近年來，人機互動對話系統在不同的領域被廣泛應用。而為了能更有效應用於心理健康照顧領域，需要額外引入情感支持的對話技巧。透過使用此對話技巧於對話系統，即使在無人陪伴的情況下，情緒低落的人也能找到與之交談的對象，並且解決遇到的情緒困擾。
本論文介紹了一種創新的情感支持對話系統，透過在三種不同的對話階段轉換來達到逐漸降低使用者情緒困擾的目的，分別為探索使用者遇到的問題、以同理心安慰使用者以及為使用者面臨的問題提供建議，並且在每個階段使用不同的策略生成對話。在現代社會的高壓環境中提供心理健康上的幫助，是此系統主要的目的以及訴求。
本研究的主要方向是將情感支持的技能融入到對話系統當中。其中對話的三個階段：探索、安慰、建議以及八種不同的策略是情感支持技能中最重要的特色，且在不同階段都有各自可以使用的策略，故兩者有著相互映射的關係。我們利用圖神經網路對於情感支持中的階段轉換以及每個階段的策略選擇進行建模，利用不同階段之間的轉換關係、階段和策略之間的共享資訊，使得可以更適合地決定當前的階段以及策略。此外，我們在每個階段除了在策略上的選擇，也為各個階段提供適合的額外知識，使得生成的對話有助於達到不同階段的目標，也能更符合當前使用者所提及的對話主題與方向，提升整體的用戶體驗。
最後，我們的實驗結果顯示，和比較基準的情感支持系統相比，我們的系統在PPL、BLEU、Rouge-L以及Distinct指標上都取得了顯著的進步。與最新的 基線系統相比，我們的系統在PPL降低了0.75，Distinct-1提升了0.05，Distinct-2提升了0.83，Bleu-2提升了1.28，Bleu-4提升了0.67，Rouge-L提升了2.3，同時也藉由統計生成樣本在95% 信賴區間在各個指標的顯著性以證明我們的系統確實和基線系統有著顯著差異。此外，在階段預測以及策略預測的準確率上都有顯著的提升。這些效能提升的結果證實了我們方法的有效性，使用圖神經網路進行階段以及策略的建模，並搭配給予每個階段所需要的額外知識，使得我們的系統可以準確決定以及表示階段和策略，並幫助我們的回話生成可以達到情緒支持的效果。

In recent years, human-machine interaction dialogue systems have been widely applied in various fields. To be more effectively utilized in the field of mental health care, additional emotional support dialogue techniques are needed. By incorporating these techniques into dialogue systems, individuals experiencing emotional distress can find someone to talk to and address their emotional concerns even in the absence of companionship.
This paper introduces an innovative emotional support dialogue system designed to gradually alleviate user emotional distress through three distinct dialogue stages: exploring user issues, providing empathetic comfort, and offering advice on user problems. Different conversation strategies are employed at each stage. The primary aim and appeal of this system lie in providing psychological assistance in today's high-pressure societal environment.
The main focus of this research is integrating emotional support skills into dialogue systems. The three dialogue stages - exploration, comfort, and advice - along with eight different strategies represent the most crucial features of emotional support skills. There is a reciprocal relationship between these stages and strategies, with various strategies available for use at different stages. We model the stage transitions and strategy selection within emotional support using graph neural networks, leveraging shared information between stages and strategies to more accurately determine the current stage and strategy. Additionally, apart from strategy selection, we provide relevant additional knowledge for each stage to ensure that the generated dialogue contributes to achieving the goals of different stages and aligns with the current user's conversation topics and direction, thereby enhancing the overall user experience.
Finally, our experimental results demonstrate significant improvements over the baseline emotion support system in terms of PPL, BLEU, Rouge-L, and Distinct metrics. Compared to the latest baseline system, our system shows improvements: PPL reduced by 0.75, Distinct-1 increased by 0.05, Distinct-2 increased by 0.83, Bleu-2 increased by 1.28, Bleu-4 increased by 0.67, and Rouge-L increased by 2.3. Additionally, statistical significance tests with 95% confidence intervals on generated samples demonstrate significant differences between our system and the baseline system. Moreover, there are significant enhancements in accuracy for both stage prediction and strategy prediction. These performance improvements validate the effectiveness of our approach, which utilizes graph neural networks for stage and strategy modeling, along with providing additional knowledge required for each stage. This enables our system to accurately determine and represent stages and strategies, aiding in achieving emotional support in our response generation.

[1] H. Rashkin, E. M. Smith, M. Li, and Y.-L. Boureau, "Towards empathetic open-domain conversation models: A new benchmark and dataset," arXiv preprint arXiv:1811.00207, 2018.
[2] C. E. Hill, Helping skills: Facilitating exploration, insight, and action. American Psychological Association, 2020.
