obtain, for a user equipment (UE) and from a network entity, one or more reference signals in accordance with a configuration associated with a beam failure procedure; and

perform the beam failure procedure, the beam failure procedure comprising initiating a beam failure timer in response to one or more measurements of the one or more reference signals by the UE, wherein the beam failure timer is autonomously determined by the UE using at least one of a machine learning (ML) model by inputting a set of parameters associated with the beam failure procedure into the ML model or a mapping that maps the set of parameters to a timer value for the beam failure timer and the beam failure timer includes one of a beam failure detection (BFD) timer or a beam failure recovery (BFR) timer.