Intel Labs

The NPU is a low power, energy efficient processor engine that elevates the game of AI model deployment on your local machine. Benjamin Consolvo, AI Engineering Manager at #Intel, discusses the new Intel® NPU Acceleration Library and how its enabling large multimodal models (LMMs) in this Hugging Face blog. https://intel.ly/3S42OwR #GenAI

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

Enable effective understanding of first-person videos captured by “always ON cameras” via a novel representation, namely ego-centric action scene graphs (EASGs). EASGs extend standard verb-noun action labels and provide a temporally evolving graph-based description of the actions performed by the camera wearer, including interacting objects, their relationships, and how actions unfold in time. Through a novel annotation procedure, we extend the Ego4D dataset, adding manually labeled Egocentric Action Scene Graphs, which offer a rich set of annotations for long-form egocentric video understanding. We show how EASG representations improve performance over state-of-the-art methods across different long-form video understanding tasks, such as action anticipation and summarization. Learn more in part 3 of this 3-part series on long-form video representation learning from Subarna Tripathi. https://intel.ly/4buWLIP #DeepLearning #ComputerVision #RepresentationLearning

Say hello to the first fully integrated optical compute interconnect (OCI) chiplet. When co-packaged with CPUs, GPUs, and other SoCs, Intel's OCI chiplet is built for scale to address AI infrastructure bandwidth needs now and in the future. Designed to support 64 channels of 32 Gbps data transmission in each direction on up to 100 meters of fiber optics, this first 4 Tbps OCI chiplet enables future scalability of CPU/GPU cluster connectivity and novel compute architectures, including coherent memory expansion and resource disaggregation. Learn more about Intel's OCI chiplet and how it's expected to revolutionize high-speed data processing for AI infrastructure. https://intel.ly/3RGKTfy #ArtificialIntelligence #SiliconPhotonics #DataCenter

Code-named Pando Tree, this groundbreaking quantum research control chip from Intel addresses quantum computing interconnect bottlenecks at the millikelvin stage. This millikelvin chip establishes Intel as the first semiconductor manufacturer to demonstrate the distribution of cryogenic silicon spin qubit control electronics inside the dilution refrigerator at different temperature stages. The Pando Tree control chip is designed to be fully compatible with the standard Intel Foundry Flip Chip Ball Grid Array 2D (FCBGA 2D) and other advanced packaging technologies for future scaling. Read the blog to learn more about Pando Tree. https://intel.ly/3KVVlw1 #Research #QuantumComputing #Foundry

Intel brings 11 research papers to the 2024 VLSI Symposium to showcase breakthrough research in various technologies. Discover the full range of research showcased, including Intel 3 process performance gains compared to Intel 4 and a millikelvin quantum research control chip to address bottlenecks. Don’t miss out on the latest developments from Intel Labs. Read the full blog post to learn more. https://intel.ly/3Vw6JDw #Technology #Research #Innovation

The Computer Vision and Pattern Recognition Conference (CVPR) 2024 is happening right now, and Intel Labs is showcasing some incredible research. We're sharing 24 research papers highlighting various breakthroughs in computer vision technology, including image sculpting, text-to-image diffusion models, long-form video understanding, and many more. Read the blog to learn more about each project. https://lnkd.in/gfCUjCWb #Research #ComputerVision #CVPR2024

Alleviate long-form video computation bottlenecks using Video Summarization with Graph Representation Learning (VideoSAGE). VideoSAGE is a novel approach to modeling video as a sparse graph, where the graph is constructed to enable interactions only between relevant nodes over time. When compared to transformers, this approach can aggregate context over 10x longer videos, consuming 3x peak memory usage and 5x faster inference. Learn more in part 1 of this 3-part series on long-form video representation learning from Subarna Tripathi. https://intel.ly/4eq8wTi #DeepLearning #Video #Developer #ComputerVision

Learn how neuromorphic computing is poised to change the landscape of everything from smartphones and PCs to autonomous vehicles and neuroprosthetics in this Diario EL PAÍS Q&A session with Mike Davies, Intel Labs Director of Neuromorphic Computing Research. https://intel.ly/3x6HRdF #Research #Computing #Technology

Explore how parallel computing and AI are helping machines improve our lives in this episode of the InTechnology podcast featuring Intel Sr. Fellow & Director of Parallel Computing Research, Pradeep Dubey, with host Camille Morhardt. https://intel.ly/4ed7oSO #ArtificialIntelligence #MachineLearning #ParallelComputing