News Story

Alba Morales’ Paper Presented at K-CAP 2021

Kiran Parmar, Monday 29 Nov 2021

KMi PhD student Alba Morales attended the 11th KCap Online Conference which took place on 2nd to 3rd December. Alba presented a paper entitled: "Reasoning on Health Condition Evolution for Enhanced Detection of Vulnerable People in Emergency Settings" which she co-authored with Dr. Enrico Daga and Prof. Enrico Motta.

The International Conference on Knowledge Capture (K-CAP) is one of the reference venues for researchers in Knowledge Engineering and its applications in areas such as Artificial Intelligence, including knowledge representation, knowledge acquisition and machine learning, as well as research in knowledge extraction, reuse and integration of data in application areas such as healthcare and cyber-infrastructures.

KMi Research Fellows, Francesco Osborne and Enrico Daga also attended the conference, as they had served on the KCap Programme Committee.

Alba is conducting her PhD research to improve how data can be used to support Smart Cities services. Specifically, her research aims to use healthcare data and provide first responders with helpful information about vulnerable people during an emergency event. In this paper, she explores the use of health records and proposes a solution for representing and reasoning on medical events. Her objective is to design a representation of health evolution that facilitates Intelligent Systems reason on medical events to devise an accurate picture of people’s health status.

The paper introduced a methodology for identifying relevant information from health records in supporting emergency services. This methodology introduces the concept of Condition Evolution Statement (CES) representation, a sophisticated and elaborated description of the recovery process, which supports the definition of the convalescence period according to the type of health issue, and therefore allowing the detection of ongoing medical events at a given point in time. Their approach integrates semantic technologies and machine learning to represent and reason on health condition evolution. The paper evaluates the precision of an Intelligent System that uses health records to identify vulnerable people during a fire emergency. They incorporate the CES representation in the Intelligent Systems, and the results of the experiments reported clear evidence of an increase in precision in all cases. Their contributions include a database of condition evolution annotations and a novel representation and reasoning of condition evolution for supporting the detection of vulnerable people in emergency response.