Learning in Text Streams: Discovery and Disambiguation of Entity and Relation Instances.

We consider a scenario where an artificial agent is reading a stream of text composed of a set of narrations, and it is informed about the identity of some of the individuals that are mentioned in the text portion that is currently being read. The agent is expected to learn to follow the narrations, thus disambiguating mentions and discovering new individuals. We focus on the case in which individuals are entities and relations and propose an end-to-end trainable memory network that learns to discover and disambiguate them in an online manner, performing one-shot learning and dealing with a small number of sparse supervisions. Our system builds a not-given-in-advance knowledge base, and it improves its skills while reading the unsupervised text. The model deals with abrupt changes in the narration, considering their effects when resolving coreferences. We showcase the strong disambiguation and discovery skills of our model on a corpus of Wikipedia documents and on a newly introduced data set that we make publicly available.

A new integrated and interactive tool applicable to inborn errors of metabolism: Application to alkaptonuria.

This paper describes our experience with the development and implementation of a database for the rare disease Alkaptonuria (AKU, OMIM: 203500). AKU is an autosomal recessive disorder caused by a gene mutation leading to the accumulation of homogentisic acid (HGA). Analogously to other rare conditions, currently there are no approved biomarkers to monitor AKU progression or severity. Although some biomarkers are under evaluation, an extensive biomarker analysis has not been undertaken in AKU yet. In order to fill this gap, we gained access to AKU-related data that we carefully processed, documented and stored in a database, which we named ApreciseKUre. We undertook a suitable statistical analysis by associating every couple of potential biomarkers to highlight significant correlations. Our database is continuously updated allowing us to find novel unpredicted correlations between AKU biomarkers and to confirm system reliability. ApreciseKUre includes data on potential biomarkers, patients' quality of life and clinical outcomes facilitating their integration and possibly allowing a Precision Medicine approach in AKU. This framework may represent an online tool that can be turned into a best practice model for other rare diseases.