The correction is the vector obtained by solving the Newton system, added to the current point to give . The inverse notation specifies the solution; a numerical implementation normally solves a linear system rather than forming a matrix inverse.
New to topics? Read the docs here!