[3] S. Suthaharan and S. Suthaharan, "Support vector machine," Machine learning models and algorithms for big data classification: thinking with examples for effective learning, pp. 207-235, 2016.
[4] K. P. Murphy, "Naive bayes classifiers," University of British Columbia, vol. 18, no. 60, pp. 1-8, 2006.
[5] W. Zaremba, I. Sutskever, and O. Vinyals, "Recurrent neural network regularization," arXiv preprint arXiv:1409.2329, 2014.
[6] A. Graves and A. Graves, "Long short-term memory," Supervised sequence labelling with recurrent neural networks, pp. 37-45, 2012.
[7] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, "Attention is all you need," Advances in neural information processing systems, vol. 30, 2017.
[8] E. Reiter and R. Dale, "Building applied natural language generation systems," Natural Language Engineering, vol. 3, no. 1, pp. 57-87, 1997.
[9] S. R. Eddy, "Hidden markov models," Current opinion in structural biology, vol. 6, no. 3, pp. 361-365, 1996.
[10] B. Roark, M. Saraclar, and M. Collins, "Discriminative n-gram language modeling," Computer Speech & Language, vol. 21, no. 2, pp. 373-392, 2007.
[11] K. W. Church, "Word2Vec," Natural Language Engineering, vol. 23, no. 1, pp. 155-162, 2017.
[12] A. Baevski, Y. Zhou, A. Mohamed, and M. Auli, "wav2vec 2.0: A framework for self-supervised learning of speech representations," Advances in neural information processing systems, vol. 33, pp. 12449-12460, 2020.
[13] J. Pennington, R. Socher, and C. D. Manning, "Glove: Global vectors for word representation," in Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), 2014, pp. 1532-1543.
[14] I. Sutskever, O. Vinyals, and Q. V. Le, "Sequence to sequence learning with neural networks," Advances in neural information processing systems, vol. 27, 2014.
[15] J. Chung, C. Gulcehre, K. Cho, and Y. Bengio, "Empirical evaluation of gated recurrent neural networks on sequence modeling," arXiv preprint arXiv:1412.3555, 2014.
[16] A. Radford, J. Wu, R. Child, D. Luan, D. Amodei, and I. Sutskever, "Language models are unsupervised multitask learners," OpenAI blog, vol. 1, no. 8, p. 9, 2019.
[17] T. Brown, B. Mann, N. Ryder, M. Subbiah, J. D. Kaplan, P. Dhariwal, A. Neelakantan, P. Shyam, G. Sastry, and A. Askell, "Language models are few-shot learners," Advances in neural information processing systems, vol. 33, pp. 1877-1901, 2020.
[18] A. Chowdhery, S. Narang, J. Devlin, M. Bosma, G. Mishra, A. Roberts, P. Barham, H. W. Chung, C. Sutton, and S. Gehrmann, "Palm: Scaling language modeling with pathways," Journal of Machine Learning Research, vol. 24, no. 240, pp. 1-113, 2023.
[19] S. Roller, E. Dinan, N. Goyal, D. Ju, M. Williamson, Y. Liu, J. Xu, M. Ott, K. Shuster, and E. M. Smith, "Recipes for building an open-domain chatbot," arXiv preprint arXiv:2004.13637, 2020.
[20] A. Ohashi and R. Higashinaka, "Post-processing networks: Method for optimizing pipeline task-oriented dialogue systems using reinforcement learning," arXiv preprint arXiv:2207.12185, 2022.
[21] T.-H. Wen, D. Vandyke, N. Mrksic, M. Gasic, L. M. Rojas-Barahona, P.-H. Su, S. Ultes, and S. Young, "A network-based end-to-end trainable task-oriented dialogue system," arXiv preprint arXiv:1604.04562, 2016.
[22] L. P. Kaelbling, M. L. Littman, and A. W. Moore, "Reinforcement learning: A survey," Journal of artificial intelligence research, vol. 4, pp. 237-285, 1996.
[23] N. Bang, J. Lee, and M.-W. Koo, "Task-optimized adapters for an end-to-end task-oriented dialogue system," arXiv preprint arXiv:2305.02468, 2023.
[24] E. M. Smith, M. Williamson, K. Shuster, J. Weston, and Y.-L. Boureau, "Can you put it all together: Evaluating conversational agents' ability to blend skills," arXiv preprint arXiv:2004.08449, 2020.
[25] Z. Lin, A. Madotto, J. Shin, P. Xu, and P. Fung, "Moel: Mixture of empathetic listeners," arXiv preprint arXiv:1908.07687, 2019.
[26] S. Sabour, C. Zheng, and M. Huang, "Cem: Commonsense-aware empathetic response generation," in Proceedings of the AAAI Conference on Artificial Intelligence, 2022, vol. 36, no. 10, pp. 11229-11237.
