Artificial intelligence trained to analyze causation

phys.org | 6/6/2018 | Staff
TwiztedGurl (Posted by) Level 3
Click For Photo: https://3c1703fe8d.site.internapcdn.net/newman/gfx/news/hires/2018/15-artificialin.jpg










The causes of real-world problems in economics and public health can be notoriously hard to determine. Often, multiple causes are suspected, but large datasets with time-sequenced data are not available. Previous models could not reliably analyze these challenges. Now, researchers have tested the first artificial intelligence model to identify and rank many causes in real-world problems without time-sequenced data, using a multi-nodal causal structure and Directed Acyclic Graphs.

When something bad happens, it is natural to try figure out why it happened. What caused it? If the cause is determined, it may be possible to avoid the same outcome the next time. However, some of the ways in which humans try to understand events, such as resorting to superstition, cannot explain what is actually going on. Neither does correlation, which can only say that event B happened around the same time as event A.

Event - Causality—how - Information - Event - Information

To really know what caused an event, we need to look at causality—how information flows from one event to another. It is the information flow that shows there is a causal link—that event A caused event B. But what happens when the time-sequenced information flow from event A to event B is missing? General causality is required to identify the causes.

Mathematical models for general causality have been very limited, working for up to two causes. Now, in an artificial intelligence breakthrough, researchers have developed the first robust model for general causality that identifies multiple causal connections without time-sequence data, the Multivariate Additive Noise Model (MANM).

Researchers - University - Johannesburg - South - Africa

Researchers from the University of Johannesburg, South Africa, and National Institute of Technology Rourkela, India, developed the model and tested it on simulated, real-world datasets. The research is published in the journal Neural Networks.

"Uniquely, the model can identify multiple, hierarchical causal factors. It works even if data with time sequencing is not available. The...
(Excerpt) Read more at: phys.org
Wake Up To Breaking News!
An enemy of death and destruction
Sign In or Register to comment.

Welcome to Long Room!

Where The World Finds Its News!