**6. Conclusions**

Memristor-based computing platforms have been proposed to give great processing efficiency for data-centric applications, in order to address the drawbacks of conventional computing systems based on the von Neumann architecture. Despite recent major advancements in this area, there are still a lot of obstacles to overcome. To address the stochasticity and CMOS compatibility difficulties, which have an impact on the underlying electrical performance, material system engineering is essential. The integration and reliability problems might have a potential answer to tackle. To solve the concerns with variability and device non-ideality, more material and device development is still necessary. However, novel materials and architecture, or circuits have not yet reached unraveling full potential of memristor for advanced technologies.
