Combined Graph and Machine Learning Solutions Power Product and Service Recommendation Engines, Supply Chain Planning, Fraud Detection Use Cases
REDWOOD CITY, CA – Sept. 30, 2020 – TigerGraph, the only scalable graph database for the enterprise, today announced that it will showcase multiple key partnerships that are democratizing the use of graph with AI and machine learning at Graph + AI World 2020. The event, organized and hosted by TigerGraph, will feature demonstrations, sessions and discussions with key partners using TigerGraph for select deployments, including hardware-accelerated machine learning and business-friendly graphical user interfaces. These deployments, in turn, are powering next-generation supply chain planning, fraud detection and customer 360/customer journey use cases.
“The Xilinx and TigerGraph partnership forms a next-generation machine learning solution that unites cutting edge hardware with graph analytics on connected data to solve real-world problems,” said Dan Gibbons, vice president, Datacenter Software and Compute Solutions Development, at Xilinx Inc. “TigerGraph and Xilinx understand that these forward-looking technologies have true human relevance — from understanding and planning a patient’s wellness journey to identifying which groups of users may be part of a fraud or money laundering ring. TigerGraph is the leading scalable graph database for the enterprise that provides in-database machine learning for delivering next-generation solutions — and solving problems — within all industries.”
TigerGraph is working with Xilinx, developer of highly flexible and adaptive processing platforms and inventor of the Field Programmable Gate Array (FPGA), to deliver two hardware- accelerated machine learning solutions that combine Xilinx’s Alveo U50(™) accelerator card with TigerGraph’s graph database and analytics platform. One solution, for recommendation engines, focuses on the Cosine similarity algorithm to find similar customers while the other solution, for fraud detection, uses community detection based on the Louvain algorithm to detect fraud and money laundering rings. Xilinix presents these joint solutions during two Graph + AI World sessions that showcase:
- An FPGA-enabled recommendation engine that performs fast parallel similarity calculations that is 400 times faster than a CPU-only alternative: The foundation of a recommendation is locating similar customers via their behavioral and purchasing patterns. However, if a company has 100 million customers, performing similarity calculations on just 200 features can take hours. FPGA allows these calculations to be done in parallel — and in under five milliseconds. Finding similar customers based on a set of attributes or features is a critical step for organizations — including healthcare, banks and retail — to connect people with the information and services they need. TigerGraph and Xilinx’s joint solution finds similar customers using the Cosine similarity algorithm, comparing around 200 features for each customer within a population of 10 million — all in under 100 milliseconds.
- A fraud detection system that locates fraud and money laundering rings — 20 times faster than a CPU-only alternative, according to Xilinx testing: Financial services organizations are always working to stay one step ahead of fraud — and finding fraud and money laundering rings is top of mind. Fraudsters and money launderers create multiple shell accounts with complex transactions to hide their activities. Xilinx and TigerGraph combine FPGA with graph database and analytics to accelerate the Louvain community detection algorithm; this, in turn, helps identify potential fraud and money laundering rings among large datasets.
The Xilinx and TigerGraph product or service recommendation engine and fraud detection solutions can be deployed today on AWS EC2 F1 instances, Azure NP VMs (preview), and on Alveo U50 Data Center Accelerator Cards.
TigerGraph’s work with partner Expero, which develops custom software exclusively for domain-expert users including scientists, healthcare professionals and government officials, further validates the company’s commitment to democratizing graph. Business users must be able to understand and work with applications via a graphical user interface; this is a key prerequisite for democratizing use of graph-based analytics. Expero has developed business GUI for two key use cases on top of TigerGraph Cloud. Expero’s business user apps, built on TigerGraph Cloud, include Fraud Alert Management and Customer 360/Customer Journey.
The TigerGraph-Expero combination unites graph and machine learning to boost fraud detection accuracy by 20 percent, while cutting costs by 10 percent. This drives organizations’ real-time decision making, allowing them to quickly identify and intervene to prevent fraud.
For more information on TigerGraph partnership sessions at Graph + AI World 2020: https://www.tigergraph.com/graphaiworld/
Helpful Links
- Graph + AI World
- TigerGraph Cloud
- TigerGraph Developer Community
- TigerGraph Website
- TigerGraph Blog
- TigerGraph on Twitter
- TigerGraph on LinkedIn
About TigerGraph TigerGraph is the only scalable graph database for the enterprise. TigerGraph’s proven technology connects data silos for deeper, wider and operational analytics at scale. Four out of the top five global banks use TigerGraph for real-time fraud detection. Over 50 million patients receive care path recommendations to assist them on their wellness journey. 300 million consumers receive personalized offers with recommendation engines powered by TigerGraph. The energy infrastructure for 1 billion people is optimized by TigerGraph for reducing power outages. TigerGraph’s proven technology supports applications such as fraud detection, customer 360, MDM, IoT, AI, and machine learning. The company is headquartered in Redwood City, California, USA. Follow TigerGraph on Twitter at @TigerGraphDB or start free with tigergraph.com/cloud or download TigerGraph Enterprise Free License.
###
Media Contact
Cathy Wright
Offleash PR for TigerGraph
650-678-1905
