Abstract
In the past decade, we have witnessed the widespread adoption of Deep Neural Networks (DNNs) in several Machine Learning tasks. However, in many critical domains, such as healthcare, finance, or law enforcement, transparency is crucial. In particular, the lack of ability to conform with prior knowledge greatly affects the trustworthiness of predictive models. This paper contributes to the trustworthiness of DNNs by promoting monotonicity. We develop a multi-layer learning architecture that handles a subset of features in a dataset that, according to prior knowledge, have a monotonic relation with the response variable. We use two alternative approaches: (i) imposing constraints on the model's parameters, and (ii) applying an additional component to the loss function that penalises non-monotonic gradients. Our method is evaluated on classification and regression tasks using two datasets. Our model is able to conform to known monotonic relations, improving trustworthiness in decision making, while simultaneously maintaining small and controllable degradation in predictive ability.

Abstract
This paper presents an effort to timely handle 400+ GBytes of sensor data in order to produce Predictive Maintenance (PdM) models. We follow a data-driven methodology, using state-of-the-art python libraries, such as Dask and Modin, which can handle big data. We use Dynamic Time Warping for sensors behavior description, an anomaly detection method (Matrix Profile) and forecasting methods (AutoRegressive Integrated Moving Average - ARIMA, Holt-Winters and Long Short-Term Memory - LSTM). The data was collected by various sensors in an industrial context and is composed by attributes that define their activity characterizing the environment where they are inserted, e.g. optical, temperature, pollution and working hours. We successfully managed to highlight aspects of all sensors behaviors, and produce forecast models for distinct series of sensors, despite the data dimension.

Abstract

Abstract

Abstract
Narrative Extraction from text is a complex task that starts by identifying a set of narrative elements (actors, events, times), and the semantic links between them (temporal, referential, semantic roles). The outcome is a structure or set of structures which can then be represented graphically, thus opening room for further and alternative exploration of the plot. Such visualization can also be useful during the on-going annotation process. Manual annotation of narratives can be a complex effort and the possibility offered by the Brat annotation tool of annotating directly on the text does not seem sufficiently helpful. In this paper, we propose Brat2Viz, a tool and a pipeline that displays visualization of narrative information annotated in Brat. Brat2Viz reads the annotation file of Brat, produces an intermediate representation in the declarative language DRS (Discourse Representation Structure), and from this obtains the visualization. Currently, we make available two visualization schemes: MSC (Message Sequence Chart) and Knowledge Graphs. The modularity of the pipeline enables the future extension to new annotation sources, different annotation schemes, and alternative visualizations or representations. We illustrate the pipeline using examples from an European Portuguese news corpus.

Abstract
ECIR 2020 1 was one of the many conferences affected by the COVID-19 pandemic. The Conference Chairs decided to keep the initially planned dates (April 14-17, 2020) and move to a fully online event. In this report, we describe the experience of organising the ECIR 2020 Workshops in this scenario from two perspectives: the workshop organisers and the workshop participants. We provide a report on the organisational aspect of these events and the consequences for participants. Covering the scientific dimension of each workshop is outside the scope of this article. A detailed account of the main conference can be found in the Spring 2020 edition of BCS Informer [BCS, 2020].