Highlight:

  • The hybrid human-machine framework will develop accurate and better AI devices.
  • The amalgamation of computer science and computer science will open new AI opportunities.

Artificial Intelligence (AI) plays a crucial role in multiple aspects of human life, from a chatbot that replies to tax queries to algorithms that diagnose medical conditions and drive autonomous vehicles. Researchers at the University of California, Irvine, suggest that developing intelligent and more accurate systems needs a hybrid human-machine approach. In a study published this month in ‘Proceedings of the National Academy of Sciences,’ the researchers have presented their latest mathematical model that can enhance performance by combining human and algorithmic predictions and confidence scores.

The researchers evaluated their framework by conducting an image classification experiment. Here, human volunteers and computer algorithms worked differently to accurately spot distorted animal and daily objects pictures like chairs, bottles, trucks, and others.

The human volunteers ranked their confidence in spotting the accuracy of each image as low, medium, and high. On the other hand, machine classifiers scored a consistent score. The results suggested a tremendous difference in confidence between human volunteers and AI algorithms across images.

“In some cases, human participants were quite confident that a particular picture contained a chair, for example, while the AI algorithm was confused about the image. Similarly, for other images, the AI algorithm was able to confidently provide a label for the object shown, while human participants were unsure if the distorted picture contained any recognizable object,” said co-author Padhraic Smyth, UCI Chancellor’s Professor of computer science.

The researchers combined the predictions and confidence scores with their new Bayesian framework. They found that the hybrid model enhanced performance that either human or machine predictions could achieve alone.

Irvine’s initiative in AI, law, and society facilitated this interdisciplinary project. The research team said that the amalgamation of cognitive sciences, which revolve around human thought processes and behavior with computer science, wherein new technologies are created, will result in better human-machine collaboration to develop accurate artificially intelligent systems.

 Experts’ view:

“Humans and machine algorithms have complementary strengths and weaknesses. Each uses different sources of information and strategies to make predictions and decisions. We show through empirical demonstrations as well as theoretical analyses that humans can improve the predictions of AI even when human accuracy is somewhat below [that of] the AI—and vice versa. And this accuracy is higher than combining predictions from two individuals or two AI algorithms,” said co-author Mark Steyvers, UCI professor of cognitive sciences.

“While past research has demonstrated the benefits of combining machine predictions or combining human predictions—the so-called ‘wisdom of the crowds’ – this work forges a new direction in demonstrating the potential of combining human and machine predictions, pointing to new and improved approaches to human-AI collaboration,” Smyth added.