rights reserved. Long History of ML at Amazon Thousands of Engineers Across the Company Focused on AI Personalized recommendations Inventing entirely new customer experiences Fulfillment automation and inventory management Drones Voice-driven interactions
ML problem framing Data collection Data integration Data preparation & Cleaning Feature engineering Model training & parameter tuning Model evaluation Are business Goals met? Model deployment Monitoring & debugging Predictions Yes No Data augmentation Feature Augmentation Re-training
Data collection Data integration Data preparation & Cleaning Feature engineering Model training & parameter tuning Model evaluation • Setup and manage Notebook Environments • Setup and manage Training Clusters • Write data connectors • Scale ML algorithms to large datasets • Distribute ML training algorithm to multiple machines • Secure model artifacts
Data collection Data integration Data preparation & Cleaning Feature engineering Model training & parameter tuning Model evaluation Are business Goals met? Model deployment Monitoring & debugging Predictions Yes No Data augmentation Feature Augmentation Re-training • Setup and manage model inference clusters • Manage and Auto-Scale model inference APIs • Monitor and debug model predictions • Models versioning and performance tracking • Automate new model version promotion to production (A/B testing)
rights reserved. A m a z o n S a g e M a k e r A managed service that provides the quickest and easiest way for your data scientists and developers to get ML models from idea to production
rights reserved. UX … or Apache Spark through EMR and the SageMaker Spark SDK... Use SageMaker‘s hosted Notebook Instances... ... or SageMaker‘s Console for a point and click experience... ... or your own device (EC2, laptop, etc.)
rights reserved. Built-in Algorithms: Supervised Training XGBoost, Factorization Machines, Linear Learner for regression and classification DeepAR for time series forecasting Image classification with convolutional neural networks Seq2seq for translation
rights reserved. Built-in Algorithms: Unsupervised Training K-Means and PCA for clustering and pre- processing LDA and NTM for topic modeling Random Cut Forest for anomaly detection abc <0.19, 0.41, 0.12> xyz <0.16, 0.22, 0.93> BlazingText for text pre-processing
rights reserved. TensorFlow and MXNet Containers … explore and refine models in a single Notebook Instance … deploy to production Sample your data… Use the same code to train on the full dataset in a cluster of GPU instances…
rights reserved. Bring Your Own Algorithm ... add algorithm code to a Docker container... Pick your preferred framework... ... publish to ECS Amazon ECS
rights reserved. Hyperparameter Optimization (Preview) Run a large set of training jobs with varying hyperparameters... ... and search the hyperparameter space for improved accuracy.
rights reserved. Demos 1. SageMaker built-in image classification without transfer learning. 2. SageMaker built-in image classification with transfer learning. 3. SageMaker MXNet and Gluon distributed training. 4. Bring your own algorithm – SciKit Learn.
SageMaker: https://aws.amazon.com/sagemaker/ • Use the Amazon SageMaker high-level SDK: • For Python: https://github.com/aws/sagemaker-python-sdk • For Spark: https://github.com/aws/sagemaker-spark • SageMaker Examples: https://github.com/awslabs/amazon-sagemaker-examples • Let us know what you build!
namiki64
shumei_ito
recruitengineers
teamlab
pfn
akhino
takonda
persolserverworks
yamanoku
kakehashi
usanchuu
gr2m
asonas
szymonslowik
tomoyanonymous
morganepeng
aki_iinuma
afnizarnur
marktimemedia
mkilby
lauravandoore
jnunemaker
sabderemane
