研究生: |
洪睿澤 Hung, Jui-tse |
---|---|
論文名稱: |
以機器學習判斷量子糾纏 Machine learning of quantum entanglement |
指導教授: |
褚志崧
Chuu, Chih-Sung |
口試委員: |
王道維
Wang, Daw-Wei 王立邦 Wang, Li-Bang |
學位類別: |
碩士 Master |
系所名稱: |
理學院 - 物理學系 Department of Physics |
論文出版年: | 2019 |
畢業學年度: | 107 |
語文別: | 中文 |
論文頁數: | 82 |
中文關鍵詞: | 量子資訊 、機器學習 、量子糾纏 、監督式學習 、雙體量子態 |
外文關鍵詞: | Werner-like state, two qubit |
相關次數: | 點閱:1 下載:0 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
我們將機器學習和量子資訊兩者做結合,由CHSH Inequality(不等式)實驗量測基底作為基礎,並使用機器學習訓練CHSH 不等式來達到對量子態進行不同類別的成分判別。
在經過機器學習的訓練之後,係數經過調動的CHSH 不等式機器學習模型可被視為一個新的判別式,判別式能夠近乎100%判斷一個雙體量子態(two qubits quantum state)的Werner-like state為entangled/separable的性質。藉由同樣的機器學習的訓練,我們做出了另一種機器學習模型,該模型能夠量化任意雙體量子態的Werner-like state的糾纏程度。
之後,並試著提出可行的實驗方案,盼驗證我們所提出機器學習的模型,經過訓練後的結果。
In this thesis, we study quantum information with machine learning. Based on CHSH inequality, we employed the machine learning to develop a new inequality to classify two different types of quantum states.
The new inequality acts as a classifier to discriminate between the entangled states and separable states from a set of Werner-like state. By altering the strategy of machine learning, the new inequality can also be used to quantify the entanglement of the Werner-like states.
Finally, a feasible experimental setup is proposed to verify the quantum-state classifier studied in this thesis.
[1] Kim Y-H, Kulik S P, Shih Y 2001 Physical Review A 63 060301
[2] Kim T, Fiorentino M, Wong F N 2006 Physical Review A 73 012316
[3] Gao J, Qiao L-F, Jiao Z-Q, Ma Y-C, Hu C-Q, Ren R-J, Yang A-L, Tang H, Yung M-H, Jin X-M 2018 Physical review letters 120 240501
[4] Ma Y-C, Yung M-H 2018 npj Quantum Information 4 34
[5] Lu S, Huang S, Li K, Li J, Chen J, Lu D, Ji Z, Shen Y, Zhou D, Zeng B 2018 Physical Review A 98 012315
[6] Wootters W K 1998 Physical Review Letters 80 2245
[7] Michal Horodecki P H, Ryszard Horodecki 1996
[8] Clauser J F, Horne M A, Shimony A, Holt R A 1969 Physical review letters 23 880
[9] Rashid T 2016 Make your own neural network (CreateSpace Independent Publishing Platform)
[10] J. B. Altepeter D F V J, and P. G. Kwiat 2004