Here is the definition of adjoint functors taken from here:
and here is a property about universal arrows:
To me the concept of natural bijection is imprecise. Does it coincide with the notion of natural isomorphism (so a natural transformation that has an inverse)? Bijection because in the associated commutative diagram we get mappings between sets which should be perhaps bijections?


If $F,G:\mathbf{C}\to\mathbf{D}$ are functors, a natural transformation $\alpha:F\to G$ has an inverse if and only if the component $\alpha_A$ is an isomorphism for each $A\in\mathbf{C}$. In $\mathbf{Set}$ the isomorphisms are precisely the bijections. So "natural bijection" just means a natural isomorphism in the case where the target category is $\mathbf{Set}$.
In particular homsets are sets, so natural isomorphisms involving them may be called natural bijections.