TensorDB: : Database Infrastructure for Continuous Machine Learning

F. Liu, A. Oehmichen, Jingqing Zhang, K. Sun, H. Dong, Y Mo, Yike Guo

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper introduces the TensorDB system, a frameworkthatfusesdatabaseinfrastructureandapplication software to streamline the development, training, evaluation and analysing machine learning models. The design principle is to track the whole model building process with database and connected different components by database query mechanism. This design produces a highly flexible framework enable each component to be updated independently. The theoretical value is that it enables continuous machine learning. TensorDB is motivated by production application of machine learning model as consolidation of many engineering practice, and is could be served as the foundation for high level tools for machine learning application.
Original languageAmerican English
Title of host publicationInternational Conference for Artificial Intelligence
StatePublished - 2017

Fingerprint Dive into the research topics of 'TensorDB: : Database Infrastructure for Continuous Machine Learning'. Together they form a unique fingerprint.

  • Cite this

    Liu, F., Oehmichen, A., Zhang, J., Sun, K., Dong, H., Mo, Y., & Guo, Y. (2017). TensorDB: : Database Infrastructure for Continuous Machine Learning. In International Conference for Artificial Intelligence https://csce.ucmss.com/cr/books/2017/LFS/CSREA2017/ICA3261.pdf