I was pleased to be on-hand for the recent (Feb.
4-5) H2O World 2019 conference in beautiful San Francisco.
This was my first H2o.
ai conference and I was looking forward to drilling down into their popular open source solutions for data scientists including the H2O machine learning platform, Sparkling Water machine learning platform on Spark, H2O4GPU accelerated AI for GPUs, as well as the groundbreaking Driverless AI automatic machine learning (AutoML) platform.
With all this leading-edge technology and resulting industry buzz, I can see why the company recently ranked #9 on insideBIGDATA’s IMPACT 50 list of the industry’s most impactful companies.
The 1,350 attendees (the most ever at an H2O World event) were excited and the energy was high.
The venue at the San Francisco Hilton on Union Square was an excellent place for technology, being in close proximity to all the tech ventures in San Francisco and nearby Silicon Valley.
Milling around in the crowd, at lunch, and waiting in line at one of the omnipresent espresso carts, I heard really positive comments about the show and the H2O technology.
It’s a good sign when everyone is beaming while downing a freshly made cup of Joe.
Plus all the tables outside of the session rooms were fully staffed with H2O personnel to answer any and all questions.
I was impressed with the breath of everyone’s knowledge of the products.
I spoke at length with some H2O people about the Driverless AI approach.
I asked who the target user was for this solution, and the consensus was “the data scientist” even though it does a lot of the work of a data scientist.
They’re not looking to replace the data scientist, just supplement him/her (since you still need the technology behind machine learning to use the product effectively).
The idea is to stretch the resources that one data scientist can bring to the table by automating a lot of the work like selecting, and testing models, feature engineering, accuracy checking, etc.
In a day where there are not enough data scientist to fill demand, a product like Driverless AI is a welcome addition to the data scientist’s arsenal.
Life of an Analyst I really enjoy attending company conferences especially those put on by leaders in the big data ecosystem.
By everything I hear on the street, H2O is definitely an innovator and leader.
During the keynote addresses, I was front center to get a sense for the vibe of the tech and the execs.
H2O thoughtfully provided a front row table for press/analysts.
I really appreciate this, as I like being close to the action for two reasons.
One, I like to see facial expressions, and body language of the presenters as it gives me a closer sense for the reality of the corporate message (e.
CEO and Co-founder SriSatish Ambati’s sincerity and passion was evident).
And also, I like being close enough to the presentation screen to take quick pics on my iPhone.
Driverless AI Marches On The company took the opportunity to announce new and innovative capabilities for its data science and machine learning platforms, H2O, Sparkling Water, and H2O Driverless AI, and to address the critical scalability and performance needs of all organizations.
As part of these new capabilities, and to further the company’s mission to democratize AI, H2O.
ai has added several new algorithms that address common use cases that customers need today.
The latest release of its award-winning “AI to do AI” automatic machine learning platform, H2O Driverless AI launched a number of new features including: Model checkpointing: This enables customers to re-train models quickly without restarting each time.
Enhanced capability to handle massive workloads: The latest version of Driverless AI improves deployment speed on existing infrastructure while significantly lowering memory footprint.
Support for new algorithms: Along with XGBoost, Tensorflow, GLM and RuleFit, LightGBM and FTRL are the latest supervised algorithms to be added to Driverless AI.
It also uses K-Means, SVD, PCA and other unsupervised algorithms for feature engineering.
The combination of these algorithms ensures that customers now have more options to solve their data science problems and can address a variety of use cases ranging from credit risk scoring, anti-money laundering, customer churn predictions, fraud detection, cyber threat prevention and more.
ai’s Driverless AI platform harnesses the power of NVIDIA Tensor Core GPUs to enable customers to see significant speedups in automated machine learning to deliver AI insights and interpretability,” said Jeff Herbst, Vice President of Business Development at NVIDIA.
“H2O Driverless AI continues to gain momentum as it offers advanced capabilities that solve real world problems for customers today.
” New Partnerships Galore The conference was the perfect time for the company to announce several new strategic partnerships: H2O.
ai announced a new strategic collaboration with Intel to accelerate AI adoption in the enterprise.
At the heart of this collaboration is Project Blue Danube, a co-innovation project focused on accelerating H2O.
ai technologies on Intel platforms, including the Intel® Xeon® Scalable processor.
In addition, the two companies will partner on community and ecosystem growth, advancing data science education through Intel® AI Academy, and support developer efforts to integrate AI into mainstream analytics workflows to deliver compelling TCO, performance and simplicity.
The combination of H2O.
ai’s automated machine learning technology and Intel Xeon Scalable processors allows enterprise organizations to quickly generate the information necessary to make critical business decisions and ultimately gain a true competitive advantage based on platforms that enterprises know and trust.
ai and Kx, a division of First Derivatives (FDP.
L) and provider of the time-series database, kdb+, announced a strategic partnership to integrate Kx technology with H2O.
ai’s automatic machine learning platform, H2O Driverless AI.
H2O Driverless AI performs the function of an expert data scientist and includes automatic feature engineering, modeling, visualization and interpretability within a single platform.
Kx technology, powered by kdb+, is used for large-scale streaming, real-time and historical data analytics.
It is widely implemented in the financial services industry for complex analytics on large-scale trading data, as well as in the manufacturing industry and other industries with high volume, high velocity time series IoT data.
The integration of kdb+ into H2O Driverless AI will enable the machine learning and data science platform to extend its ability to quickly process vast data sets, allowing faster identification of AI models and more performant predictive capabilities.
Fireside chat between Alteryx CEO, Dean Stoecker and Sri Ambati, CEO and founder at H2O.
ai announced a strategic collaboration with Alteryx, Inc.
(NYSE: AYX), revolutionizing business through data science and analytics, that will integrate H2O.
ai’s enterprise automatic machine learning platform, H2O Driverless AI, with the Alteryx Platform.
This new integration, announced at H2O World San Francisco and discussed during a fireside chat between Alteryx CEO, Dean Stoecker and Sri Ambati, CEO and founder at H2O.
ai, enables users of both platforms to simplify and operationalize their end-to-end data science workflows – from data acquisition and prep and blend, to advanced predictive modeling and machine learning, to model deployment.
ParallelM, a leader in MLOps, announced integration with H2O open source to drive the adoption of AI across industries.
The integrations will allow H2O.
ai customers to quickly deploy and manage models in ParallelM MCenter and manage ongoing lifecycle needs like model health monitoring and model retraining for models running in production.
My only complaints about H2O World is that it could have been 1 day longer.
There was definitely enough content to expand into 3 full days.
Nevertheless, H2O has sold me in their inventiveness in the industry, and I’d love to return to H2O World again next year.
Contributed by Daniel D.
Gutierrez, Managing Editor and Resident Data Scientist of insideBIGDATA.
In addition to being a tech journalist, Daniel also is a practicing data scientist, author, educator and sits on a number of advisory boards for various start-up companies.
Sign up for the free insideBIGDATA newsletter.