SiliconANGLE reports on the introduction of a host of generative AI features to the Oracle HeatWave database

0
52
SiliconANGLE reports on the introduction of a host of generative AI features to the Oracle HeatWave database



Oracle Corp. is enhancing its managed MySQL database service Heatwave with generative artificial intelligence capabilities, dubbed HeatWave GenAI. These new features include large language models in the database, automated storage of vectors, scalable vector processing, natural language queries, and chat application creation tools. HeatWave GenAI is now available in all Oracle Cloud Infrastructure regions and other cloud platforms at no extra cost to HeatWave customers.

Developers can now create a vector store for unstructured content with a single SQL command using built-in embedding models. This vector warehouse is optimized for storing high-dimensional vector data types used in recommendation systems and similarity searches. Oracle emphasizes that data never leaves the database, and training and inference tasks can be done without the need for graphical processing units.

The database LLMs enable users to search, generate, and summarize content, perform augmented recovery generation, and combine generative AI with other HeatWave capabilities like machine learning. Oracle has been steadily expanding HeatWave’s features since its introduction in 2020, with recent updates including advanced automation and increased machine learning support and lake house features in 2022. HeatWave is now available on Amazon Web Services Inc. and Microsoft Corp. Azure.

Oracle has reshaped HeatWave as its flagship autonomous database and positioned it as a cost-effective solution for companies with less complex processing needs. Despite its focus on affordability, Oracle highlights the ease of use of HeatWave’s new features, such as creating a vector warehouse database with simple automation.

The database’s smaller LLMs are comparable in accuracy and performance to larger models and can be used in conjunction with other functions like machine learning. Quantization techniques help reduce model size and computational demands while maintaining performance. Oracle claims HeatWave’s ability to process similarity searches faster than competitors like Snowflake Inc. and Databricks Inc., offering faster results at a lower cost.

Oracle’s continuous innovation with HeatWave demonstrates its commitment to providing advanced database solutions to customers. With HeatWave GenAI, users can leverage generative AI capabilities within the database for enhanced data processing and analysis. Oracle’s strategic approach to integrating AI technologies into HeatWave aims to deliver high performance and cost-effective solutions for businesses of all sizes.

Article Source
https://siliconangle.com/2024/06/26/oracle-heatwave-gets-host-generative-ai-features/