going! How do you see LLMs reshaping problem-solving in your field?  #AcademicInnovation #WolframAlpha #AI #Creativity #LLMs #ProblemSolving #ArtificialIntelligence #Innovation #Physics #ComputationalThinking #DataScience #Research #Collaboration #FutureTech #STEM

12

🤖 Reminder: tomorrow Tue Dec 3, Professor Amanda Prorok (University of Cambridge) will give a talk on AI+Robotics in our ETH AI Center Academic Talk Series (AICATS), hosted by HEHUI ZHENG! This is a public talk with mandatory registration. Don’t miss this last academic talk of the semester! Last but definitely not least! ℹ️ More info, talks, updates, and registration in the comments.

15

1 Kommentar

AI or teacher? A new EPFL paper has found that students are cautious towards #AI feedback, highlighting the complexity of integrating it into educational feedback systems. #education #research #trust #ML4ED Tanja Käser Tanya Nazaretsky

5

1 Kommentar

#CYDCampusConference2024 We were pleased to welcome Prof. Torsten Hoefler, Head of Scalable Parallel Computing Laboratory at ETH Zürich and Chief Architect for ML at CSCS, for a captivating session. His expertise in machine learning and high-performance computing brought valuable insights into the evolving landscape of AI and computational technologies. Prof. Hoefler starts his speech by announcing the end of the Data Age, and prophesizing the Age of Computation, using the rise of Nvidia as the 2nd largest company in the world as its poster-child, and stressing Switzerland’s pivotal role with the Alps supercomputer, which is the 6th largest in the world and 2nd largest in Europe. But as the need for computing power increases, the risks of connecting supercomputers, which were not necessarily built with security in mind, to unsecure systems and networks become crucial. Once these secure systems have been built, AI and other systems can be built on top of it.

53

3 Kommentare

nAG Principal Scientist and RWTH Aachen University Professor of Computer Science Uwe Naumann presents 'Matrix-Free Differential Inversion. A Lesson from Adjoint Algorithmic Differentiation' at QuantMinds today - reach out to Uwe or the nAG team at #QuantMinds to learn more or send us a note  #quantfinance #computation #algorithmicdifferentiation #data #innovation

26

📰 💡 Are you interested in our ideas for more adversarial robustness of machine learning? Thanks to The Julius Maximilians University of Würzburg for this press release on my joint project with Leo Schwinn, funded by Deutsche Forschungsgemeinschaft (DFG) - German Research Foundation. #AI #Research #Robustness #Machinelearning #Mathematics

18

Thanks to Dr. Wilfried Kubinger, Peter Rössler, Roman Beneder, and Florian Wimmer for the enlightening exchange on research and study programs around #Embedded & #CyberPhysical Systems at the Department Electrical Engineering & Entrepreneurship at Fachhochschule Technikum Wien . The connection between theory, physical fabrication, programming and simulation is impressive. By combining electromagnetic & IT topics with everyday life questions, you implemented so many things to make 🔌 #ElectricalEngineering ⚡ attractive for students. To name only one, you let them design and program microcontrollers where students can instantaneously ‘see’ the impact, namely with their commands they are controlling an Uncrewed Aerial Vehicle. In that little example they learn programming, sensor fusion, control technology and many other things. With Dimitris Tzagkas we discussed further ways of virtualization. Looking forward to further collaboration! https://lnkd.in/dqxJbgVq Ansys CADFEM (D-A-CH) #Academic    

