Multi-access mobile edge computing (MA-MEC) has been envisioned as one of the key approaches for enabling computation-intensive yet delay-sensitive services in future wireless systems. This chapter leverages non-orthogonal multiple access (NOMA) for computation offloading in MA-MEC and studies a joint optimization of the multi-access multi-task computation offloading, NOMA transmission, and computation-resource allocation, with the objective of minimizing the total energy consumption of wireless device to complete its tasks. This chapter firstly focuses on a static channel scenario and proposes a layered algorithm to solve the joint optimization problem. Furthermore, this chapter considers a dynamic channel scenario in which the channel power gains from the wireless device to the edge-computing servers are time-varying. To tackle with the difficulty due to the huge number of different channel realizations in the dynamic scenario, this chapter proposes an online algorithm, which is based on deep reinforcement learning (DRL), to efficiently learn the near-optimal offloading solutions for the time-varying channel realizations. Numerical results are provided to validate the proposed layered algorithm for the static channel scenario and the DRL-based online algorithm for the dynamic channel scenario. The chapter is organized as follows. Section 5.1 illustrates this considered system model and problem formulation. Section 5.2 presents the layered energy-efficient multi-task multi-access offloading algorithm. Section 5.3 illustrates the performance evaluation. We review the related studies in Sect. 5.4. Finally, we conclude this chapter in Sect. 5.5 and discuss the future directions.