ETH IVIA Lab

Thanks to everybody who stopped by our poster at #neurips2024! We are also presenting two workshop posters this weekend: 🧫 Explainable AI-based analysis of human pancreas sections identifies traits of type-2 diabetes at the Interpretable AI Workshop 👥 With Sebastian Ziegler, Felicia Gerst, Yanni Morgenroth, Karol Gotkowski, Paul F. Jaeger, Fabian Isensee, Michele Solimena, Robert Wagner and more. ⛓️💥Interactive Semantic Interventions for VLMs: A human-in-the-loop approach to interpretablity for VLM model failure detection at the Red Teaming GenAI Workshop. 👥 With Kenza Amara, Carsten Lüth, Hendrik Strobelt, Mennatallah El-Assady and Paul F. Jaeger. Happy to discuss both topics at the workshops!

Poster Session "Fundamentals of Web Engineering" Join us for the poster session of the “Fundamentals of Web Engineering” course, hosted by the Interactive Visualization & Intelligence Augmentation Lab (IVIA). Student teams will showcase their web development projects on a variety of topics, presenting their results through engaging posters. These projects highlight the development of full-stack web applications, created in collaboration with research and industry partners. Explore the interactive visual interfaces, engage with the student teams, and discuss potential directions for future work. The poster session will take place on: 🗓️ Date: Thursday, 19th June 2024 ⏰ Time: 12:00 am - 1:00 pm 🧭 Place: ETH Zürich, OAT Floor 14, S 15 (Andreasstrasse 5, 8050 Zürich) 👉 Registration link (open until Dec 15): https://lnkd.in/dKSwgmEX

🔍 Picking the right explainable AI method for your computer vision task? Wondering about its evaluation reliability? Check out our latest #neurips2024 D&B publication on LATEC, a (meta-)evaluation benchmark for XAI methods and metrics. #xai 🚀 Through LATEC, we showcase the risk of conflicting metrics causing unreliable rankings and propose a more robust evaluation scheme. We critically evaluated 17 XAI methods across 20 metrics in 7,560 unique setups, including varied architectures & input modalities. 🎯 Curiously, the emerging top-performing method is not examined in any relevant related study. Dive into our findings: Paper: https://lnkd.in/eApfZJvg Benchmark: https://lnkd.in/ePYMWv3h   👥 This work was a Team effort involving Carsten Lüth, Udo Schlegel, Till Bungert, Mennatallah El-Assady, and Paul F. Jaeger.

Explore the intersections of human and artificial intelligence in today's SeminarSeries#3: Intelligence Augmentation: Bridging Human and Artificial Intelligence by Prof. Dr. Mennatallah El-Assady 🗓 Date: 5th December 2024  ⏰ Time: 17:30 – 19:00  📍 Location: ETH Zürich Hönggerberg HIB E 52, ITA Open Space 2 Organized by Design++, ETH Zurich, hosted at the Institute of Technology in Architecture, ETH Zurich, Department of Architecture (D-ARCH) ETH Zurich. See you there!  Organized by: Dr. Andreas Müller, Anton Savov, Fabio Scotto, Nijat Mahamaliyev, Erika Marthins 

🌍 Global science network at TUM Campus Heilbronn! The world was a guest at TUM Campus Heilbronn: In mid-November, researchers from the excellence universities of Stanford University, the University of Oxford, ETH Zürich, The Hebrew University of Jerusalem, the two universities from Singapore (National University of Singapore and Nanyang Technological University Singapore) celebrated their premiere, as did the Canadian representatives from the University of Waterloo, met at the TUM Global Technology Forum in the buildings of the ‘Bildungscampus’ in Heilbronn. The scientists could exchange ideas and research results on various topics such as Artificial Intelligence (AI) or quantum computing. Prof. Ali Sunyaev, Vice President of TUM Campus Heilbronn, warmly welcomed the assembled scientists: “It's great to have you all here. At TUM Campus Heilbronn, management and computer science are combined in a unique way. Take this chance; join forces and take research to the next level.” For three inspiring days, Prof. Helmut Krcmar led many keynote lectures, with a unifying element of AI connecting the individual talks. 💡   🔑 Key Takeaways from the sessions: 🔹AI can be a real game-changer in developing drugs and decoding the human genome. 🔹Sustainability is the key to the future: AI can help act efficiently and improve results in business processes and using natural resources. 🔹Voice cloning is developing into a strategically valuable marketing tool. Rare languages can be decoded and preserved using algorithms. 🔹The future of software engineering lies in human-centered AI collaboration. 🔹AI can spread conspiracy theories but also help us identify them. 🔹Mentors are experienced, take a structured approach, and point out new courses of action – AI can take over some of these tasks. 🔹AI has also long since found its way into medicine. Doctors must be able to understand and apply it without giving up their freedom of decision.   To sum up, AI has long since entered the most diverse areas of life, and it is no longer possible to imagine life without it. It is important to seize the opportunities it offers and overcome initial reservations. Or as Prof. Helmut Krcmar puts it: “After people have used an AI tool for the first time, their trust often increases rapidly.”   The TUM Global Technology Forum has established itself as a ‘knowledge marketplace’ – and we’re already looking forward to next year! 🚀   #TUMCampusHeilbronn #GlobalTechnologyForum #Innovation #Sustainability #FutureOfWork #KnowledgeExchange #AI #DigitalAge

