ML/AI Vendor Roundup October 2018
Hello, everyone! It’s time for another roundup of machine learning and AI news. Product enhancement, customer reference, partnership, acquisition, or other significant contributions. No hype!
Vendors listed below in alphabetical order.
AWS Machine Learning
AWS announces enhancements to SageMaker’s built-in image classifier.
AWS announces the availability of a PyTorch 1.0 preview in SageMaker and Amazon AMIs.
Big Squid seems to be struggling commercially. No new hires since April. Layoffs rumored.
Meanwhile, morale at Cloudera has plunged since I departed. Really, people — you can still function without me.
Under its Feature Factory initiative for automated feature engineering, Dataiku releases an Events Aggregator plugin. The EventsAggregator plugin works with data where each row records a specific action, or event, and the timestamp of when it occurred. The plugin can generate more than 20 different types of aggregated features per aggregation group.
Databricks partners with RStudio to add support for R to the MLflow machine learning toolkit.
DataRobot lands a Series D round of $100 million. Meritech and Sapphire Ventures led the round. DFJ Growth, New Enterprise Associates, and IA Ventures also participated. (I work for DataRobot.)
Domino Data Lab
Domino Data Lab announces Domino 3.0. The new release includes new capabilities for model deployment. These include automatic infrastructure provisioning via Docker; a portal to catalog model products, and automated versioning.
Google Cloud Platform
Google releases a 60-minute training module on machine learning fairness.
H2O.ai releases Driverless AI Version 1.4, which includes minor enhancements and bug fixes. Multi-node training added to the roadmap for Version 1.5.
H2O.ai announces the availability of H2O and Driverless AI in the Azure Marketplace.
H2O.ai and Snowflake announce that they will jointly support a connection between Snowflake Database, H2O and Driverless AI.
H2O.ai “sells out” 70% of the seats for H2O AI World London, according to a source who attended. The event included a surprise speaker, Dr. Dan Rubenstein of the Princeton Institute, who spoke about saving zebras with image recognition. Dr. Rubenstein may be saving zebras, but he’s not saving them with Driverless AI. Driverless AI does not work with image data.
IBM announces that it plans to incorporate the RAPIDS open source library for GPU-accelerated machine learning into several products, including PowerAI on IBM POWER9, IBM Watson Studio, and compute instances in IBM Cloud.
Mathworks announces support for NVIDIA’s DGX HPC boxes and NVIDIA Cloud.
NVIDIA announces RAPIDS, an open source GPU-acceleration platform for data science and machine learning. Among those endorsing the project: Anaconda, Cisco, Dell EMC, H2O.ai, IBM, Lenovo, MapR, SAP, and — get this — SAS. Walmart signs on as a “corporate community adopter.”
Oracle announces plans to support the NVIDIA HGK-2 platform for machine learning and AI on Oracle Cloud. Oracle also announces plans to support the RAPIDS open source library for machine learning on NVIDIA GPUs.
RapidMiner introduces RapidMiner AI Cloud, a SaaS offering that delivers a subset of RapidMiner platform capabilities. The first application in AI Cloud is an automated modeling service, currently in beta. The press release does not disclose planned general availability or pricing.
SAP announces the expansion of SAP Leonardo Machine Learning capabilities, including upgrades to the SAP Conversational AI service, and five new services planned for 2018. SAP’s machine learning offering will be available on Google Cloud Platform in November 2018.
SAS announces SAS for Containers, for some reason. The product’s been available for two years.