San Francisco Data Management Association (SFDAMA) Presents

DAMA DAY 2018.

Data Lake Applications and Management

Wednesday, 13th Dec 2018

8:00 AM – Breakfast and Networking

9:00 AM to 5:00 PM – DAMA Day Conference (Lunch Included)

Key Note Address:

Knowledge Graphs, Semantic Data Lakes and AI: The Next Generation of Data

By

David Newman
SVP, Strategic Planning Manager, Innovation R&D Group, Wells Fargo

In recent years, data management professionals are coming to the realization that the growing costs and inefficiencies of conventional data management solutions indicates that legacy information technologies may not be able to support the massive explosion of data within the enterprise. The enterprise drive towards big data and data lake solutions has only highlighted the scale of the challenge. This presentation will describe semantic knowledge graph technology, the "next generation of data". We will discuss how knowledge representation and reasoning capabilities using ontologies is the way forward to tame the data management monster. The Financial Industry Business Ontology (FIBO) will be described as an exemplar. We will also discuss what a semantic data lake architecture looks like. Lastly, we will discuss what a knowledge graph is and how ontologies and machine learning can operate together for maximum benefit.

AI for Self-Service Data

By

Ayush Parashar, 
Co-Founder and Vice President of Engineering at Unifi Software

Artificial Intelligence will fundamentally change the way that business users across an enterprise interact with data and gain business insights. AI will shift the balance of power from a few highly skilled and highly paid workers to the masses. Anyone with access to big data will be able to quickly comprehend what has happened with their business and predict their business outcomes, then visualize or consume that insight on a wide and growing range of devices and display types. In this brief introduction to AI applied to Big Data analytics, Ayush will describe some of the trends and what to expect from AI in 2018 and beyond and offer a look into Unifi's Data as a Service Platform for self-service access to data powered by AI.

Powering Machine Learning and Analytics with Logical Data Lake

By

Avinash Deshpande 
Chief Software Architect, Logitech

In the past few years, as Logitech ventured into smart home, video conferencing, gaming and mobile technologies, they needed to invest in data visualization, machine learning, sentiment analysis, IoT analytics, predictive analytics, natural language processing and data mining, to be able to offer their end customers the best user experience. While data became the most critical component for their growing business, it also grew in volume, variety and complexity. Even though Logitech moved their analytics to AWS cloud and created S3 based data lake, exponential growth of data and data silos required a different data lake solution. Logitech created a logical data lake with Denodo Platform for data virtualization for all of their analytics, reporting and machine learning purposes and the solution has since seen wider applications across the enterprise. Attend this presentation to learn:

· What is logical data lake and why it is important

· Why data virtualization and data visualization plays the most important role in Logitech’s modern data architecture

· Importance of logical data lakes in the future of machine learning and analytics

Applying Knowledge Based AI to Modern Data Management

By

Mani Keeran,
Head of Information Management, Franklin Templeton Investments

Gi Kim,
Information Architect, Franklin Templeton Investments

Preeti Sharma,
Information Architect, Franklin Templeton Investments

The last financial crisis and current regulatory needs highlight the importance of knowing the total exposures to counter parties. Knowing the exposures presupposes actionable metrics on consolidated transactions and risks, which may not be readily available in reality due to the fragmented nature of counter party data.

The case study will present an operational database that has multiple master tables for counterparties with different contexts from various business processes. Most of the regulatory requirements include aggregated numbers from various interactions with legal entities but implementing them is generally tedious work as the Relational data model does not fully describe relationships between siloed tables. This leads to expensive analysis of data and development of ETL logic regularly.

This PoC tests the hypothesis that the Knowledge based Artificial Intelligence technique using a Semantic Database can solve this problem effectively in less time and cost. A demo of the Graph DB based solution will be shown.