SimpleMind for more human-like cognitive AI

What is it about?

This paper introduces a new open source software environment for Cognitive AI that improves the reliability and trustworthiness of deep learning by adding human-like knowledge and reasoning. Deep learning is data driven and cannot understand or reason with its outputs, so despite strong overall performance metrics it is vulnerable to common sense errors. Cognitive AI refers to an emerging class of artificial intelligence systems that can mimic and replicate human thought processes, enabling them to perceive, reason, learn, and make decisions in a manner similar to human cognition. SimpleMind uses Cognitive AI to guide deep neural networks and check and correct their outputs when needed. It enables users to integrate deep learning with a knowledge graph and logical reasoning algorithms to make AI decisions that are reliable, interpretable, and explainable. The paper provides example SimpleMind applications that demonstrate improvements in trustworthiness, encompassing reliability, interpretability and explainability of decisions.

Featured Image

Photo by Nick Fewings on Unsplash

Photo by Nick Fewings on Unsplash

Why is it important?

Currently, AI in medicine is dominated by powerful, but difficult-to-interpret black-box deep neural network approaches. These often perform well in preliminary studies but are found to lack robustness in real-world performance and explainability, resulting in a lack of trust and adoption by clinicians for critical tasks. It also means that when they are used, deep learning systems require a human in the loop to check results, which limits their effectiveness in reducing workloads. By adding human-like knowledge and reasoning to check and correct deep learning outputs, SimpleMind can increase trustworthiness and thereby adoption and automation in medical practice. This could enable AI to reach its full potential in improving decision making and reducing time and cost in healthcare.

Perspectives