Pre:Invent 2019 — Time to catch up!

AWS re:Invent is almost upon us, so here’s your last chance to catch up with recent AI and Machine Learning launches. Once the waves of new announcements start rolling down, it will be too late ;)

So here goes: this is everything you may have missed since mid-October 2019.

Happy to answer questions here or on Twitter.

Winter is coming. Good! Plenty of time to read AWS docs.

Applied ML

Use ML models in Amazon Quicksight.

Add ML predictions using Amazon SageMaker models in Amazon QuickSight

Use ML models in Amazon Aurora.

New for Amazon Aurora - Use Machine Learning Directly From Your Databases | Amazon Web Services

AI Services

Amazon Comprehend: Natural Language Processing

Real-time custom classification is now supported (documentation).

Amazon Comprehend launches real time custom classification

Amazon Forecast: Forecasting

You can now use any quantile, not just p10 / p50/ p90.

Amazon Forecast can now support generating predictions in any quantile

Increasing visibility into AutoML jobs.

Now get additional details/metrics around all your algorithm runs with AutoML

Amazon Lex: Chatbots

Sentiment analysis is available (documentation).

Amazon Lex Now Supports Sentiment Analysis

Lex complies with SOC 1/2/3.

Amazon Lex is Now SOC Compliant

Web and mobile chat integration is supported in Amazon Connect (documentation).

Amazon Connect Introduces Web & Mobile Chat for a True Omnichannel Contact Center Experience |…

Amazon Personalize: Personalization and Recommendation

Batch mode predictions are now supported (documentation).

Now available: Batch Recommendations in Amazon Personalize | Amazon Web Services

Amazon Polly: Text To Speech

A new ‘conversational’ style has been released for Neural TTS.

Amazon Polly launches Conversational Speaking Style voices

Two new voices are supported with Neural TTS: US Spanish and Brazilian Portuguese (documentation).

Amazon Polly Launches two new voices in Neural Text-to-Speech technology

We released a Wordpress plugin, with Amazon Cloudfront integration.

AWS for WordPress plugin now available and with new Amazon CloudFront workflow

Amazon Rekognition: Image And Video Analysis

You can now train models on custom labels (feature will be available on December 3rd).

Announcing Amazon Rekognition Custom Labels | Amazon Web Services

Amazon Translate: Neural Translation

22 new languages and 6 new regions (documentation).

22 New Languages And Variants, 6 New Regions For Amazon Translate | Amazon Web Services

Amazon Transcribe: Automatic Speech Recognition

15 new languages have been added in two batches (documentation).

Amazon Transcribe Now Supports Speech-to-text in 7 Additional Languages

Amazon Transcribe Now Supports Speech-to-text in 8 Additional Languages

Australian English is supported in streaming mode.

Amazon Transcribe Now Supports Australian English Speech-to-Text in Real Time

Alternative transcriptions are available, i.e. different interpretations of the transcribed audio.

Amazon Transcribe now Supports Alternative Transcriptions

Regional expansion: Asia Pacific (Tokyo), Asia Pacific (Hong Kong) and Middle East (Bahrain) regions.

Machine Learning Services

Amazon SageMaker

Multi-model endpoints (documentation)!

Save on inference costs by using Amazon SageMaker multi-model endpoints | Amazon Web Services

A new version of XGBoost is available in Amazon SageMaker (documentation).

New version of SageMaker XGBoost algorithm available

SageMaker is now available in the Middle East (Bahrain) region.

SageMaker Region expansion to Bahrain

AWS DeepRacer

AWS DeepRacer Update - New Features & New Racing Opportunities | Amazon Web Services

Infrastructure and Frameworks

Amazon Elastic Inference: Fractional GPU Acceleration

New accelerators with more GPU memory, woohoo (documentation).

Amazon Elastic Inference introduces new Accelerators with higher GPU memory

Deep Learning Containers

PyTorch is now supported (documentation).

AWS Deep Learning Containers now support PyTorch