人工智慧是時下非常熱門的議題，而隨著技術的進步，將演算法應用的晶片上的研究也愈來愈多，而演算法中的重要因素之一即是權重值，學習完的權重值需要長久地存在非揮發性類比記憶體中，讓學習的成果得以保存下來以便下次的使用而不需要重新學習。
本論文主要研究如何在台積電0.18 µm標準製程下設計非揮發性類比記憶體，本記憶體主要的設計原則為操作簡單、快速寫入和抹除、高精準度(八位元解析度)、以及演算法相容。我們的目標主要讓記憶體可以寫入到特定值之外，並可以結合脈衝時序依賴可塑性演算法的實現，進而實現在記憶體運算的想法。
本論文會先簡介脈衝時序依賴可塑性演算法，以及如何能夠在Hspice模擬中建立類比記憶體的特性，並以此特性開始去設計記憶體電路。最後再經由下線並量測，驗證其是否有符合模擬結果以及我們所推論出的結果。

Artificial intelligence is a very popular topic in recent year. With the technological advances, there are more and more researches about how to apply algorithm into chips. One of the most important factors in algorithms is weight value. The learned weight values are essential to be stored in non-volatile memory permanently so that the learning results can be saved and the results can be used next time. Besides, there is no need to learn again.
The main purpose of this research is to design an analogue non-volatile memory in TSMC 0.18 µm standard process. There are a few designing issues of this memory, which are easy operation , rapid programming/erasing, high resolution(8-bit resolution), and compatible with STDP algorithms. Our memory can be programmed to a target value, and can be implemented into STDP algorithm to achieve computing in memory.
This research will introduce STDP algorithm shortly first. Second, we will explain how to build up the behavior of the analogue memory in Hspice simulation so that we can design the circuit with this behavior. Finally, we tape out the chip and measure the result to see whether it is same as the simulation result.

[1] S. TAM, P.-K. KO, and C. HU, “Lucky-electron model of channel hot-electron injection inmosfet’s,” IEEE transctions on electron Devices, vol. ED-31, no. 9, pp. 1116–1125, 1984.
[2] K.-H. Lee, S.-C.Wang, and Y.-C. King, “Self-convergent scheme for logic-process-based multilevel/analog memory,” IEEE transctions on electron Devices, vol. 52, no. 12, pp. 2676–2681, 2005.
[3] T.-C. ONG, P.-K. KO, and C. HU, “Hot-carrier current modeling and device degradation insurface-channel p-mosfet’s,” IEEE TRANSACIORS ON ELECTRON DEVICES, vol. 37, no. 7,pp. 1658–1666, 1990.
[4]黃正達,高解析度非揮發性類比記憶體研究發展,碩士論文,國立清華大學,2007
[5]鍾國峻,嵌入式可雙向線性調適之非揮發性類比記憶體,碩士論文,國立清華大學,2011
[6]Kang, D.-H, et al., “Emulation of spike-timing dependent plasticity in nano-scale phase change memory.’ Neurocomputing,2015. 155: p. 153-158
[7] Shubha Ramakrishnan, Paul Hasler and Christal Gordon, “Floating gate synapses with spike timing dependent plasticity,” Proceedings of 2010 IEEE International Symposium on Circuits and Systems, 2010, pp. 369-372
[8]陳珮琳,嵌入式電流模式非揮發性類比記憶體研發,碩士論文,國立清華大學,2010