I am happy to share that I successfully defended my PhD Proposal last week! This milestone marks the official start of my PhD candidate-journey. My research, "Humanizing Recommender Systems: Interactive and Explainable Recommendations through Data Humanism," explores how we can make explanations not just an add-on to check a box for regulatory compliance but an integral part of the 'personalized experience' that recommender systems aim to deliver. Thanks to my supervisors, Juergen Bernard and Mennatallah El-Assady, for their support over the past two years and to my proposal committee, Miriah Meyer and Francesco Ricci, for their feedback on the proposal. I look forward to diving deeper into my research and contributing to the cross-section of explainable recommender systems, data visualization, and HCI! Future skill to learn: how to be good in pictures while presenting. For now, here are some of my slides.

🔍 Picking the right explainable AI method for your computer vision task? Wondering about its evaluation reliability? Check out our latest #neurips2024 D&B publication on LATEC, a (meta-)evaluation benchmark for XAI methods and metrics. #xai 🚀 Through LATEC, we showcase the risk of conflicting metrics causing unreliable rankings and propose a more robust evaluation scheme. We critically evaluated 17 XAI methods across 20 metrics in 7,560 unique setups, including varied architectures & input modalities. 🎯 Curiously, the emerging top-performing method is not examined in any relevant related study. Dive into our findings: Paper: https://lnkd.in/eApfZJvg Benchmark: https://lnkd.in/ePYMWv3h   👥 This work was a Team effort involving Carsten Lüth, Udo Schlegel, Till Bungert, Mennatallah El-Assady, and Paul F. Jaeger.

🚀 𝐀𝐫𝐭𝐢𝐟𝐢𝐜𝐢𝐚𝐥 𝐢𝐧𝐭𝐞𝐥𝐥𝐢𝐠𝐞𝐧𝐜𝐞 𝐦𝐞𝐞𝐭𝐬 𝐢𝐧𝐝𝐮𝐬𝐭𝐫𝐲 What happens when leading AI professors from the world's most renowned universities meet with an industry pioneer like ZIEHL-ABEGG? The answer: an exciting dialog about the future of artificial intelligence in industry! As part of the Global Technology Forum, organised by Technische Universität München Campus Heilbronn, we welcomed a high-caliber delegation of AI experts from universities of Paris, Oxford, Stanford, Zurich, Jerusalem, Waterloo and Singapore. Prof. Dr. Dr. h.c. Helmut Krcmar, Founding Dean of the TUM Campus Heilbronn, emphasizes the importance of the cooperation: “We scientists are always looking for contact with industry. I am delighted that ZIEHL-ABEGG has allowed us to look deep behind the scenes. Interlinking research and teaching with industry is a key concern of mine - because both benefit from this.”  The focus of the visit was a tour of the state-of-the-art laboratory with the world's largest combined measurement and test rig for fans. Our CTO Wolfgang Mayer summed up the visit enthusiastically: “It was a great honor for us to be able to give the AI experts from the renowned universities an insight into our industrial company. I am particularly proud of how amazed the scientists were at the complexity of a supposedly simple fan.”  #AI #artificialintelligence #science #industry

We're #hiring a new Administrative Assistenz (m/w/d) am Institut for Visual Computing | 50% in Zurich, Zurich. Apply today or share this post with your network.

Congratulations! Juergen Bernard, head of the Interactive Visual Data Analysis Group at the UZH Department of Informatics, has been awarded a Schweizerischer Nationalfonds SNF Grant on «Personalized Visual Analytics: Human Preference Elicitation for Ranking-Based Multi-Criteria Decision Support». Together with Mennatallah El-Assady, head of the Interactive Visualization and Intelligence Augmentation Lab at ETH Zürich, he will explore multi-criteria decision support through interactive item ranking, as an understudied research direction for people lost in choosing between thousands of relevant items (cars, houses, stocks, etc.). Their Visual Analytics approach will support this complex human decision-making task by bringing together human intellectuality with the computational power of algorithms and AI in effective human-in-the-loop approaches. Learn more: https://lnkd.in/esF8w8Rm #SNFGrant #ifiuzh