邁向無處不在的人工智能時代.pdf

《邁向無處不在的人工智能時代.pdf》由會員分享，可在線閱讀，更多相關《邁向無處不在的人工智能時代.pdf（28頁珍藏版）》請在三個皮匠報告上搜索。

1、Toward the Era of AI EverywhereLOKWON KIMCEODEEPX2 2023 DEEPX Co.,Ltd.Rise of Edge AIEverything with wheels has been demonstrated to move autonomously.All camera or sensor companies demonstrated object detection solutions.45 2023 DEEPX Co.,Ltd.Blocking Points forAI EverywhereMismatch Between AI and

2、Edge devices Processing AI requires greater data and computations than any other algorithm created by humans.AI algorithms are rapidly evolving and becoming more intelligent,adding new mathematical operators and requiring more data and computations.By nature,AI is ill-suited for edge device applicat

3、ions.Need a solution that goes beyond limits of theory!2023 DEEPX6#1:Conventional Solution 2023 DEEPX7#2:Edge Applications Require SOTA AI ModelsSome NPU solutions improved performance and power efficiency,but 8#3:A 1%Accuracy Loss Is Too Much!Using 8-bit integer instead of 32-bit floating point is

4、the key for power and area efficiency on resource-constrained edges.Normally,results in AI accuracy drop compared to GPUs.With GPUs,edge application developers create AI models for their own applications.The actual intelligence of the application is determined by edge NPUs.This incurs another loop o

5、f the AI accuracy validation for commercial applications which incurs longer development time and effort.A 1%AI accuracy drop is too much for easy deployment of edge AI.2023 DEEPX9#4:A Single One-Size-Fits-All Solution?There are various AI edge device applications.Each edge device AI application has

6、 specific function,performance,and cost requirements.A family of chips with different functions,price,and performance is necessary for different applications.#5:Developing Edge AI Software TakesToo Much Time and Effort Edge AI application developers demand an easy-to-use SDK.(In the Computer Vision

7、Developer Survey conducted in late 2022 by the Edge AI and Vision Alliance,“algorithm implementation/optimization”was named as one of the most challenging aspects of product development(37%).)Some SDKs require several months to learn how to use.Some SDKs require understanding of hardware architectur

8、e and deep technical features.Not easy to deploy AI!12 2023 DEEPXDEEPX Disruptive StrengthsDEEPX MissionDEEPXs Key Differentiators 2023 DEEPX14densnetgooglenetmnasnetmobileNetResNetSSDYolov3,v4,v5,v7EfficientNet/DetBiseNetShelfNetPIDNetSFA3D+Other AI models(Model Zoo:170 models)A Successful Customer

9、 CollaborationThe power/performance efficiency of DEEPX NPU is 10X higher than GPU.The AI accuracy of DEEPX NPU is higher than GPUs.15A Comparison with a Reference Platform10X more performance-cost efficient10X more performance-power efficient16Product Roadmap(2023)DX-L1 Targets Single-camera AI App

10、licationsDX-M1 Targets Clustered Cameras or Workloads19AI Inference Server Solution for ESG/TCODXNN SDK for DEEPX NPUsEnjoy the SOTA AI on All DEEPX SolutionsRepresentative Project#1 2023 DEEPX23Representative Project#2 2023 DEEPX24Representative Project#3 2023 DEEPX25Product Strategy 2023 DEEPX26Su

11、mmaryEdge AI requires performance,low power,low cost,SOTA AI algorithms,AI accuracy,and ease of use.DEEPX NPU can:1.Run SOTA AI models.2.Get the best AI accuracy.3.Achieve the highest performance,power efficiency,and cost efficiency(including BOM cost).Please visit our demo booth and check it out!Thank you!Company Info,Demo and Additional Talk2023 Embedded Vision SummitAdditional Talks from DEEPX“State-of-the-Art Model Quantization and Optimization for Efficient Edge AI”(Hyunjin Kim,Wednesday,May 24,12:00 pm)281.Demo Booth:#1032.DEEPX Homepagehttps:/www.deepx.ai3.Linkedin&Youtube 2023 DEEPX