We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 - 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!


TigerGraph, maker of a graph analytics platform for data scientists, during its Graph & AI Summit event today introduced its TigerGraph ML (Machine Learning) Workbench, a new-gen toolkit that ostensibly will enable analysts to improve ML model accuracy significantly and shorten development cycles. 

Workbench does this while using familiar tools, workflows, and libraries in a single environment that plugs directly into existing data pipelines and ML infrastructure, TigerGraph VP Victor Lee told VentureBeat. 

The ML Workbench is a Jupyter-based Python development framework that enables data scientists to build deep-learning AI models using connected data directly from the business. Graph-enabled ML has proven to have more accurate predictive power and take far less run time than the conventional ML approach. 

Conventional machine learning algorithms are based on the learning of systems by training sets to develop a trained model. This pre-trained model is used to classify or recognize the test dataset; this typically can take days or weeks to finalize for a particular use case. Graph-based ML sometimes can take minutes to build an algorithmic model.

Event

Transform 2022

Join us at the leading event on applied AI for enterprise business and technology decision makers in-person July 19 and virtually from July 20-28.

Register Here

Value of ML high, but so is the learning curve

“Graph is proven to accelerate and improve ML learning and performance, but the learning curve to use the APIs (application programming interfaces) and libraries to make that happen has proven very steep for many data scientists,” Lee said in a media advisory. “So we created ML Workbench to provide a new functional layer between the data scientists and the graph machine-learning APIs and libraries to facilitate data storage and management, data preparation, and ML training. 

“In fact, we have seen early adopters gaining a 10-50% increase in the accuracy of their ML models as a result of using ML Workbench and TigerGraph,” he said.

TigerGraph’s whole way of thinking is around the definition of human identity, which is based on how you interact with others, Lee told VentureBeat. 

“The same thing holds true with graphs in data modeling, and this is just now extending to neural networks.” Lee said. “Every node in a graph is interrelated, like people. Graphs are great for querying pattern-matching algorithms. Workbench will help you deploy machine learning based on the information inside the graph, but the real power comes with graph neural networks, which are regular graphs on steroids. 

“In our DGL (deep graph library), for example, there’s an extension of (Meta’s) Pytorch geometric that supports graph neural networks,” he said. “This is a great feature, and it shows we’re going to where the data scientists are; we’re not trying to make them learn something new. We’re using the tools that they already know and are comfortable with, because we’re trying to cut down the learning curve.”

Optimal for fraud, prediction use cases

The ML Workbench enables organizations to determine improved insights in node-prediction applications, such as fraud, and edge-prediction applications, which include product recommendations, Lee said. The ML Workbench enables AI/ML practitioners to explore graph-enhanced machine learning and graph neural networks (GNNs) because it is fully integrated with TigerGraph’s database for parallelized graph data processing/manipulation, Lee said. 

The ML Workbench is designed to interoperate with popular deep learning frameworks such as PyTorch, PyTorch Geometric, DGL, and TensorFlow, providing users with the flexibility to choose a framework with which they are most familiar. The ML Workbench is also plug-and-play ready for Amazon SageMaker, Microsoft Azure ML, and Google Vertex AI, Lee said.

The ML Workbench is designed to work with enterprise-level data. Users can train GNNs – even on very large graphs – due to the following built-in capabilities:

  • TigerGraph DB’s distributed storage and massively parallel processing;
  • Graph-based partitioning to generate training/validation/test graph data sets;
  • Graph-based batching for GNN mini-batch training to improve performance and to reduce HW requirements; and
  • Subgraph sampling to support leading edge GNN modeling techniques.

ML Workbench is compatible with TigerGraph 3.2 onward, available as a fully managed cloud service and for on-premises use. Currently available as a preview, ML Workbench will be generally available in June 2022, Lee said.

TigerGaph competes with Neo4J, ArangoDB, MemGraph and a few others in the graph database space.

‘Million Dollar Challenge’ winners selected

At the Graph & AI Summit, TigerGraph unveiled the winners of the Graph for All Million Dollar Challenge — awarding $1 million in cash to game-changing, graph-powered projects that analyze and address many of today’s biggest global social, economic, health, and climate-related concerns.

The winning projects, announced at this week’s Graph + AI Summit, were hand-selected by the global judging committee from more than 1,500 registrations from 100-plus countries. Mental Health Hero claimed the $250,000 Grand Prize for creating an application to help provide greater access and personalization to mental health treatment.

Author
Topics