Top 5 most ROI driven ML releases from AWS re:Invent
The Best of AI from AWS re:Invent 2018.
AWS re:Invent is the super bowl of Infrastructure and cloud computing.
Every year, Amazon rolls out amazing features for their Amazon Web Services ecosystem. Since Artificial Intelligence and Machine Learning is getting more bigger every single day, at AWS re:Invent this year, there were some interesting announcements in this space. I prepared a list of the top ML releases for the Data Science/AI community from re:Invent this year.
I have categorized these into three main areas, Application Services Layer, ML Services Layer and Data Layer.
Application Services Layer
Amazon Personalize and Amazon Forecast— Amazon Personalize is a real-time personalization recommendation service, while Amazon Forecast offers time-series forecasting. This is purely Amazon using their expertise in AI and ML and building models for you. You provide Amazon with your data, stats about your business, amazon will select the top algorithms and models that would be suitable to your data, run them and provide the results via a recommendation API. The models remain private to you.
Amazon Personalization and Amazon Forecast are two services aimed at democratizing Machine Learning in the enterprise.
Amazon Forecast is time series forecasting for Amazon.com. Amazon Forecasting works like Personalization. A customer gives Amazon historical data like sales information. They also provide all the variables that could impact their sales forecast. Amazon does all the work, like selecting hyper-parameters and training the models providing a time series forecast.
To know more: Amazon Personalize, Amazon Forecast
ML Services Layer
This has the maximum number of interesting releases this year.
Amazon Sagemaker — Neo and RL: Without any manual intervention, SageMaker Neo optimizes models deployed on Amazon EC2 instances, Amazon SageMaker endpoints and devices managed by AWS Greengrass. Neo supports leading frameworks and algorithms in the ML world including TensorFlow, Apache MXNet, PyTorch, ONNX, and XGBoost. The biggest advantage is the speed boost that you get without any loss of accuracy while running your models. The AI community knows about the use cases for Reinforcement Learning(RL). RL is particularly suitable for complex, unpredictable, environments that can be simulated and where building a prior dataset would either be infeasible like autonomous vehicles, portfolio management, inventory management and robotics. SageMaker RL (Reinforcement Learning) builds on top of the existing SageMaker, adding pre-packaged RL toolkits and making it easy to integrate any simulation environment. The biggest advantage of RL is that you can focus on the RL problem and not on managing servers.
To know more: SageMaker Neo, SageMaker RL
AWS Marketplace for ML — This is the App store for ML Models and solutions. Anyone can build and package models and algorithms and sell them on the AWS marketplace with a self service option. AWS will provide the licensing and payment infrastructure to support the listings.
AWS Marketplace for ML will truly enable a model built by a Data Scientist in a small city in Africa to be deployed on a Dataset in the US.
Data Layer
Data Lakes
Everybody wants a data lake — AWS CEO Andy Jassy
As a Data Scientist, we are all aware that there is inherent value in consolidating data. We all want to consolidate and have all the data on S3. For that, we would need to configure S3, move the data from different sources, clean it up and find an optimized strategy to move it. Once the setup is done, we would need access and permission setup for it. AWS Lake formation automates this with a few clicks. It takes the menial work out of the process and lets us concentrate on the harder problems.
Conclusion
AWS also released other ML tools like Elastic Inference, SageMaker Ground Truth and DeepRacer. The top five mentioned above have the most ROI driven impact to a business or a Data Science team. All the ML tools are aimed to help Data Science Managers and Data Scientists with saving time dedicated to managing infrastructure and helping them by complementing their work. The messaging is clear, AWS has entered into the ML infrastructure game with a heavy arsenal against its competitors. In the end, it is the Data Scientists and Data Science Managers that will gain the most out of this competition.
Subscribe to our Acing AI newsletter, I promise not to spam and its FREE!
Thanks for reading! 😊 If you enjoyed it, test how many times can you hit 👏 in 5 seconds. It’s great cardio for your fingers AND will help other people see the story.
All data is sourced from online public sources. I aim to make this a living document, so any updates and suggested changes can always be included. Please provide relevant feedback.
