Hi Rong, I am VP of product management @ MariaDB. I have a few follow-up questions for you:
1. What types of AI models have you seen work well in estimating cardinality and ndv and how have you seen them work in production? I think the what-if scenarios setup in the correct way could help AI models align to relatively well optimized outputs since data can be created to assist in AI training. 2. What are the limitations of VIDEX simulations in representing real-world database workloads that you have seen? What have you seen work relating to ensuring that AI models trained on VIDEX data generalize well to production environments? 3. Has ByteDance explored using LLMs or other natural language processing techniques in conjunction with VIDEX for database optimization tasks? If so, what were the outcomes, and are there recommendations for incorporating such models into an index automation system? Thanks Adam _______________________________________________ discuss mailing list -- discuss@lists.mariadb.org To unsubscribe send an email to discuss-le...@lists.mariadb.org