[27] S. Liu, C. Zheng, O. Demasi, S. Sabour, Y. Li, Z. Yu, Y. Jiang, and M. Huang, "Towards emotional support dialog systems," arXiv preprint arXiv:2106.01144, 2021.
[28] Q. Tu, Y. Li, J. Cui, B. Wang, J.-R. Wen, and R. Yan, "Misc: a mixed strategy-aware model integrating comet for emotional support conversation," arXiv preprint arXiv:2203.13560, 2022.
[29] A. Van Den Oord and O. Vinyals, "Neural discrete representation learning," Advances in neural information processing systems, vol. 30, 2017.
[30] J. Cheng, S. Sabour, H. Sun, Z. Chen, and M. Huang, "Pal: Persona-augmented emotional support conversation generation," arXiv preprint arXiv:2212.09235, 2022.
[31] J. Zhou, Z. Chen, B. Wang, and M. Huang, "Facilitating multi-turn emotional support conversation with positive emotion elicitation: A reinforcement learning approach," arXiv preprint arXiv:2307.07994, 2023.
[32] N. Shazeer, A. Mirhoseini, K. Maziarz, A. Davis, Q. Le, G. Hinton, and J. Dean, "Outrageously large neural networks: The sparsely-gated mixture-of-experts layer," arXiv preprint arXiv:1701.06538, 2017.
[33] W. Peng, Y. Hu, L. Xing, Y. Xie, Y. Sun, and Y. Li, "Control globally, understand locally: A global-to-local hierarchical graph network for emotional support conversation," arXiv preprint arXiv:2204.12749, 2022.
[34] F. Scarselli, M. Gori, A. C. Tsoi, M. Hagenbuchner, and G. Monfardini, "The graph neural network model," IEEE transactions on neural networks, vol. 20, no. 1, pp. 61-80, 2008.
[35] Y. Liu, M. Ott, N. Goyal, J. Du, M. Joshi, D. Chen, O. Levy, M. Lewis, L. Zettlemoyer, and V. Stoyanov, "Roberta: A robustly optimized bert pretraining approach," arXiv preprint arXiv:1907.11692, 2019.
[36] W. Hamilton, Z. Ying, and J. Leskovec, "Inductive representation learning on large graphs," Advances in neural information processing systems, vol. 30, 2017.
[37] J. D. Hwang, C. Bhagavatula, R. Le Bras, J. Da, K. Sakaguchi, A. Bosselut, and Y. Choi, "(comet-) atomic 2020: On symbolic and neural commonsense knowledge graphs," in Proceedings of the AAAI Conference on Artificial Intelligence, 2021, vol. 35, no. 7, pp. 6384-6392.
[38] J. S. Bridle, "Probabilistic interpretation of feedforward classification network outputs, with relationships to statistical pattern recognition," in Neurocomputing: Algorithms, architectures and applications: Springer, 1990, pp. 227-236.
[39] C. Szegedy, S. Ioffe, V. Vanhoucke, and A. Alemi, "Inception-v4, inception-resnet and the impact of residual connections on learning," in Proceedings of the AAAI conference on artificial intelligence, 2017, vol. 31, no. 1.
[40] J. L. Ba, J. R. Kiros, and G. E. Hinton, "Layer normalization," arXiv preprint arXiv:1607.06450, 2016.
[41] X. Glorot, A. Bordes, and Y. Bengio, "Deep sparse rectifier neural networks," in Proceedings of the fourteenth international conference on artificial intelligence and statistics, 2011: JMLR Workshop and Conference Proceedings, pp. 315-323.
[42] R. Caruana, "Multitask learning," Machine learning, vol. 28, pp. 41-75, 1997.
[43] M. Lewis, Y. Liu, N. Goyal, M. Ghazvininejad, A. Mohamed, O. Levy, V. Stoyanov, and L. Zettlemoyer, "Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension," arXiv preprint arXiv:1910.13461, 2019.
[44] K. Papineni, S. Roukos, T. Ward, and W.-J. Zhu, "Bleu: a method for automatic evaluation of machine translation," in Proceedings of the 40th annual meeting of the Association for Computational Linguistics, 2002, pp. 311-318.
[45] C.-Y. Lin, "Rouge: A package for automatic evaluation of summaries," in Text summarization branches out, 2004, pp. 74-81.
[46] J. Li, M. Galley, C. Brockett, J. Gao, and B. Dolan, "A diversity-promoting objective function for neural conversation models," arXiv preprint arXiv:1510.03055, 2015.
[47] L. Ferrer and P. Riera, "Confidence Intervals for evaluation in machine learning." [Online]. Available: https://github.com/luferrer/ConfidenceIntervals.