Intel Labs’ Post

LiSA, the first semantic-aware AI framework for highly accurate 3D visual localization, has been recognized as a highlight paper at CVPR 2024. By leveraging pre-trained semantic segmentation models, LiSA can significantly improve 3D visual localization accuracy without introducing computational overheard during inference. Zhipeng Cai goes into detail about the collaborative work with Xiamen University here. https://intel.ly/3Wow2I4 #AI #ComputerVision #Lidar #CVPR2024

Catch up on the latest news from our Visual Compute & Graphics Lab! Anton Kaplanyan, Intel Labs VP & Director of Graphics research, breaks down what’s been happening, including presenting a paper at SIGGRAPH 2024, developments in generative AI, and a progress update on Intel’s open ecosystem of graphics libraries. Read the article: https://intel.ly/4ccMqkQ #SIGGRAPH2024 #IntelEmbree #GenerativeAI

A breakthrough method of using concurrent binary trees (CBT) for large-scale game components and the generation of bisection-based terrain tessellations was recently revealed at HPG'24. By expanding the role of CBT, Intel Labs researchers Anis Benyoub and Jonathan Dupuy were able to achieve a real-time, planetary-scale level of detail rendering on console-level GPUs for geometries that would not otherwise fit on SSDs. Learn more about the approach and how it works: https://intel.ly/3WmE0lb #ConcurrentBinaryTrees #GameDesign #Triangulation

While it comes with 16 petabytes per second (PB/s) of total memory bandwidth, 3.5 PB/s of inter-core communication bandwidth, and 5 terabytes per second (TB/s) of inter-chip communication bandwidth, Hala Point isn’t just fast; it's also efficient and scalable. Learn more about the world's largest neuromorphic system and our work with Sandia National Laboratories here. https://lnkd.in/gAiyCrmy #Developer #NeuromorphicComputing #Research

Congratulations to our very own Ilke Demir for being part of a select group of innovative early-career engineers invited to participate in the 2024 Grainger Foundation Frontiers of Engineering Symposium! Hosted by the National Academy of Engineering, the annual event brings together exceptional engineers across a variety of disciplines to share knowledge, facilitate collaboration, and build professional networks among the next generation of engineering leaders. Read more: https://intel.ly/3y6MrZQ #Engineering #GraingerFoundation #FrontiersOfEngineering

A novel online evaluation protocol for test time adaptation (TTA) methods? A novel task-centric angle for the pre-trained weights of large language models (LLMs)? A general form of dynamic convolution that redefines the basic concepts of kernels? These are but a few of the innovative research topics being presented by Intel Labs researchers at this year's ICML. Learn more about our event activities in this blog. https://intel.ly/3YfWbvA #ICML #ML #AI #Research

Transforming a non-differentiable rasterizer into a differentiable one without critical performance or qualitative penalties might be easier than you think. Intel Labs researchers Thomas Deliot, Eric Heitz, and Laurent Belcour explain this in their award-winning research paper. https://intel.ly/4d7EDpr #3D #Developer #GameDev

ICYMI, two new Intel Labs research papers on neuromorphic computing were published in the June issue of Nature Magazine Machine Intelligence! Drawing insights from biological signal processing, the first paper is key in enabling robots to reliably navigate their environment. The novel approach uses neuromorphic computing hardware for fast and power-efficient visual odometry, and the related task of simultaneous localization and mapping. The companion paper explores how a neuromorphic resonator network can be applied to visual scene understanding. Learn more about our collaborative research efforts to improve visual perception in robotics here: https://intel.ly/4bEeaPa #ArtificialIntelligence #NeuromorphicComputing #VisualOdometry

Is neuromorphic computing the solution to addressing rapidly increasing energy consumption driven by the AI revolution? Hear from industry experts like our very own Mike Davies in this article from the BBC. https://intel.ly/3Lej6iO #Research #Science #NeuromorphicComputing