The identification of different types of symmetries in an unknown quantum process is a crucial step in understanding its dynamics. Typically, this is achieved by iteratively performing the following steps: preparation of an initial state, evolution of the state over the process, and measurement of the evolved state. Such conventional algorithms require an exponential number of accesses to the process in terms of the number of qubits. Recently, a more efficient quantum-enhanced algorithm has been proposed. This algorithm stores quantum information about the evolved state using a quantum memory. This enables the possibility of multiple accesses to the quantum process before measurement, unlike conventional algorithms that require a measurement after each access. Building on previous work, this thesis extends the quantum-enhanced algorithm to classify different forms of time-reversal and particle-hole symmetry in an unknown quantum process. We theoretically prove that this algorithm has constant complexity in terms of number of accesses, offering an exponential advantage over conventional algorithms. By implementing it on a quantum simulator, we validate its functionality and demonstrate that even in the presence of noise in the quantum process, we can achieve symmetry classification with a reasonable number of accesses. These results highlight how utilizing quantum computational resources dramatically enhances our ability to learn symmetries.