Truth and Consequences: Strata NYC Focuses on Promise and Potential Pitfalls of AI and Machine Learning

Truth and Consequences: Strata NYC Focuses on Promise and Potential Pitfalls of AI and Machine Learning

This year’s Strata New York marked the 9th year for the event, focusing on how data-driven businesses can harness the value of Big Data. The keynote speakers lead many interesting discussions on artificial intelligence and machine learning applications, from saving snow leopards in the Himalayas to Carnegie Mellon’s fascinating efforts at making robots able to communicate and interact with humans.

Data Doesn’t Lie, But Sometimes It’s Biased

There were also some sobering thoughts on the dangers of relying too much on what the machine tells us. Author Cathy O’Neil reminded us that math models can be biased and subjective. Models trained on past data can easily propagate existing prejudices. She gave various examples about how these models are going to have a big impact on people’s lives by determining who gets a job, a credit card, insurance, tenure at a teaching job, etc.

The New Inquiry editor Sam Lavigne echoed that sentiment when talking about predictive policing being a self-fulfilling prophecy. Princeton Professor Joanna Bryson suggested that artificial intelligence be regulated to follow established policies, similar to building code regulations for example.

Download our eBook: The New Rules for Your Data Landscape

Aligning Technology with Business Value

Ted Malaska shared some great lessons learned in his new role at Blizzard Entertainment, where there is limited amount of time to make decisions, and mistakes can be damaging to the project and to careers. When faced with multiple competing voices from business requirements, technology hype, a desire to play with new tools/build up your resume, and competing messages from vendors, his advice was simple: The only thing that should matter is the alignment of technology choices with business value. He cautioned that your passions may mislead you.

At Syncsort we take that advice to heart. Syncsort was founded in 1968 and has remained relevant and vibrant by listening to our customers and helping them meet new challenges and create more value for their customers in turn.

Keith Kohl, Vice President, Product Management at Syncsort, discussed the pressing challenges of creating a trusted, single view of data that can meet compliance requirements. By building data quality into the Enterprise Data Hub, duplicate records can be identified and standardized, giving the business a complete view of their customers and also ensuring that the right information will be deleted when there is a GDPR request to erase personal data.

Truth and Consequences: Strata NYC Focuses on Promise and Potential Pitfalls of AI and Machine Learning

This slide from Keith Kohl’s Strata presentation depicts the challenges of creating a trusted, single view of data that can meet compliance requirements.

By sharing lineage information with Metadata Catalog utilities, data elements can be traced and their origins can be proven, helping with data trust, compliance and audit needs.

Truth and Consequences: Strata NYC Focuses on Promise and Potential Pitfalls of AI and Machine Learning

Also from Keith’s Strata NYC presentation: An example of Data Lineage with Syncsort DMX-h

Technical Innovation and Announcements

On the technical side, Cloudera announced its Fast Forward Labs, aimed at making it easier to build Machine Learning applications and determine the best algorithm for a given business problem.

The other main announcements and discussions were around security, governance, and cloud. Mike Olson, Chief Strategy Officer at Cloudera, announced SDX-Shared Data Experience, which gives a consistent picture of data across all deployments, on-premise and in the cloud. It includes a data catalog, security, governance, workload management, and flexible ingestion and replication. The framework is extensible, so Cloudera-certified applications such as DMX-h can be brought into the cluster.

Cloudera also announced that Altus now supports Azure along with AWS.

Europe’s General Data Protection Regulation (GDPR) is scheduled to go into effect on May 25, 2018. Security, data quality, and data governance were brought to the forefront of the discussions at Strata New York. Nick Curcuru from Mastercard pointed out the growing role of the cloud to create a distributed Enterprise Data Hub in order to keep data from leaving a country’s borders. Mark Donsky and Steven Ross also had a very good session discussing the challenges of how to find the data, how to demonstrate compliance, and how to ensure data is erased when consent is revoked by individuals.

Truth and Consequences: Strata NYC Focuses on Promise and Potential Pitfalls of AI and Machine Learning Truth and Consequences: Strata NYC Focuses on Promise and Potential Pitfalls of AI and Machine Learning
Two members of Syncsort’s executive leadership team made appearances on theCUBE during Strata NYC. CEO Josh Rogers also sat down with host John Furrier to discuss the Big Iron to Big Data challenge and how Syncsort is growing and investing to take advantage of market opportunities. He also touched on working with the company’s new owners, Centerbridge Partners.

During her interview, Syncsort CTO, Tendü Yoğurtçu, PhD talked about the value Syncsort brings to enterprises that need to access and integrate quality data in a trusted manner, whether on premise or in the cloud, bridging the gap between legacy platforms and modern data architectures. She also discussed the recent announcement of Syncsort’s new Trillium Quality for Big Data capability.

For additional insights into how business alignment to Big Data is changing the rules, download our eBook: The New Rules for Your Data Landscape today!

Fernanda Tavares

Authored by Fernanda Tavares

Vice President, Data Integration R&D
1 comment

Leave a Comment

*