Rotterdam Capital Days is back – Demystifying the world of capital in Rotterdam on September 18th–22nd!Check it out!
Rotterdam Capital Days is back!


Loading Events

« All Events

  • This event has passed.

Break Out: The need for Explainable AI in Healthcare – Landscape

May 16, 18:15 - 19:00

The Health sector is one of the most promising areas where Artificial Intelligence will continue to increase its impact.

An abundance of data, high workloads with a high administrative burden and many repetitive tasks in a focussed domain: these factors provide fertile grounds for AI and Machine Learning to alleviate some of this workload, both in the medical and operational domain. Though some expectations are overblown, AI is here to stay.

There are, however, some issues to overcome in a sector that requires high confidence and accountabilty. To be able to implement AI in healthcare, we will need to be able to trust that our AI models don’t make mistakes. It is very hard to trust a black box model we don’t understand, even when it might outperform us humans.

Landscape is developing AI solutions to lower the administrative burden in this sector. We will explore the ways we solve the issues of explainability and trust in a recent project of ours: automated classification of hospital admissions with ICD-10 codes.

All Dutch hospital admissions require labeling of the main and secondary diagnoses. This labeling process is expensive and understaffed in many hospitals. Together with DHD, the institution that is responsible for collecting this information, we have created an AI-based solution that automates part of this process and reduces workload for the remainder. Quality, explainability and trust in the model were important requirements in this project.

Landscape is a Data Science Agency that focusses on creating real business value through AI and ML. We create solutions, provide education through training and workshops, and help with creating a roadmap on increasing data impact.

Details

Date:
May 16
Time:
18:15 - 19:00
Event Tags:
, ,
Hidden