Datenschutzerklärung|Data Privacy
Impressum

07.12.2018
K. Forster

The Paper "Continuous Deployment of Machine Learning Pipelines" was Accepted for Publication at EDBT 2019

Continuous Deployment of Machine Learning Pipelines, Behrouz Derakhshan, Alireza Rezaei Mahdiraji,
Tilmann Rabl, Volker Markl
, 2018. To be presented on 22nd International Conference on Extending Database Technology (EDBT), Lisbon, Portugal, 2019.

Abstract
Today machine learning is entering many business and scientific 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 themodel 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 offer 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.

The paper is available here.


I f you want to learn more about EDBT/ICDT 2019 Joint Conference visit: http://edbticdt2019.inesc-id.pt/