Decision Management Community

Explainability of decisions produced by machines is one of the hottest topic these days (see XAI). Explainable AI usually makes decisions using a complicated black box model, and uses a second (posthoc) model created to explain what the first model is doing. Interpretable AI concentrates on models that can themselves be directly inspected and interpreted by human experts. The recent paper “Stop explaining  black box machine learning models for high stakes decisions and use interpretable models instead” shows the difference between explainability and interpretability, and states that the former may be problematic. Link

On Dec. 4 Andrew Ng listed several important techniques  that can persuade people to trust an algorithm. “Trust isn’t just about convincing others that our solution works. I use techniques like these because I find it at least as important to convince myself that a solution works, before I ask a customer to rely on it.” Here they are: Continue reading →