Graph + AI Summit – Day 2 Recap

• Day 2 General Session with Pinterest, NewDay, and WarnerMedia – Here are the key insights from Day 2’s general session:
  • Dr. Jure Leskovec, Chief Scientist at Pinterest and Professor at Stanford University shared GraphSAGE, a new way to improve machine learning for new domains with sparse training data. Dr. Leskovec shared a couple of examples – the application of the approach at Pinterest (PinSAGE) and application in pharmaceuticals to predict side effects of combining multiple drugs. 
  • Danny Clark, Head of Fraud Strategy at NewDay presented their deployment of TigerGraph to connect and analyze various types of data including prior fraud cases, account, IVR, e-servicing, and digital wallets information from Experian, Threat Matrix, RSA Security, emailage, TeleSign, and FICO Falcon Intelligence Network.
    • “Initial rollout of TigerGraph has reduced the undetected fraud cases by 10%-15% and we are planning additional enhancements in fraud detection and rollouts in other areas of the business.” – Danny Clark, Head of Fraud Strategy, NewDay
    • “We evaluated multiple leading graph database vendors and conducted a proof of concept with three suppliers. We chose TigerGraph because of the scalability, simplicity, analytical capability, flexibility of the platform along with excellent support and engagement from TigerGraph team.”  – Danny Clark, Head of Fraud Strategy, NewDay
  • Matt Teshera, Sr. VP Enterprise Data Solutions, WarnerMedia covered the crucial business drivers shaping the media and entertainment industry and recommended looking at an identity graph based on Graph as a way to understand customers better and improve engagement. Matt shared that they have implemented an identity graph at Xandr, and are using Graphs for identity graph and recommendations at the HBO business unit of WanerMedia. 

• • We had two executive roundtables following the general session on day 2:
    • The Future of Graph Query Languages – A lively discussion on the evolution of ISO standard, GQL for Graph
    • Accelerating Graph-based Analytics and Machine Learning with Next Generation Hardware –  Intel, Dell, and Xilinx leaders shared their insights on how to accelerate these workloads by 10-100x with purpose-built hardware

• • Neena Sathi of Applied AI Institute shared her insights on building a business value use case for Graph + AI in the session Ramping up on knowledge graphs with TigerGraph and AI

We had an amazing lineup of technical breakout sessions on April 22, with five tracks running in parallel: 

• • 1. Graph + AI In the Cloud – This track had two amazing AWS speakers – Jack Secord (Analytics Guru) and Phi Nguyen (ML guru) who shared how to combine AWS, Amazon SageMaker with TIgerGraph Cloud for various use cases
    2. Visualization of Connected Data for Analytics and AI – Bethany Lyons, Sr. Product Manager of Tableau presented a session, “Visualizing Connected Data” with Dr. Victor Lee showcasing the joint Tableau and TigerGraph solution. Web App Hackathon winners presented their solution.
    3. Data Science with Graph Algorithms and Machine Learning –  Xinyu Chang of TigerGraph shared the new machine learning feature sets for fraud detection, Gojek, the shared economy giant from South Asia shares real-time fraud detection with Graphs, including benchmark results comparing multiple Graph vendors and key considerations in the final product selected. We also have sessions on improving cybersecurity, data lineage, and an overview of key graph analytics patterns for financial crimes detection from Graham Ganssle of Expero. 
    4. Combining NLP with Graph Analytics and AI -Amazing sessions from the TigerGraph community on integrating RASA pipeline with TigerGraph Cloud and how to create a semantic graph from the medical documents with SpaCy, open-source NLP tool + TigerGraph
    5. Enterprise Ecosystem for Graph + AI – Rayees Pasha and Xinyu Chang demonstrate the new connector for Snowflake,  We had in-depth sessions on hardware-accelerated graphs from Intel and Dell, containerization of Graph workloads, and building a no-code machine learning pipeline for Graph 

We have an exciting Day 3 for Graph + AI Summit tomorrow with three live workshops on analytics and machine learning with TigerGraph Cloud. 

TigerGraph released an updated “What is TigerGraph” video, which explains the core value proposition of a graph database in three minutes – Connect, Analyze and Learn from Data and share the core differentiators of the TigerGraph platform.