Nvidia Corporation · NASDAQ:NVDA
Nvidia — J.P. Morgan Healthcare Conference 2020
22 slides from Nvidia's conference presentation, Jan 16, 2020
Slide 1: NVIDIA Healthcare
NVIDIA HEALTHCARE Kimberly Powell | VP & GM Healthcare | JP Morgan Healthcare 2020
Slide 2: Safe Harbor: Forward-Looking Statements
SAFE HARBOR Forward-Looking Statements Except for the historical information contained herein, certain matters in this presentation including, but not limited to, statements as to: our strategies, growth, market opportunities and position; the performance, benefits and impact of our products and tec
Slide 3: NVIDIA at the Intersection of Graphics, HPC, AI
NVIDIA AT THE INTERESCTION OF GRAPHICS, HPC, AI GRAPHICS | HPC | AI 3
Slide 4: One Architecture — Multiple Domains
NVIDIA ONE ARCHITECTURE — MULTIPLE DOMAINS Gaming Cloud Computing Automotive Manufacturing Healthcare $150B $230B $10T $14T $7T GRAPHICS | HPC | AI 4
Slide 5: AI Breakthroughs in Graphics
NVIDIA AI PROJECT SOL: MINECRAFT RTX: BREAKTHROUGHS A Showcase for the Power of NVIDIA RTX Real-time Ray Tracing in the World’s Most Popular Game IN GRAPHICS OMNIVERSE: NASA MARS LANDER: A Powerful Collaboration Platform for 3D Design Visualizing NASA’s Supercomputer Simulations 5
Slide 6: AI Computing Platform
NVIDIA AI COMPUTING PLATFORM DRIVE METRO ISAAC CLARA AERIAL Cars Factory Robot Health 5G GPU DGX HGX RTX GFN HPC AI ML CUDA-X CPU CUDA EGX AGX CSP OEM X-FACTOR / SPEED-UP FULL STACK CLOUD TO EDGE SYSTEMS EVERYWHERE 6
Slide 7: NVIDIA Is the Engine of AI
NVIDIA IS THE ENGINE OF AI #1 ON MLPERF AI Compute Growing Exponentially From 600 to 2 Hours in 5 Years K80 600 Hours SERVER DGX 2 Hours Time to Train (ResNet-50) COMPUTE DOUBLING EVERY 3.4 MONTHS SUPER MOORE’S LAW 7
Slide 8: AI Renaissance
AI RENAISSANCE WRITING IMAGE GENERATION DIALOG 3D POSE CLASSIFICATION DENOISING TRANSLATION SEGMENTATION SUMMARIZATION ALEXNET OBJECT RECOGNITION BERT CNN TRANSFORMER Q&A CLASSIFICATION 2012 “See” 2018 “Understand” 8
Slide 9: A Decade in Healthcare
NVIDIA A DECADE IN HEALTHCARE AI MITOSIS DETECTION CLARA AGX DEVELOPER KIT DESKTOP SEQUENCING PROJECT CLARA CLARA AI HIV-1 CAPSID FRAMEWORK AMNIOSCOPIC RENDERING GANS MRI CCDS GRAPHICS ULTRASOUND 4D BRATS CHALLENGE AI & HPC EJECTION CT RECON FRACTION 2010 2015 9
Slide 10: Artificial Intelligence Is Domain Specific
ARTIFICIAL INTELLIGENCE IS DOMAIN SPECIFIC World Sense See, Understand Automation AI Program Self-Driving Computer 10
Slide 11: Artificial Intelligence Is Domain Specific
ARTIFICIAL INTELLIGENCE IS DOMAIN SPECIFIC World Sense See, Understand Automation AI Program Self-Driving Computer AI Program Manufacturing Computer 11
Slide 12: Artificial Intelligence Is Domain Specific
ARTIFICIAL INTELLIGENCE IS DOMAIN SPECIFIC World Sense See, Understand Automation AI Program Self-Driving Computer AI Program Manufacturing Computer AI Program Radiology Computer 12
Slide 13: Clara: Domain Targeted AI Application Framework
NVIDIA CLARA DOMAIN TARGETED AI APPLICATION FRAMEWORK Pre-trained NVIDIA CLARA AI APPLICATION FRAMEWORK AI Models Transfer NVIDIA AIAA Learning TENSORRT Tool Your Data LABEL ADAPT COMPILE AI Application Clara Runtime HGX, EGX, AGX 13
Slide 14: Smart Medical Device Revolution
SMART SMART SMART SENSING IMAGING TREATMENT SMART MEDICAL DEVICE REVOLUTION 14
Slide 15: Clara: Federated Learning for Collaborative AI
NVIDIA CLARA FEDERATED LEARNING FOR COLLABORATIVE AI w NVIDIA CLARA AI APPLICATION FRAMEWORK Transfer Federated NVIDIA AIAA Learning Learning TENSORRT Hospital Tool Tool A, B, C LABEL ADAPT COMPILE COLLABORATE AI Application Clara Runtime HGX, EGX, AGX 15
Slide 16: Healthcare Industry Adopting Federated Learning
HEALTHCARE INDUSTRY ADOPTING FEDERATED LEARNING PHARMA PARTNERS PUBLIC PARTNERS This project has received funding from the Innovative Medicines Initiative 2 Joint Undertaking under grant agreement N831472. This Joint Undertaking receives support from the European Union’s Horizon 2020 research and in
Slide 17: Clara AI Application Framework for Imaging
NVIDIA CLARA AI APPLICATION FRAMEWORK FOR IMAGING NGC Ecosystem Partners CUDA-X PRE-TRAINED MODELS DOMAIN TARGETED AI ECOSYSTEM ADOPTION 17
Slide 18: Computational Genomics Inflection Point
COMPUTATIONAL GENOMICS INFLECTION POINT Cost Decreasing, Throughput Increasing $35,000 25000 $30,000 20000 Genomes/Sequencer $25,000 $/Genome 15000 $20,000 2x Genomics Data $15,000 Every 7 Months 10000 $10,000 5000 $5,000 $0 0 2010 2015 2020 200M WGS by 2025 2 Trillion CPU Hours for Variant Calling
Slide 19: Parabricks: GPU Accelerated Genomics Analysis Toolkit
NVIDIA PARABRICKS GPU ACCELERATED GENOMICS ANALYSIS TOOLKIT PARABRICKS CUDA-X Equivalent Results to Industry Standard GATK4 | 30-50X Faster Processing from NGS to Variant Calling Pipeline Integrates DeepVariant AI Variant Calling Parabricks Available Now on NGC 19
Slide 20: Breakthroughs in Computational Genomics
BREAKTHROUGHS IN COMPUTATIONAL GENOMICS LIFE SCIENCE SUPERCOMPUTER CANCER GENOMICS NMF FOR MUTOGRAPHS 60 Genomes/Day Faster, More Accurate 30X Faster AI BASECALLER DE NOVO ASSEMBLY SINGLE CELL ATAC-SEQ NVIDIA Speech AI Basecaller GPU Accelerated Mapping, Alignment, Polishing AI Denoising 20
Slide 21: Healthcare Accelerated Computing
HEALTHCARE ACCELERATED COMPUTING NVIDIA Clara NVIDIA Parabricks DGX HGX EGX AGX SMART MEDICAL DEVICES GRAPHICS | HPC | AI ECOSYSTEM BREAKTHROUGHS 21
Slide 22: .
NVIDIA.