Machine Learning Powers Accurate Pandemic Predictions

• Researchers at The Florey have developed a human-machine hybrid model that outperforms human predictions in forecasting future events.
• The model combines the wisdom of crowds with AI technology to analyze extensive data from COVID-19-related questions posted on a prediction market.
• The model has been shown to be effective in forecasting events or risks, especially in situations where reliable data sources are lacking.

In a groundbreaking development, researchers and data scientists at The Florey have harnessed the power of artificial intelligence (AI) to enhance the accuracy of forecasting future events.

Their innovative approach combines the wisdom of crowds with AI technology, resulting in a human-machine hybrid model that outperforms human predictions, particularly in the context of COVID-19-related events.

The team’s findings, detailed in a paper published in eBioMedicine, emphasize the critical role of accurate forecasting in decision-making, whether it involves responding to a pandemic, predicting election outcomes, or assessing economic trends. Professor Anne-Louise Ponsonby, the senior author of the paper, noted that forecasting in the realm of public health is especially challenging, as highlighted by the COVID-19 pandemic.

Prediction markets, which tap into the collective intelligence of diverse participants to forecast specific outcomes, have shown promise in previous studies compared to traditional methods such as surveys, expert panels, and polls. The researchers at The Florey leveraged artificial intelligence to analyze extensive data from COVID-19-related questions posted on the Almanis forecasting platform, managed by Dysrupt Labs.

Their AI-driven approach involved identifying unique characteristics, patterns, and past performance of forecasters to generate real-time assessments of their prediction accuracy, referred to as “trade quality,” on the prediction market. Forecasts with better track records were given greater weight, resulting in significantly improved prediction accuracy.

This innovative method yielded enhanced event prediction capabilities across various independent datasets, including the Next Generation Social Science Program. Notably, when the hybrid model and human-only predictions diverged by 5 percentage points or more in terms of event likelihood, the hybrid model consistently outperformed the human-only model. This was reflected in an impressive Area Under the Curve (AUC) accuracy score of 0.90 for the hybrid model compared to 0.77 for the human-only model (with a score of 1 indicating perfect prediction and 0.5 equivalent to chance).

Lead author and Florey data scientist, Alex Gruen, highlighted the potential applications of this hybrid approach in forecasting events or risks, especially in situations where reliable data sources are lacking or uncertainties regarding human behavior are significant. Gruen emphasized the daily decision-making impact of predicting future events on both individual and collective levels and how this hybrid model could substantially enhance prediction accuracy. Moreover, he emphasized its potential to bolster responses to emerging risks such as pandemics and climate change.

Also Read

AI in Photography: Most Photographers Worried About the Impact

Most Photographers Believe Traditional Photography Remains Necessary Despite AI Advances. RPS Survey:...

To stay informed about current events, please like our Facebook page https://www.facebook.com/BOLUrduNews/.
Follow us on Twitter https://twitter.com/bolnewsurdu01 and stay updated with the latest news.
Subscribe to our YouTube channel https://bit.ly/3Tv8a3P to watch news from Pakistan and around the world