Probabilistic protection refers to cyber security strategies that apply an algorithm (often a ML algorithm) to existing data on cyber attacks with the goal of determining the odds (probability) that a given online behavior constitutes an attack. This probability then informs the tool's decision on whether to treat behaviors that it has not seen before as security threats. A key difference between probabilistic and deterministic protection is that probabilistic solutions do not always treat the same behavior the same way, often leading to a higher rate of false positives.
