The vulnerability of the bernoulli_p function in the decompose() function of the PyTorch machine learning framework allows attackers to compromise the confidentiality of the protected information.

The vulnerability of the bernoullip function in the PyTorch machine learning framework’s decompose method is related to inefficient calculations by the central processor. Exploiting this vulnerability could allow a malicious actor to compromise the confidentiality of the protected information...