On the path to productivity

AQO 2.0which was released a year ago. It has improved performance and usability due to the optimization and simplification of many complex elements of the system (the corresponding changes can be viewed in presentations).

Conclusion

The result is simple – ML in general and neural networks in particular can successfully cope with the challenges that real businesses face. Many of the approaches considered are already being developed both in open-source and commercial DBMSs, which is good news. It remains to analyze several models that are unique in nature, which take another important step in the development of neural network query optimizers – we will talk about this in a new article.