跳到主要內容

臺灣博碩士論文加值系統

(44.211.26.178) 您好!臺灣時間:2024/06/24 22:43
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

: 
twitterline
研究生:Akanksha Gund
研究生(外文):GUND, AKANKSHA
論文名稱:T5 和 ChatGPT 餐廳評論匯總的比較研究
論文名稱(外文):A Comparative Study of T5 and ChatGPT for Restaurant Review Summarization
指導教授:邱垂昱邱垂昱引用關係
指導教授(外文):CHIU, CHUI-YU
口試委員:施博洲鄭辰仰邱垂昱
口試委員(外文):SHIH, PO-CHOUCHENG, CHEN-YANGCHIU, CHUI-YU
口試日期:2023-06-13
學位類別:碩士
校院名稱:國立臺北科技大學
系所名稱:工業工程與管理系
學門:工程學門
學類:工業工程學類
論文種類:學術論文
論文出版年:2023
畢業學年度:111
語文別:英文
論文頁數:54
中文關鍵詞:自然語言處理文本摘要情感分析T5 語言模型OpenAIChatGPT
外文關鍵詞:Natural language processingText summarizationsentiment analysisT5 language modelOpenAIChatGPT
相關次數:
  • 被引用被引用:0
  • 點閱點閱:400
  • 評分評分:
  • 下載下載:100
  • 收藏至我的研究室書目清單書目收藏:1
本研究通過結合情感分析和 自然語言處理的技術解決了中文的餐廳評論文本摘要需求。 從在線評論中提取有意義的見解對企業來說是一個挑戰。進行情感分析並將評論分為正面、負面或中性情緒。本研究使用微調的 T5 模型和ChatGPT,並使用 ROUGE-1 指標評估摘要任務。研究結果證明了該方法在情感分析和總結方面的有效性。 結果表明,兩種模型都表現良好,T5 模型取得了較佳的 ROUGE-1 分數。 模型之間的比較分析揭示了它們的相對性能。此外,人工評估表明 ChatGPT 優於對生成的摘要的定性評估。 這項研究有助於理解在現實場景中使用高級語言模型進行情感分析和摘要任務。
This study addresses the need for text summarization in restaurant reviews by combining sentiment analysis and NLP techniques. Extracting meaningful insights from online reviews is a challenge for businesses. Sentiment analysis was conducted, classifying reviews into positive, negative, or neutral sentiments. The fine-tuned T5 model and ChatGPT were used and evaluated using the ROUGE metric for summarization tasks. The evaluation demonstrates the methodology's effectiveness in sentiment analysis and summarization. Results indicated that both models performed well, with the T5 model achieving promising ROUGE-1 scores. A comparative analysis between the models revealed their relative performance. Additionally, the human evaluation indicated that ChatGPT outperformed the qualitative assessment of the generated summaries. This research contributes to understanding using advanced language models for sentiment analysis and summarization tasks in real-world scenarios.
Table of Content
Chinese Abstract --- i
ABSTRACT --- ii
Acknowledgments --- iii
Table of Content --- iv
Table of Tables --- vi
Table of Figures --- vii
1. INTRODUCTION --- 1
1.1 Research background and Motivation --- 1
1.2 Purpose of the study --- 3
1.3 Research Architecture ---3
2. Literature Review --- 5
2.1 Sentiment Analysis --- 5
2.2 T5 Model for Summarization --- 7
2.2.1 Text-To-Text Transfer Transformer (T5) --- 9
2.2.2 Model Objective --- 9
2.2.3 Downstream task --- 10
2.2.4 Attention Mask Pattern --- 10
2.3 OpenAI ChatGPT for Summarization --- 11
2.3.1 Architecture and Reinforcement Learning Process of ChatGPT --- 13
3. Research Methodology --- 15
3.1 Data Collection --- 16
3.2 Data Pre-Processing --- 17
3.3 Sentiment Analysis --- 18
3.4 Fine-tuning of T5 model and summarization --- 20
3.5 ChatGPT for Summary generation --- 24
3.6 Evaluation --- 25
3.6.1 Automated Evaluation matric --- 25
3.6.2 Human Evaluation --- 27
4. Evaluation result --- 28
4.1 Dataset Description --- 28
4.2 Sentiment Analysis Evaluation --- 28
4.3 Summarization Evaluation --- 31
4.3.1 T5 Evaluation --- 31
4.3.2 ChatGPT Evaluation --- 33
4.3.3 ChatGPT and T5 Evaluation --- 34
4.3.4 Additional Evaluation Metrics --- 36
4.3.5 Manual Evaluation --- 36
4.4 Evaluation Summary --- 39
5. Conclusion and Future scope --- 42
5.1 Future Scope --- 42
5.2 Challenges --- 43
References --- 44
Appendix --- 47
Agarwal, A., Xie, B., et al. (2011). Sentiment analysis of twitter data. Paper presented at the Proceedings of the workshop on language in social media (LSM 2011).
Alashhab, Z. R., Anbar, M., et al. (2021). Impact of coronavirus pandemic crisis on technologies and cloud computing applications. Journal of Electronic Science and Technology, 19(1), 100059.
Bornet, P., Barkin, I., et al. (2021). Intelligent automation: Welcome to the world of hyperautomation: learn how to harness artificial intelligence to boost business & make our world more human: World Scientific.
Brown, T., Mann, B., et al. (2020). Language models are few-shot learners. Advances in neural information processing systems, 33, 1877-1901.
