Balancing health, equity, and technology for the Hispanic/Latino community.

One equation at a time.

The Framingham Risk Score, a transparent model, uses specific risk factors like age and cholesterol to calculate a patient's 10-year cardiovascular risk. Each factor is weighted by its significance (β_i), providing a clear, interpretable outcome for clinicians. This model allows healthcare professionals to easily understand how each factor contributes to overall risk, supporting informed decisions:

In a neural network, the inputs are the same patient data (age, cholesterol, blood pressure, etc.), but the data is processed through multiple hidden layers. The output is a cardiovascular risk score, similar to the Framingham model, but the relationships between the inputs and the final prediction are less interpretable:

Neurosymbolic AI bridges these two approaches. It combines the interpretability of symbolic models like the Framingham Risk Score with the learning power of neural networks. This creates systems that produce accurate cardiovascular risk predictions while maintaining transparency, helping clinicians understand not just the result, but also how different risk factors contributed to it. The solutions at Equ allow us to arrive to personalized models that adjust for hispanic/latino needs: