Key points are not available for this paper at this time.
As large code models become increasingly central to software development, concerns about their opacity, reliability, and resource-intensive nature are mounting. This article introduces neurosymbolic software engineering, an innovative paradigm that blends neural learning, symbolic reasoning, and controlled stochasticity.
Antonio Mastropaolo (Wed,) studied this question.