Some functions are special cases of others, e.g. a0 * x and x are identical, but the latter is the special case a0=1. The second function is also simpler than the first, so should be preferred. Once we do the parameter optimisation for a0 * x, the code should check whether there are any integer values of a0 which give a lower description length (integers to consider can be identified as being within some number of sqrt(I_00)), much like how we have a "zero snap" which checks whether any parameter can be set to zero. This could also reduce the number of optimisations performed as more equations can be matched onto the same unique equation.
Some functions are special cases of others, e.g.
a0 * xandxare identical, but the latter is the special casea0=1. The second function is also simpler than the first, so should be preferred. Once we do the parameter optimisation fora0 * x, the code should check whether there are any integer values ofa0which give a lower description length (integers to consider can be identified as being within some number of sqrt(I_00)), much like how we have a "zero snap" which checks whether any parameter can be set to zero. This could also reduce the number of optimisations performed as more equations can be matched onto the same unique equation.