Cognitive Computing for the delivery of healthcare

About the talk

The delivery of care relies on information coming from systems of record (e.g. EMRs), information collected by professionals (e.g. assessments) and information collected by IoT devices, as well as a vast corpus of knowledge from the medical and social care domains.

How do we connect this information and use it to improve the health of citizens and the work of care professionals, while providing for a more efficient health system? The answer lies in combining techniques from a variety of fields, including the Semantic Web, NLP and Artificial Intelligence. There is a variety of specific areas where these technologies help, spanning core healthcare, care management and social care.

Spyros will present a use-case in the Child Welfare space. Child welfare intake workers around the world answer millions of calls regarding children and youth who may be abused or neglected. Their job is to determine whether the child is safe. These calls may come from families, law enforcement officials, or other mandated reporters such as teachers or medical professionals. IBM is working on a solution which applies artificial intelligence and machine learning to assist in prioritizing high risk intake calls, improve intake worker’s workflow and streamline information gathering and documentation, all with a view of enabling better decisions for children and youth in need.

During the intake, the system captures initial information to identify the urgency of the call. Once a connection between a reporter and an intake worker is established, it “listens in” on the conversation, helping with documentation and providing valuable insights on-the-fly.

Using speech-to-text and natural language understanding, the artificial intelligence system identifies and surfaces relevant case information providing a complete view of the child and family if the data is available. Using machine learning, knowledge graphs and dialog systems, it analyses the case and the stage in the intake process to make suggestions to the intake worker on critical path questions and important topics to discuss with the reporter to scale the expertise of the intake worker and providing continuous training.

In the background, the system is collecting information that will assist the intake worker in completing their required intake assessments transforming how intake workers capture information. This gives care workers the time to focus on what they do best – helping children and keeping them safe.

About the speaker


Spyros Kotoulas is a Research Scientist at IBM Research Ireland and the Manager of the Health and Person-Centered Systems research group. He is working on improving Healthcare and Social Care for vulnerable populations, aiming a better experience for the individual, increased effectiveness and reduced cost. His aspiration is to build personalised and person-centric systems that will positively disrupt the management and delivery of care.


His research interests lie in artificial intelligence and data management for semi-structured data, parallel methods for data intensive processing, knowledge representation and reasoning, flexible data integration methods and the Internet of Things.


He is the IBM PI for the ProACT project.

He has been part of the teams that won the third prize at the Billion Triple Challenge in 2008, the first prize at the IEEE Scale Chalenge in 2010 and the second prize at the Semantic Web Challenge in 2012.

Spyros was previously at the Knowledge Representation and Reasoning Group at the Vrije Universiteit Amsterdam, where he has also completed his PhD in Scalable Peer-to-Peer Data Discovery and his MSc in Parallel and Distributed Computer Systems.