Publikation
Continuous Deployment of Machine Learning Pipelines
Behrouz Derakhshan; Alireza Rezaei Mahdiraji; Tilmann Rabl; Volker Markl
In: International Conference on Extending Database Technology. International Conference on Extending Database Technology (EDBT-2019), March 25-29, Lisbon, Portugal, ISBN 978-3-89318-081-3, OpenProceedings, 2019.
Zusammenfassung
Today machine learning is entering many business and scientic
applications. The life cycle of machine learning applications consists of data preprocessing for transforming the raw data into
features, training a model using the features, and deploying the
model for answering prediction queries. In order to guarantee
accurate predictions, one has to continuously monitor and update
the deployed model and pipeline. Current deployment platforms
update the model using online learning methods. When online
learning alone is not adequate to guarantee the prediction accuracy, some deployment platforms provide a mechanism for
automatic or manual retraining of the model. While the online
training is fast, the retraining of the model is time-consuming and
adds extra overhead and complexity to the process of deployment.
We propose a novel continuous deployment approach for updating the deployed model using a combination of the incoming realtime data and the historical data.We utilize sampling techniques to
include the historical data in the training process, thus eliminating
the need for retraining the deployed model. We also o er online
statistics computation and dynamic materialization of the preprocessed features, which further reduces the total training and data
preprocessing time. In our experiments, we design and deploy
two pipelines and models to process two real-world datasets. The
experiments show that continuous deployment reduces the total
training cost up to 15 times while providing the same level of quality when compared to the state-of-the-art deployment approaches