During the Graph & AI Summit event, TigerGraph, a new-gen toolkit that enables analysts to improve the model accuracy significantly and shorten development cycles.
The ML Workbench is a Java-based Python development framework that allows data scientists to develop deep-learning AI models using connected data from the business. The graph-enabled ML has proven to have a greater predictive power and an effect that is far less time compared to the conventional ML approach.
In a similar way, graph-based ML may take a few minutes to develop an algorithmic model.
The value of ML is high, but there is also the learning curve.
According to Lee, the learning curve for graph applications to improve ML learning and performance has been very severe for many data scientists. So we created a ML Workbench that allows the data scientists to create a new functional layer between the graph machine learning APIs and libraries, allowing for data storage and management, and data preparation.
According to the study, early adopters have increased the accuracy of their ML models by ten to 50 percent as a result of using ML Workbench and TigerGraph.
The same thing is true with graphs in data modeling, and this is just now expanding to neural networks. Lee said. Graphs are great for querying pattern-matching algorithms. However, graph neural networks are the norm.
According to the DGL (deep graph library), there is a (Metas) Pytorch geometric that supports graph neural networks. It shows that the data scientists were going to where they are, and they weren''t trying to make them learn something new. Were using the tools they already know and understand, because we are trying to reduce learning curve.
Optimal for fraud, prediction use cases
The ML Workbench is designed to enable organizations to discern enhanced insights in node-prediction applications, such as fraud, and edge-prediction applications, which include product recommendations. Lee said the ML Workbench is well-integrated with the TigerGraphs database for parallel graph data processing and manipulation.
Lee said that the ML Workbench is designed to interoperate with popular deep learning tools such as PyTorch, PyTorch Geometric, DGL, and TensorFlow, thus giving users the flexibility to select a framework with which they are most familiar. The ML Workbench is also available for Amazon SageMaker, Microsoft Azure ML, and Google Vertex AI.
Due to the following built-in capabilities, the ML Workbench is designed to work with enterprise data. GNNs can be even taught on very large graphs.
- TigerGraph DBs 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 now compatible with TigerGraph 3.2 onward, which is now a fully managed cloud service and for on-premises use. Lee said the company will generally be available in June 2022.
In the graph database space, TigerGaph vs Neo4J, ArangoDB, MemGraph, and other companies.
Million Dollar Challenge winners selected
TigerGraph presented the winners of the Graph for the All Million Dollar Challenge, which awarded $1 million in cash to game-changing, graph-powered projects that will analyze and address today''s biggest social, economic, health, and climate-related concerns.
The awards, which were announced at this weeks Graph + AI Summit, were hand-selected by the global judging committee, which included more than 1,500 registrations from 100-plus countries. Mental Health Hero received the $250,000 Grand Prize for a program to enable greater access and personalization to mental health treatment.