Discovery Science 

Abstract: Detecting concept drift is important for dealing with realworld online learning problems. To detect concept drift in a small number of examples, methods that have an online classifier and monitor its prediction errors during the learning have been developed. We have developed such a detection method that uses a statistical test of equal proportions. Experimental results showed that our method performed well in detecting the concept drift in five synthetic datasets that contained various types of concept drift.

Abstract: The FAIR Guiding Principles, published in 2016, aim to improve the findability, accessibility, interoperability and reusability of digital research objects for both humans and machines. The FAIR principles are also directly relevant to research software. In this position paper “Towards FAIR principles for research software”, we summarised and developed a basis for community discussion. At the start, we discussed what makes software different from data concerning the application of the FAIR principles, and which desired characteristics of research software go beyond FAIR. Then, we presented an analysis of where the existing principles can directly apply to software, and where they need to be adapted or reinterpreted. Our next step after the position paper is to prompt for community-agreed identifiers for FAIR research software.Acknowledgments To all the authors of Towards FAIR principles for research software https://doi.org/10.3233/DS-190026, and the numerous people who contributed to the discussions around FAIR research software at different occasions preceding the work on this paper. References Lamprecht, Anna-Lena, et al. (2019) Towards FAIR principles for research software. Data Science. https://doi.org/10.3233/DS-190026 ABOUT THE AUTHOR(S) Dr Paula Andrea Martinez is leading the National Training Program for the Characterisation Community in Australia since 2019. She works for the National Image Facility (NIF). Last year she worked at ELIXIR Europe coordinating the Bioinformatics and Data Science training program in Belgium and collaborated with multiple ELIXIR nodes in the development of Software best practices. Her career, spanning Sweden, Australia and Belgium nurtured her experience in Bioinformatics and Research Software development for complex and dataintensive science. She started a career in Computer Science, later on, interested in research methods development and now outreach and advocacy in data and software best practices

Abstract: In a multidatabase system, schematic conflicts between two objects are usually of interest only when the objects have some semantic affinity. In this paper we try to reconcile the two perspectives. We first define the concept of semantic proximity and provide a semantic taxonomy. We then enumerate and classify the schematic and data conflicts. We discuss possible semantic similarities between two objects that have various types of schematic and data conflicts. Issues of uncertain information and inconsistent information are also addressed.

Abstract: Neural network classifiers are known to be able to learn very accurate models. In the recent past, researchers have even been able to train neural networks with multiple hidden layers (deep neural networks) more effectively and efficiently. However, the major downside of neural networks is that it is not trivial to understand the way how they derive their classification decisions. To solve this problem, there has been research on extracting better understandable rules from neural networks. However, most authors focus on nets with only one single hidden layer. The present paper introduces a new decompositional algorithm – DeepRED – that is able to extract rules from deep neural networks.