Probabilistic Matchmaking Methods for Automated Service Discovery

Abstract

Abstract-Automated service discovery enables human users or software agents to form queries and to search and discover the services based on different requirements. This enables implementation of high-level functionalities such as service recommendation, composition, and provisioning. The current service search and discovery on the Web is mainly supported by text and keyword based solutions which offer very limited semantic expressiveness to service developers and consumers. This paper presents a method using probabilistic machine-learning techniques to extract latent factors from semantically enriched service descriptions. The latent factors are used to construct a model to represent different types of service descriptions in a vector form. With this transformation, heterogeneous service descriptions can be represented, discovered, and compared on the same homogeneous plane. The proposed solution is scalable to large service datasets and provides an efficient mechanism that enables publishing and adding new services to the registry and representing them using latent factors after deployment of the system. We have evaluated our solution against logic-based and keyword-based service search and discovery solutions. The results show that the proposed method performs better than other solutions in terms of precision and normalised discounted cumulative gain values.

Large-Scale Indexing, Discovery, and Ranking for the Internet of Things (IoT)

“…For example, [Cassar et al 2014] utilise Ontology Web Language for Services (OWL-S) to describe sensor services semantically and apply LDA (a generative probabilistic unsupervised machine-learning technique) to map service descriptions into latent factors (topics). A ranking process is performed on latent factors based on their similarities with user queries.…”

A ranking method for sensor services based on estimation of service access cost

“…It mostly employs content-oriented ranking paradigms, e.g., by analysing the relevance of Web service descriptions with regard to queries [27]. To improve the discovery performance, machine learning based methods can be applied to perform deep semantic analysis on the service descriptions, for example, the work in [12] first applies Latent Dirichlet Allocation [39] to derive a latent factor model from service descriptions, and then computes service matchmaking using the latent model. The experiments show that the discovery results are more promising than the existing methods.…”

“…Matching queries with sensor service descriptions is only able to retrieve a list of relevant services (which are semantically ''equivalent'' to each other). In order to improve ranking, the work in [13] proposes a hybrid method which applies both probabilistic inference [12] and semantic reasoning on service descriptions. However, in reality, it is reasonable to assume that detailed descriptions on sensor services are always available for performing sophisticated analysis.…”

“…This is very different from ranking Web services and semantic Web services, which contain abundant features for effective algorithm design (e.g., textual descriptions and quality of service information) [27,12,43]. Characteristics of the sensor services are also different as the exposing sensors are mostly capability constrained, location dependent, heterogeneous and operating in highly dynamic physical environments.…”

SWEFT: Semantic Web Service Engine for Telecommunication – An Automatic Discovery and Composition through Genetic Approach

“…Cassar et. al., [3] suggested a method which uses probabilistic machine-learning techniques to abstract the underlying components from semantically enriched service annotations. The model is constructed using latent factors and helps to exhibit the different service annotations by a vector form.…”