From what I can understand from the source code, the circuit is converted into a DAG before the decomposition transpiler is performed onto the DAG circuit.
How does converting to a DAG circuit help us recognise the gate operations needed to reach a certain state say for example the entangled state $\frac{1}{\sqrt{2}}(|00\rangle +|11\rangle)$. Or basically, how does that decomposition work?