MLaaS: Machine Learning as a Service

Abstract

Abstract-The demand for knowledge extraction has been increasing. With the growing amount of data being generated by global data sources (e.g., social media and mobile apps) and the popularization of context-specific data (e.g., the Internet of Things), companies and researchers need to connect all these data and extract valuable information. Machine learning has been gaining much attention in data mining, leveraging the birth of new solutions. This paper proposes an architecture to create a flexible and scalable machine learning as a service. An open source solution was implemented and presented. As a case study, a forecast of electricity demand was generated using real-world sensor and weather data by running different algorithms at the same time.

Performance analysis of binary and multiclass models using azure machine learning

“…The objective of the proposed work is to analyse the performance of various algorithms and investigate their training time, prediction time, attack detection rate and false alarm rate by considering network instances of UNSW NB-15 dataset on a sophisticated Machine learning as a service (MLaaS) platform called Microsoft Azure Machine Learning Studio(MAMLS).A modern and a comprehensive dataset is essential to evaluate the effectiveness of the proposed approach and UNSW NB-15 dataset serves the purpose [17][18][19]. A significant advantage of any MLaaS offering is its ability to save computational resources that involve exceesive costs [20,21].The novelty of the proposed approach is that the false alarm rate generated by two class decision forest model is quite negligible and the attack detection capability of multiclass decision forest model is definitely desirable. It is worthwhile to mention that the results of classification tasks are quite superior than existing state of the art techniques.Some existing studies in the literature have explored the performance of different machine learning algorithms on UNSW NB-15 dataset as elucidated below.…”

Parallelizing Machine Learning as a service for the end-user

“…Similarly, Chan et al [8] describe the distributed architecture exploited in ML applications in Uber, with a focus on model training and features selection. A complete architecture for MLaaS is also described by Ribeiro et al [2] who present a specific analysis on three ML classifiers (multi-layer perceptrons, support vector machines, K-nearest neighbors). In each of these works, however, the parallelization of the final prediction stage is only superficially addressed, or ignored altogether, whereas we believe that a ML tool provided as-a-service to a wide public of end-users cannot disregard the parallelization of this last step (although it is generally less computationally expensive then the previous ones).…”

Challenges and opportunities in the integration of the Systems Engineering process and the AI/ML model lifecycle

“…The execution environment of a software system including AI/ML functions must consider the specific requirements regarding data and technology as well as the feedback loops (if necessary). In general, current approaches are considering an AI/ML system as a black-box, a type of service, Artificial "Intelligence as a Service" (AIaaS) and "Machine Learning as a Service" (MLaaS) (Ribeiro, Grolinger & Capretz 2015). However, it is necessary to highlight that the operation of the AI/ML system can also imply the need of exchanging data, information and knowledge, cooperation with other software modules, requiring then a standardized access to data and operations (e.g.…”