Conceptmissing probability mass
Missing probability mass
A concept in information theory and statistical learning referring to the phenomenon in which low-frequency events in a training distribution have probability mass close to zero and are therefore ignored or insufficiently modeled during learning.
Timeline (last 90 days)
Nov 3
The study proposes from the perspective of "missing probability mass" that monofacts are the key to triggering scientific discovery.
Unverified50%