Artificial intelligence (AI) can evaluate more data faster than anyone else. With such a huge pool of information, AI should be able to consider past data, process all the impacts, and make reliable predictions better than humans -- right? That may not always be the case, according to a multi-institutional team of researchers who have studied the synergies between how humans and ARTIFICIAL intelligence make predictions.
They published their results in the Aug. 23 issue of the Journal of Social Computing.
"The predictive task is ubiquitous - in any field or aspect of life, any decision involves predicting the consequences of alternative choices before making them," said Scott E. Page, author of the paper and a professor at the University of Michigan's Ross School of Business. "Understanding the combination of danger and promise, and striking the right balance between the two, is the main focus going forward."
According to Page, this concern stems from the recent shift from predictions based on experience, some data and intuition to programming considerations based on data and ai systems.
"Increasing accuracy by applying ever more powerful algorithms to ever larger databases begs the question: should humans remain in the field of prediction, or should predictions be left entirely to algorithms?" Ask the page.
The answer, researchers found, was a resounding no. The way humans predict is much more subtle than artificial intelligence methods, which have a crucial impact on accurate predictions.
Page said AI can handle big data well, while humans can better analyze what researchers call "thick" data. Unlike big data, which consists of many data points of the same type of data, fewer data points in thick data can tell a richer story. For example, years of statistics could allow ai to predict how many home runs a baseball player is likely to hit, but humans are more likely to understand that a popular team player is likely to have a longer career.
"Big data and thick data working together will produce more accurate collective predictions," Page said. "Thick data captures and draws attention to groups of factors that might slip through the cracks between separate big data variables. While big data casts a wider net, there are holes in that net."
Researchers tested this idea by measuring human and AI inputs through mathematical tests that might produce different predictions. They found that in typical cases, where future outcomes depend on past outcomes, AI could make accurate predictions without human input. However, in atypical cases where there are more unknown or unexpected factors, humans help ai reduce potential errors.
"As long as humans can continue to identify different attributes, that is, continue to build thicker data, or better understand atypical cases, they will continue to improve accuracy," Page said. "The hybrid predictor of the future will be a complex exploration of symbiosis rather than a competition between humans and computers."
The researchers plan to continue exploring how ai and human cooperative systems can help improve predictions, including how multiple systems work together to potentially give more accurate results.
"The details are unclear, but we can almost certainly predict that the roles and contributions of participants will adapt to growing data and greater computing power," Page said. "Cognitive work now and in the future will definitely involve humans, algorithms, data sets, subjects, objects and domains. As they seek to understand the work, these mixed groups will also shape it."
Other contributors include first author Lu Hong, Department of Finance, Loyola University; And PJ Lamberson of the UCLA Department of Communications.
Manly Battery which located in Shenzhen,China. A leading Lithium battery manufacturer over 12
years ,widly used for Robostic industry,if there is any project need to evluate ,pls feel free to send
email to info@manlybatteries.com
