Today Kinetica unveiled a natural language interface for its high-speed analytics database that has been powered by GPUs, making complex SQL queries easier to understand for those less technically minded. The interface, called ChatGPT, was made possible through the OpenAI large language model.
Kinetica was originally created for the U.S. Army to rapidly inspect large bits of geo and temporal data for potential terrorist activity. At the time, it relied on the processing capacity of GPUs, with other databases having to heavily limit data using indexes and other functions. With the introduction of Conversational Query, Kinetica’s power is now available to those who may not possess the same technical abilities usually required to understand traditional SQL queries.
ChatGPT will convert full phrases in plain language into SQL, with the OpenAI process first converting the statement before handing it the finalized command over to Kinetica. This move allows non-technical employees of organizations to ask ad-hoc questions, which demand more of a wide scan of records rather than pre-defined queries.
Kinetica co-founder and CEO Nima Negahban suggested that the surge in ad-hoc queries is increasingly sought after. To further this goal and in order to get good results from ChatGPT, Kinetica has implemented a hydration process to earlier process data sets. This allows the language model to access native analytical functions, joining with time-series and spatial results to get the best out of the query.
Though OpenAI’s model is far from perfect, Negahban has assured that what it does understand is enough for ad-hoc queries. The immense power of the database also comes makes it applicable to these types of queries, with other computer databases lacking such ability when presented with more than one join.
Kinetica recently introduced Conversational Query, a feature available on both their cloud and on-premises versions. This new capability was announced alongside the appointment of Robert DeMartino as Kinetica’s Chief Revenue Officer (CRO). With over 25 years of experience, Robert has a wealth of knowledge in developing various aspects of customer-focused organizations, including sales for major accounts, enterprises, and mid-market segments, as well as channels, customer success, and services. Throughout his career, Robert has worked with successful early-stage startups and has successfully built sales teams from scratch, growing them to over a thousand members. He has served enterprises worldwide and has led sales teams in the Americas, Europe, and the Asia Pacific region during his time at Sun Microsystems. In his most recent role, Robert served as CRO at TigerGraph, and prior to that, he was part of the executive team responsible for introducing innovative data and analytics technologies to the market with Greenplum/Pivotal. These ventures were later acquired by Dell/EMC and Workday, respectively.
OpenAi’s ChatGPT large language model has proven itself to be quite adept with natural language and computer understanding. This technology breakthrough has made Kinetica’s high speed analytics database accessible for non-technical users to pose questions written plain English. Primarily used in the past for identifying terrorist activity, the use of ChatGPT will make it easier for users from all fields to understand complex SQL queries and find the answers they need from Kinetica’s database quickly and easily.