33

1 Kommentar

🎉 Congratulations to all the recipients of the Wolfram Innovator Award 2024! 🏆 Special congratulations go to Professor Thomas Wallek from Technische Universität Graz on winning the Wolfram Innovator Award for his groundbreaking work in Monte Carlo simulations! 👏 🏆 Professor Wallek and his team have developed an innovative 𝗚𝗶𝗯𝗯𝘀 𝗘𝗻𝘀𝗲𝗺𝗯𝗹𝗲 𝗠𝗼𝗻𝘁𝗲 𝗖𝗮𝗿𝗹𝗼 (GEMC) algorithm for simulating fluid phase equilibria using 𝗪𝗼𝗹𝗳𝗿𝗮𝗺 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 and 𝗠𝗮𝘁𝗵𝗲𝗺𝗮𝘁𝗶𝗰𝗮. Their GEMC software includes advanced features such as: - The OPLS force field for intermolecular interactions - The Widom ghost particle method for chemical potential - Enhanced molecule transfer algorithms for binary mixtures - Long-range and tail corrections for energy and pressure This tool enables precise phase equilibrium calculations, setting new benchmarks in the field! 🚀 #WolframInnovatorAward #MonteCarlo #TUGraz #WolframLanguage #Mathematica #Innovation https://lnkd.in/d32p_XXS

17

2 Kommentare

Join us for an interactive session with Dr Stefan Leichenauer Vice President of Engineering SandboxAQ Palo Alto, USA Registration Link- https://cutt.ly/5egshCHQ About Dr STEFAN LEICHENAUER Dr. Stefan Leichenauer, a trailblazer in AI and Quantum technologies, serves as the Vice President of Engineering and lead scientist at SandboxAQ. As the inaugural member of the Sandbox team at Alphabet, Stefan played a pivotal role in transforming theoretical concepts into tangible real-world applications. His visionary approach and technical acumen led to the initiation of groundbreaking projects, shaping the foundation of SandboxAQ solutions. About SandboxAQ In the tech and data intelligence worlds, a sandbox is where innovation is born. It’s a place where the brightest free-thinking minds from across disciplines come together to reimagine what’s possible. A collaborative environment where the whole is infinitely greater than the sum of the parts. SandboxAQ is an enterprise SaaS company, providing solutions at the nexus of AI and Quantum technology (AQ) to address some of the world's most challenging problems. The company's core team and inspiration formed at Alphabet Inc., emerging as an independent, growth-capital-backed company in 2022. For more information, visit www.sandboxaq.com #tribaldesignforum #tdftalks #tribalfutures #sandboxaq #ai #innovation #engineering #paloalto #USA #techology #dataintelligence #saas #quantumtechnology #advancedcomputing #future #science #data #tribaldesign

1

In our newest video, Professor Robert Wille from Technische Universität München explains why quantum computing requires a multidisciplinary approach. It should not only respect the contributions of physicists but also harness the power of computer science to develop robust software solutions. His talk is part of the seminar series "Quantumness" in cooperation with the Institute of Advanced Studies (TUM IAS - www.ias.tum.de), from which we post new talks every two weeks. Here you can find the video: https://lnkd.in/dUJtu4W2 #QuantumComputing #SoftwareDevelopment #DesignAutomation #InterdisciplinaryResearch #TechnologyInnovation #ComputerScience #QuantumTechnology #ResearchCollaboration #FutureTech #TechInsights #QuantumSoftware #InnovativeSolutions #TechCommunity #QuantumAlgorithms #ScienceAndTechnology #TUM #ResearchAndInnovation #Philosophy

28

1 Kommentar

Exciting News! After 1.5 years of review, our paper titled “Cognitive Load-based Affective Workload Allocation for Multi-human Multi-robot Teams” has been accepted for publication in IEEE Transactions on Human-Machine Systems (#THMS)! This paper dives into the fields of #HumanMultiRobotSystems and #HumanRobotTeams, focusing on how humans and robots can collaborate more effectively as a team. We developed a system leveraging deep reinforcement learning (#DRL) to dynamically adjust workloads based on operator performance, combining both self-reported and real-time deep learning-based cognitive workload predictions. Our work included extensive real-world user studies in a CCTV monitoring scenario with multi-humans and multi-robots, demonstrating the proposed system's ability to dynamically adapt operators' cognitive workloads. Open-access preprint (arXiv): https://lnkd.in/gtCq68KZ Supplementary video: https://lnkd.in/g8Wpsb2u Paper website: https://lnkd.in/guqGKV4s We are deeply grateful to the reviewers and editors for their insightful feedback and comments, which significantly improved our paper. A huge congratulations to my fantastic co-authors, Wonse Jo, Ruiqi Wang, Baijian Yang, Dan Foti, and Mo Rastgaar, and especially to our first author, Wonse, for his hard work and dedication to this exciting project. I’m so proud to see this research come to life! #HumanMachineSystems #MultiHumanMultiRobot #HumanRobotInteraction #AI #ReinforcementLearning #DeepLearning #NSFCAREER #Purdue #PurdueSMARTLab

