Authentication from Weak PRFs with Hidden Auxiliary Input

In this work, we study a class of randomized weak pseudorandom functions, which we call weak PRFs with hidden auxiliary input (HIwPRF). Compared to Learning Parity with Noise (LPN) or Learning with Errors (LWE) based randomized weak PRFs, it provides less algebraic structure such that many known techniques and constructions do not translate to this class. We investigate the potential of HIwPRFs for secure message and user authentication. We construct a protocol that gives as strong security guarantees when instantiated with a HIwPRF as known from weak PRF, LPN or LWE based protocols.