I am trying to prove that the automorphisms of the Klein four-group is isomorphic to $S_3$.
The first insight is the Klein group is an abelian group of four elements, three of which are order $2$, and must be mapped to other elements of order $2$, and the identity must of course be mapped to itself.
So any automorphism "permutes" the non-identity elements.
So there are exactly $3! = 6$ of them, matching the number of elements of $S_3$. Then if I call those non-identity maps $f_2, f_2, \ldots, f_6$, I can naturally map them to $S_3$.
For example, if the non-identity elements of the Klein group are $a$, $b$, $c$, I can regard these as $1$, $2$, and $3$ (I believe the labelling doesn't matter) and if I swapped $a$ and $b$ and fixed $c$, I can regard this as the transposition $(12)$.
By repeating this, I get a natural bijection of sets.
I don't know how to "prove" it's a bijection other than to observe the mapping and that it hits everything in $S_3$ exactly once.
I don't know, in particular, how to prove that this is a homomorphism, other than by checking every possible pair of isomorphisms, which is a bit laborious.
I assume there is a better, cleaner way to do it.
UPDATE: After working on this problem more, I have an updated attempt. It's incomplete, but I think it makes some progress.
We realize the Klein 4-group as $V = \{e, a, b, c\}$, where $a,b,c$ are the non-identity elements, each of which has order $2$. Any automorphism $f$ of $V$ must send elements of order $2$ to elements of order $2$, so necessarily it must permute the elements $a,b,c$. I claim that such a permutation is necessarily an automorphism. We observe that the Klein four-group has the property that if we multiply two non-identity elements, we get the third. Since $f$ maps from $V$ to $V$ bijectively, it is also the case that when $f(g)f(g') = f(g'')$, where $g$ and $g'$ are distinct elements and $g''$ is the third. That is, for any $g \neq g' \neq e$, we have \begin{align*} f(gg') = f(g'') = f(g)f(g''). \end{align*} If we take an arbitrary $g \in V$ and then consider multiplying with the identity $e$, we have \begin{align*} f(ge) = f(g) = e f(g) = f(e) f(g), \end{align*} since $f(e) = e$. This proves that $h$ is a homomorphism, and since it is bijective by definition, it is an automorphism. Since every automorphism must send elements of order $2$ to elements of order $2$ and there are exactly $3! = 6$ permutations of $\{a,b,c\}$, there are exactly six elements of Aut(G).
An automorphism is, therefore, uniquely determined on how it permutes the elements ${a,b,c}$, giving a natural bijective map to $S_3$, relabelling these non-identity elements as $1,2,3$. I call $e$ the identity permutation and $f_{abc}$ as the automorphism that maps ${a,b,c} \to {a,b,c}$ (i.e., the identity map) and so forth. The map is: \begin{align*} & f_{a,b,c} \longmapsto e & & \text{both have order $1$} \\ & f_{a,c,b} \longmapsto (23) & & \text{both have order $2$} \\ & f_{b,a,c} \longmapsto (12) & & \text{both have order $2$} \\ & f_{c,b,a} \longmapsto (13) & & \text{both have order $2$} \\ & f_{b,c,a} \longmapsto (123) & & \text{both have order $3$} \\ & f_{c,a,b} \longmapsto (132) & & \text{both have order $3$} \end{align*} This is clearly a bijection of sets: every permutation in $S_3$ was hit exactly once. I still do not know how to show that this is a homomorphism. I constructed the map so that it preserved the order of elements, but is that enough? I could check that $\varphi(xy) = \varphi(x) \varphi(y)$ for each pair, but there must be a way to do it that I'm missing.
Good thing about $V$ is that $V \cong \mathbb{Z}_{2}\times\mathbb{Z}_{2}$. And there is a nice approach, because once you determined where $(1,0)$ and $(0,1)$ go the whole automorphism is defined. So let's represent automorphism as a matrix $$\left(\begin{array}{cc} a & b\\ c& d \end{array}\right),$$ where $(1,0)\mapsto (a,b)$ and $(0,1)\mapsto (c,d)$. It easy to show that such matrix is invertible. More than that, composition of two automorphism is given by multiplication of their matrices. So let $G$ be a group of invertible $2\times 2$ matrices with $\{0,1\}$ coefficients. $|G|=6$ and we can easily list its elements: $$\left(\begin{array}{cc} 1 & 0\\ 0& 1 \end{array}\right),\left(\begin{array}{cc} 1 & 1\\ 0& 1 \end{array}\right),\left(\begin{array}{cc} 1 & 0\\ 1& 1 \end{array}\right),\left(\begin{array}{cc} 1 & 1\\ 1& 0 \end{array}\right), \left(\begin{array}{cc} 0 & 1\\ 1& 1 \end{array}\right),\left(\begin{array}{cc} 0 & 1\\ 1& 0 \end{array}\right) $$ From here we can easily determine "transpositions", because they are of order 2. We can match them with matrices of order 2 and then extend to isomorphism (because transpositions generate symmetric group and such matrices satisfy relations). This is a bit laborious too, but gives some kind of systematic approach.