Henri Prade

Analogical proportions in probability distributions can significantly improve classification accuracy by revealing deeper relationships between data profiles.

Binarized neural networks can be understood through the lens of Sugeno integrals, revealing a structured way to interpret neuron decisions and input interactions.