Avicenna, AI, and the data-driven fight against COVID-19
We are currently in a unique place in the history of global population health. At a time when there is a pandemic raging, we have at our disposal advanced information technology – and most notably artificial intelligence – that can be leveraged to help fight it, and quickly. Unfortunately, most organizations are not equipped with the tools and information to understand the impact of the pandemic and to respond strategically.
At Lucd we are doing our part to help in this global fight by making Avicenna, our own highly accurate, predictive infectious disease spreading model, free to the public. Avicenna has been able to maintain 94 percent accuracy in predicting the spread of COVID-19. This has enabled researchers and leaders from not just federal, state, or local agencies, but also academic and commercial organizations to more easily generate faster and actionable insights that are intended to help with limiting the spread of the disease as well as its impact on the global workforce and commerce.
What has enabled the unique success of Avicenna in predicting disease spread?
The genesis of this predictive model came in response to the avian influenza pandemic in 2006. Working with various federal agencies in the US, this model was able to be trained on incredibly complex, population scale datasets and successfully forecast the spread of the disease down to the level of the individual.
Avicenna's primary model is the epidemiological model – which is a mathematical representation of the transmission of disease between individuals as they interact at different locations over time. This modeling approach for disease spread has been implemented on the small scale as “agent-based models” for decades. Avicenna is unique in that the fidelity of the scenarios is nearly limitless given the scalability of the underlying engine. That fidelity is what really enables Avicenna to produce results that are unprecedented in the field of infectious disease spread.
Contributing to the success of Avicenna has been the lead time that the model has had for continued improvements, as well as access to data. With 13 years of research and development, Avicenna’s mature modeling engine is able to generate robust predictions about the pandemic spread. It has been used to advise public and private organizations on the likelihood of where viruses will hit and the extent of workforce impact, all while increasing the accuracy of its predictive capabilities.
What data has Avicenna been training with?
Over the course of its evolution, Avicenna’s overall 94% accuracy level has been carefully data-driven thanks to training on extremely large population datasets. A typical epidemiological model incorporates factors to adjust for variations in different diseases, social distancing such as school closures, or work from home policies, which are now variations that most everyone is intimately familiar with. Avicenna goes much further, by including variations in population density, variations in population demographics, variations in population mobility over time, and how individuals respond to the spread of disease in their local area over time. Further Avicenna does this at the highly granular level of census tracts (i.e., neighborhoods).
With data from nearly 40 countries covering 3 billion people globally so far (we continue to extend the model to all the countries), Avicenna has been instrumental in assisting in pandemic planning and preparedness while also predicting its impact.
What are some of the immediate insights Avicenna has delivered?
As the pandemic began to take root in the US, Avicenna was able to predict the 45-day spread of the virus in cities such as New York, NY and Dallas, TX with an accuracy level reaching as high as 95%. These outputs have provided the public sector, as well as commercial leaders, with new insights intended to help with mitigating the spread of the virus, as well as how to better understand and prepare for local and regional downstream impacts such as supply chain disruptions and potential workforce planning.
What are Lucd’s next steps for Avicenna?
As stated, Avicenna is available for free to the public at https://avicenna.lucd.ai. Additionally, Lucd plans to continually expand the Avicenna platform to include new data sets and new population data to further increase the integrity of results in the short-term, as well as analyzing the long-haul effects of COVID-19 in years to come. These insights are expected to have an impact on a variety of industries, including manufacturing, healthcare, pharmaceutical and insurance.