[AWS Dev Day] 실습 워크샵 | AWS IoT와 SageMaker를 활용한 예지 정비의 구현하기
0 likes1,641 views
The document discusses the implementation of predictive maintenance using AWS IoT and SageMaker, emphasizing AWS Greengrass for edge computing and AWS IoT Analytics for data analysis. It highlights the capabilities for local data processing, machine learning model execution, and the seamless integration of cloud and edge resources. Additionally, it details the benefits of Amazon SageMaker for building, training, and deploying machine learning models at scale, supporting various programming languages and flexible model structures.
![[AWS Dev Day] 실습 워크샵 | AWS IoT와 SageMaker를 활용한 예지 정비의 구현하기](https://image.slidesharecdn.com/awsdevdayseoul2019iotpredictivemainternanceworkshop-190930065537/85/AWS-Dev-Day-AWS-IoT-SageMaker-1-320.jpg)
![[AWS Dev Day] 실습 워크샵 | AWS IoT와 SageMaker를 활용한 예지 정비의 구현하기](https://image.slidesharecdn.com/awsdevdayseoul2019iotpredictivemainternanceworkshop-190930065537/85/AWS-Dev-Day-AWS-IoT-SageMaker-14-320.jpg)
![[AWS Dev Day] 실습 워크샵 | AWS IoT와 SageMaker를 활용한 예지 정비의 구현하기](https://image.slidesharecdn.com/awsdevdayseoul2019iotpredictivemainternanceworkshop-190930065537/85/AWS-Dev-Day-AWS-IoT-SageMaker-17-320.jpg)
[AWS Dev Day] 실습 워크샵 | AWS IoT와 SageMaker를 활용한 예지 정비의 구현하기
- 2. AWS IoT와 SageMaker를 활용한 예지 정비의 구현하기 권신중 솔루션즈아키텍트 최원근 솔루션즈아키텍트 이종화 솔루션즈아키텍트 송규호 솔루션즈아키텍트 김영진 솔루션즈아키텍트 김민성 솔루션즈아키텍트 현륜식 솔루션즈아키텍트 김준형 솔루션즈아키텍트
- 6. AWS Greengrass는 AWS Cloud를 Edge로 확장합니다. IoT 디바이스에서 발생하는 데이터는 로컬에서 처리하고 데이터 관리, 분석, 저장은 AWS Cloud에서 처리합니다. Data processed in the cloud Data processed locally AWS 클라우드 기능을 엣지까지 확장 AWS Greengrass
- 7. Features AWS Greengrass 데이터와 상태 동기화 보안 Over-the-air updates 프로토콜 어뎁터Local actions 로컬 메시지 브로커 머신러닝 유추 로컬 리소스 억세스 Local device shadows Lambda functions Local message broker High-quality AWS security Easily update AWS Greengrass core Local execution of ML models Lambda interacts with peripherals Easy integrations with local protocols ʥ A
- 8. AWS IoT Analytics 는 대규모 IoT 데이터에 대한 정교한 분석을 손쉽게 실행 및 운용할 수 있게 해주는 완전관리형 서비스 입니다. 디바이스 데이터에서 가치를 창출 AWS IoT Analytics IoT data는 잡음과 상당한 격차 및 잘못읽은 경우가 많음 이 데이터를 필터, 처리, 변환 및 보강 임시 쿼리 또는 정교한 IoT 분석 및 시각화 Raw 데이터를 저장하고 데이터 처리 수행 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 0 1 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 0 1 0 0 1 0 ENRICHMENT 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 0 1 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 1 0 1 0 0 1 1 0 0 1 0 0 1 0 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 1 0 1 0 0 1 1 0 0 1 0 0 1 0 0 1 0 1 0 1 0 1 1 0 0 1 0 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 1 0 1 0 0 1 1 0 0 1 0 0 1 0 0 1 0 1 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0101001101001 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 01 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 0101001010 101001 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 1 0 1 0
- 9. AWS IoT Analytics AWS IoT Analytics에서의 데이터 흐름
- 10. AWS IoT Analytics Collect AnalyzeStore Visualize 저장하고 분석하고 싶은 데이터만 수집 Raw 데이터를 의미있는 데이터로 변환 디바이스 데이터 분석을 위해 시계열로 저장되는 데이터 스토어에 저장 빌트인 된 IoT 분석 SQL 쿼리 엔진 혹은 Jupyter Notebook으로 데이터셋 분석 IoT 데이터셋을 빠르게 시각화하여 분석 Process Channels DatasetsPipelines Data stores Jupyter Notebooks AWS IoT Analytics에서의 데이터 흐름
- 11. Amazon SageMaker: Build, Train, and Deploy ML Models at Scale 1 2 3
- 12. Jupyter Notebook
- 13. ML HOL overview Classification using Mxnet and Gluon by HasAnomaly label Load data & Preprocessing - normalization Divide dataset to Training & Test – 8:2 Build Neural network Training model Prediction Make Inference
- 15. Native distributed training supported Supports distributed training on multiple CPU/GPU machines to take advantage of cloud scale Flexible programming model Supports both imperative and symbolic programming maximizing efficiency and productivity Portable from the cloud to the client Runs on CPUs or GPUs, on clusters, servers, desktops, or mobile phones Multi-lingual | No need to learn a new language Python, R, Scala, Julia, C++, Matlab, or Javascript Performance 0ptimized Optimized C++ backend engine parallelizes both I/O regardless of source language
- 16. Introducing Gluon Simple, easy-to- understand code Flexible, imperative structure Dynamic graphs High performance § Neural networks can be defined using simple, clear, concise code § Plug-and-play neural network building blocks—including predefined layers, optimizers, and initializers § Eliminates rigidity of neural network model definition and brings together the model with the training algorithm § Intuitive, easy-to-debug, familiar code § Neural networks can change in shape or size during the training process to address advanced use cases where the size of data feed is variable § Important area of innovation in natural language processing (NLP) § There is no sacrifice with respect to training speed § When it is time to move from prototyping to production, easily cache neural networks for high performance and a reduced memory footprint