TY - RPRT TI - Variation Due to Regularization Tractably Recovers Bayesian Deep Learning AU - James McInerney AU - Nathan Kallus PY - 2025 UR - https://arxiv.org/abs/2403.10671 ID - 2403.10671 ER -