83

6 Kommentare

I am very excited to report that our Project „GeniusRobot“ joint with Björn Ommer (Ludwig-Maximilians-Universität München), Wolfram Burgard (Technische Universität Nürnberg), and Roberto Calandra (Technische Universität Dresden) was approved by the Bundesministerium für Bildung und Forschung, see also https://lnkd.in/dUtsnt7m. Recent successes in #artificialintelligence (#AI) significantly impact our daily lives through systems for #autonomousdriving, #speechrecognition, #imageprocessing, and #robotics. However, #robots are typically only consumers of the technology, and engineers need to train and curate the models each time anew for the tasks at hand. Deploying robots in dangerous scenarios requires flexible models which can handle multi-modal data and be flexibly adapted based on their environments. In addition, certainly such robot models have to be safe, robust, and interpretable. In our Project "GeniusRobot", we aim to lay the foundation for enabling robots to learn models from multi-modal data; this also includes computer vision, sense of touch, and forces through interaction with the environment. Our approach shall allow robots to perform closed-loop grasping and manipulation actions, leading also to a substantial increasement of their robustness and reliability. baiosphere Research Services LMU Munich Institute of Robotics and Machine Intelligence (MIRMI) at the Technical University of Munich relAI Konrad Zuse School of Excellence

93

2 Kommentare

Make sure to checkout this recent T-RO paper by researchers from Karlsruher Institut für Technologie (KIT) on automatically extracting an object-centric task representation from human demonstration videos recently presented at #ICRA2024. https://lnkd.in/gyr8XdTt Paper title: K-VIL: Keypoints-Based Visual Imitation Learning Authors: Jianfeng Gao, Zhi Tao, Noémie Jaquier, and Tamim Asfour Paper link: https://lnkd.in/gyr8XdTt Image caption: Overview of the K-VIL approach. (a) Human demonstration videos of manipulation actions involving categorical objects with shape, pose, and trajectory variations. (b) Sampling of dense candidate points from the object surface. (c) Extraction of sparse keypoints k1andk2 subject to certain types of geometric constraints (point-to-point and point-to-curve), their associated local frames F , and the movement primitives, which represent the demonstrated keypoint motions. (d) Adaptation of the learned generalizable geometric task representation to a new scene and execution by the robot. #TaskAnalysis #ManipulationPlanning #VisualLearning

35

Technische Universität Berlin researchers present a method for the direct heteroepitaxial growth of high-quality InGaAs quantum dots on silicon: https://ow.ly/F5e850SaKHo Through the use of a GaP buffer and meticulous mitigation of defects at the III/V–Si interface, the method enables the fabrication of scalable and cost-effective quantum photonic devices compatible with silicon technology. The research, published in #OPG_OpticaQ, could pave the way for future chips combining electronic and quantum photonic functionality. Written by: Imad Limame, Peter Ludewig, Ching-Wen Shih, Marcel Hohn, Chirag C. Palekar, Wolfgang Stolz, and Stephan Reitzenstein #quantum #quantumscience #technology #silicon

34

3 Kommentare

Our paper "Generating PLC Code with Universal Large Language Models" has been accepted to ETFA'24. Joint work with Jérôme Pfeiffer, Jingxi Zhang, and K. Tran kf ISW Universität Stuttgart and Bianca Wiesmayr of JKU - LIT Open Innovation Center ⇒ Lesson learned: Train your own language model or use ChatGPT4. Preprint follows.

61

5 Kommentare