Hernán C. Vázquez

Hernan C. Vazquez, J. Andres Diaz Pace, Claudia Marcos et al.

The selection of software technologies is an important but complex task. We consider developers of JavaScript (JS) applications, for whom the assessment of JS libraries has become difficult and time-consuming due to the growing number of technology options available. A common strategy is to browse software repositories via search engines (e.g., NPM, or Google), although it brings some problems. First, given a technology need, the engines might return a long list of results, which often causes information overload issues. Second, the results should be ranked according to criteria of interest for the developer. However, deciding how to weight these criteria to make a decision is not straightforward. In this work, we propose a two-phase approach for assisting developers to retrieve and rank JS technologies in a semi-automated fashion. The first-phase (ST-Retrieval) uses a meta-search technique for collecting JS technologies that meet the developer's needs. The second-phase (called ST-Rank), relies on a machine learning technique to infer, based on criteria used by other projects in the Web, a ranking of the output of ST-Retrieval. We evaluated our approach with NPM and obtained satisfactory results in terms of the accuracy of the technologies retrieved and the order in which they were ranked.

Hernán Ceferino Vázquez, Jorge Sanchez, Rafael Carrascosa

Automated Machine Learning (AutoML) has become increasingly popular in recent years due to its ability to reduce the amount of time and expertise required to design and develop machine learning systems. This is very important for the practice of machine learning, as it allows building strong baselines quickly, improving the efficiency of the data scientists, and reducing the time to production. However, despite the advantages of AutoML, it faces several challenges, such as defining the solutions space and exploring it efficiently. Recently, some approaches have been shown to be able to do it using tree-based search algorithms and context-free grammars. In particular, GramML presents a model-free reinforcement learning approach that leverages pipeline configuration grammars and operates using Monte Carlo tree search. However, one of the limitations of GramML is that it uses default hyperparameters, limiting the search problem to finding optimal pipeline structures for the available data preprocessors and models. In this work, we propose an extension to GramML that supports larger search spaces including hyperparameter search. We evaluated the approach using an OpenML benchmark and found significant improvements compared to other state-of-the-art techniques.