本論文目的在建構一語意分類系統，使指令不單局限在單一詞彙或單一描述，例如指令旋轉之後前進及前進之前旋轉視為同樣意思，使機器人不是單純的判斷關鍵詞的順序而是使機器人能夠自行判斷語意後執行動作，讓所接受的指令更為靈活且多樣。語意分類系統建構是先以文字語句作為訓練資料，將詞彙透過詞嵌入的方式轉為數據。接著，使用神經網路進行分類訓練，主要以卷積類神經網路(Convolutional Neural Network, CNN)、長短期記憶(Long Short-Term Memory, LSTM)這兩種神經網路進行建模，CNN 具有優秀的特徵擷取及處理能力，LSTM 則在序列表現異，透過實驗比較這兩種方法，並選擇結果較好的架構應用於兩輪機器人。

The purpose of this paper is to construct a semantic classification system.The instructions are not limited to a single vocabulary or a single description, and the instructions that the robot can accept are more flexible. The construction of semantic classification system mainly needs to be modeled with two types of neural networks: Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM). CNN has excellent feature extraction and processing power, and LSTM has excellent performance in sequence. For comparison, some experiments are performed for two neural networks. Finally, the semantic classification is implemented in a two-wheeled robot.

[1] B. Pang, L. Lee, and S. Vaithyanathan , ‘‘Thumbs up?: Sentiment Classification Using Machine Learning Techniques,’’ Annual Meeting of the ACL Proceedings of the ACL-02 Conference on Empirical Methods in Natural Language Processing, Vol. 10, pp.79-86, 2002
[2] Q. Ye, R. Law, B. Gu, and W. Chen. ‘‘The influence of user-generated content on traveler behavior: An empirical investigation on the effects of e-word-of-mouth to hotel online bookings.’’ Computers in Human Behavior, 27, 634-639, 2011
[3] Q. Ye, Z. Zhang, and R. Law, ‘‘Sentiment Classification of Online Reviews to Travel Destinations by Supervised Machine Learning Approaches,’’ Expert Systems with Applications, Vol. 36, No. 3, pp. 6527-6535, 2009
[4] J. P. Turian, L. Ratinov, and Y. Bengio, “Word representations: A simple and general method for semi- supervised learning,” in ACL 2010, Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics, July 11- 16, 2010, Uppsala, Sweden, pp. 384–394, 2010.
[5] F. Liu, F. Liu, and Y. Liu, “A supervised framework for keyword extraction from meeting transcripts,” IEEE Trans. on Audio, Speech, and Language Processing, VOL. 19, NO. 3, March 2011.
[6] M. Xu, L. He, and X. Lin, ‘‘A refined TF-IDF algorithm based on channel distribution information for Web news feature extraction,’’ in Proc. 2nd Int. Workshop Educ. Technol. Comput. Sci., pp. 15–19, 2010
[7] G. E. Hinton, “Learning distributed representations of concepts,” in Proceedings of the eighth annual conference of the cognitive science society, vol. 1, p. 12,Amherst, MA, 1986.
[8] M. Li, Q. Lu, Y. Long, and L. Gui, “Inferring affective meanings of words from word embedding,” IEEE Trans. Affective Comput., vol. 8, no. 4, pp. 443–456,
2017.
[9] L. Xiang, J. Yu, C. Yang, D. Zeng, and X. Shen, “A word-embeddingbased steganalysis method for linguistic steganography via synonymsubstitution,” IEEE Access, vol. 6, pp. 64 131–64 141, 2018.
[10] Q. Wang et al., “Privacy-preserving collaborative model learning: The case of word vector training,” IEEE Trans. Knowl. Data Eng., vol. 30, no. 12, pp. 2381– 2393 , doi: 10.1109/TKDE. 2018.2819673, 2018.
[11] T. Mikolov, K. Chen, G. Corrado, and J. Dean, “Efficient estimation of word representations in vector space,” CoRR, vol. abs/1301.3781, 2013.
[12] J. Devlin, M.W. Chang, K. Lee, and K. Toutanova, “ Bert: Pre-training of deep bidirectional transformers for language understanding.” arXiv preprint arXiv:1810.04805, 2018.
[13] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N Gomez, L. Kaiser, and I. Polosukhin “Attention is all you need.” In Advances in Neural Information Processing Systems, pp. 6000–6010, 2017.
[14] Q. V. Le and T. Mikolov, “Distributed representations of sentences and documents,” in Proceedings of the 31th International Conference on Machine Machine Learning, ICML 2014, Beijing, China, 21-26 June 2014, pp. 1188–1196, 2014.
[15] T. He, W. Huang, Y. Qiao, and J. Yao, “Text-attentional convolutional neural network for scene text detection,” IEEE transactions on image processing, vol. 25, no. 6, pp.2529-2541, 2016.
[16] X. Ren, Y. Zhou, Z. Huang, J. Sun, X. Yang, and K. Chen, ‘‘A novel text structure feature extractor for Chinese scene text detection and recognition,’’ IEEE Access, vol. 5, pp. 3193–3204, 2017.
[17] Y. Kim, “Convolutional neural networks for sentence classification,” in Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing, EMNLP 2014, October 25-29, 2014, Doha, Qatar, A meeting of
SIGDAT, a Special Interest Group of the ACL, pp. 1746–1751, 2014.
[18] S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural Computation, vol. 9, no. 8, pp. 1735–1780, 1997.
[19] Martin Sundermeyer, Hermann Ney, and Ralf Schluter, ¨ “From feedforward to recurrent LSTM neural networks for language modeling,” IEEE Trans. Audio, Speech and Language Processing, vol. 23, no. 3, pp. 517–528, 2015.
[20] H. Palangi et al., ‘‘Deep sentence embedding using long short-term memory networks: Analysis and application to information retrieval,’’ IEEE/ACM Trans. Audio, Speech, Language Process., vol. 24, no. 4, pp. 694–707, Apr. 2016.
[21] B. Shuai, Z. Zuo, and G. Wang. Quaddirectional 2d-recurrent neural networks for image labeling. IEEE Signal Processing Letters, 22(11):1990–1994, 2015.
[22] P. Squartini, Paolinelli, “Comparing different recurrent neural architectures on a specific task from vanishing gradient effect perspective,” in Networking, Sensing and Control, pp. 380–385, 2006.
[23] 周建華， 具影像特徵之 LSTM 深度遞迴類神經網路之日射量預測 ， 碩士論文，國立臺灣師範大學電機工程學系，台灣，2019
[24] 林哲宇， 以 CNN 為基礎之語音辨識系統及應用於兩輪平衡車的控制 ， 碩士論文，國立臺灣師範大學電機工程學系，台灣，2019
[25] https://cloud.google.com/speech-to-text/ , 2020
[26] Y. Zhang and B. Wallace, ‘‘A sensitivity analysis of (and practitioners’ Guide to) convolutional neural networks for sentence classification.’’ 2016 [Online].
Available: https://arxiv.org/abs/1510.03820