Clavister’s Post

My coauthors and I just proudly presented our paper, "This Probably Looks Exactly Like That: An Invertible Prototypical Network," at #ECCV2024 in Milan, Italy! It's been great to see familiar faces and see all the new AI innovation, especially in the Explainable AI space. Special thanks to Walter Scheirer and University of Notre Dame for supporting my research and travels! Check out the paper and poster here: https://lnkd.in/e7XZhUB5

Excited to announce the release of our new preprint titled "AGDIFF: Attention-Enhanced Diffusion for Molecular Geometry Prediction"! 📄 In this work, we present AGDIFF, a cutting-edge computational framework that utilizes diffusion models for precise and efficient molecular structure prediction. By augmenting the GeoDiff model with advanced attention mechanisms across global, local, and edge encoders, AGDIFF exhibits remarkable performance improvements on the GEOM-QM9 and GEOM-Drugs datasets when compared to traditional benchmarks. 💡 A huge shoutout to my fantastic collaborators, Dr. Alex Dickson and Fatemeh Fathi Niazi, whose contributions were invaluable to this research. Together, we've made significant strides in molecular prediction using AI. For the full details, you can access the paper and code here: 📄 Paper: https://lnkd.in/gRA4epS2 💻 Code: https://lnkd.in/gR9TZQpq #diffusionmodel #attention #GNN #conformer #generativeai #machinelearning #ai

Can your architecture fit an entire multi-billion parameter transformer in just ONE kernel, too?

Great to see the progress in applying large transformer models to scientific research.

My second paper presentation on the topic of Edge ai in Kongunadu college of engineering and technology.... Conducted on 21/09/2024... Really new experience and knowing the knowledge about edge ai... Conducted by the department of Information and technology....

At this year’s Trillion Parameter Consortium (TPC) in Barcelona, SambaNova’s Urmish Thakker spoke of enabling extremely fast inference and training performance using dataflow and SN40L during the AI Hardware Acceleration Strategies at Scale session. The session concluded with a discussion between Satoshi Matsuoka, Director of the RIKEN Center for Computational Science and Mateo Valero, Director of Barcelona Supercomputing Center, on how we can contribute to AI for science and the goal of TPC. #AI #GenerativeAI #Inference

Data Augmentation in AI for Science: An Earth Science Case Study Augmenting observational data with physical simulations improves the performance of AI models in scientific research.Continue reading on Towards Data Science »... https://lnkd.in/ecgn7j3R #AI #ML #Automation

We are thrilled to share that our research paper, "MYFix: Automated Fixation Annotation of Eye-Tracking Videos," has been published! 🎉 In this work, we introduce MYFix, a fully automatic approach for annotating eye-tracking videos. Leveraging the power of Yolov8 and Mask2Former in a late fusion technique, MYFix excels at detecting a wide variety of objects. Whether it's within an urban environment or a common context, MYFix can handle closely located objects and those at different depths with impressive accuracy. Our innovative method is set to enhance the capabilities of eye-tracking studies by providing precise and reliable fixation annotations automatically. This opens up new possibilities for research and applications in many fields including #GIScience. Feel free to check out the full paper here: https://lnkd.in/duZ25gMk. For a quick guide on how to use MYFix for your studies, check this #Medium_article: https://lnkd.in/dQ2kTXvu or visit our website: https://lnkd.in/dK4nzxht #Research #EyeTracking #ComputerVision #AI #Yolov8 #Mask2Former #Innovation #Technology #Science #Automation

Dive into the future of AI: Quantum AI Explained! In this recorded webinar, we delve into the fascinating world of Quantum AI and explore its potential to revolutionize the field of Artificial Intelligence. watch now: https://lnkd.in/g8nj-36X Our experts unpack the complexities of quantum computing and its implications for machine learning algorithms. Watch now and gain insights into: The fundamentals of Quantum AI How quantum computing can overcome the limitations of classical AI Potential applications of Quantum AI across various industries And much more! Don't miss out on this opportunity to explore the cutting edge of AI. Watch the recorded session now! #quantum #quantumai #quantumcomputing #artificialintelligence #quantummission

🌐 Under the title "AI Research Meets Practice: Perspectives for Companies and Society", the consortium project SAIL Research Network invited to the Applied AI Day at the Hochschule Bielefeld! 🌐 In addition to contributions from research and industry on the topic of AI, a demonstrator exhibition showed possible applications of AI. inIT employees Julian Knaup and Jörn Tebbe were also there and presented their PhD projects. 👉 Read the full article at: https://lnkd.in/einuZtzq #AppliedAIDay #DemonstratorExhibition