Cao, J., Li, M., et al. (2023). A study on Prompt Design, Advantages and Limitations of ChatGPT for Deep Learning Program Repair. arXiv preprint arXiv:2304.08191.
Choudhary, A., Alugubelly, M., et al. (2023). A Comparative Study on Transformer-based News Summarization. Paper presented at the 2023 15th International Conference on Developments in eSystems Engineering (DeSE).
Clark, K., Luong, M.-T., et al. (2020). Electra: Pre-training text encoders as discriminators rather than generators. arXiv preprint arXiv:2003.10555.
Devlin, J., Chang, M.-W., et al. (2018). Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.
Farahani, M., Gharachorloo, M., et al. (2021). Leveraging ParsBERT and Pretrained mT5 for Persian Abstractive Text Summarization. Paper presented at the 2021 26th International Computer Conference, Computer Society of Iran (CSICC).
García-Saisó, S., Marti, M., et al. (2021). The COVID-19 infodemic. In (Vol. 45, pp. e56): SciELO Public Health.
Goyal, T., Li, J. J., et al. (2022). News summarization and evaluation in the era of gpt-3. arXiv preprint arXiv:2209.12356.
Hughes, A. (2023). ChatGPT: Everything you need to know about OpenAI’s GPT-3 tool. BBC Science Focus Magazine, 16.
Iwasaki, Y., Yamashita, A., et al. (2019). Japanese abstractive text summarization using BERT. Paper presented at the 2019 International Conference on Technologies and Applications of Artificial Intelligence (TAAI).
Kaur, H., & Mangat, V. (2017). A survey of sentiment analysis techniques. Paper presented at the 2017 International conference on I-SMAC (IoT in social, mobile, analytics and cloud)(I-SMAC).
Kieuvongngam, V., Tan, B., et al. (2020). Automatic text summarization of covid-19 medical research articles using bert and gpt-2. arXiv preprint arXiv:2006.01997.
Lan, Z., Chen, M., et al. (2019). Albert: A lite bert for self-supervised learning of language representations. arXiv preprint arXiv:1909.11942.
Lewis, G., & Zervas, G. (2016). The welfare impact of consumer reviews: A case study of the hotel industry. Unpublished manuscript.
Lewis, M., Liu, Y., et al. (2019). Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension. arXiv preprint arXiv:1910.13461.
45
Li, H., Liu, H., et al. (2020). Online persuasion of review emotional intensity: A text mining analysis of restaurant reviews. International Journal of Hospitality Management, 89, 102558.
Li, M., Chen, L., et al. (2021). Sentiment analysis of Chinese stock reviews based on BERT model. Applied Intelligence, 51(7), 5016-5024.
Lin, C.-Y., & Och, F. (2004). Looking for a few good metrics: ROUGE and its evaluation. Paper presented at the Ntcir workshop.
Liu, Y., Han, T., et al. (2023). Summary of chatgpt/gpt-4 research and perspective towards the future of large language models. arXiv preprint arXiv:2304.01852.
Liu, Y., Ott, M., et al. (2019). Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692.
Meek, S., Wilk, V., et al. (2021). A big data exploration of the informational and normative influences on the helpfulness of online restaurant reviews. Journal of business research, 125, 354-367.
Min, B., Ross, H., et al. (2021). Recent advances in natural language processing via large pre-trained language models: A survey. arXiv preprint arXiv:2111.01243.
Moussa, M. E., Mohamed, E. H., et al. (2018). A survey on opinion summarization techniques for social media. Future Computing and Informatics Journal, 3(1), 82-109.
Nakayama, M., & Wan, Y. (2019). The cultural impact on social commerce: A sentiment analysis on Yelp ethnic restaurant reviews. Information & Management, 56(2), 271-279.
Nallapati, R., Xiang, B., et al. (2016). Sequence-to-sequence rnns for text summarization.
Nguyen, D. Q., Vu, T., et al. (2020). BERTweet: A pre-trained language model for English Tweets. arXiv preprint arXiv:2005.10200.
OpenAI, T. (2022). Chatgpt: Optimizing language models for dialogue. OpenAI.
Ouyang, L., Wu, J., et al. (2022). Training language models to follow instructions with human feedback. Advances in neural information processing systems, 35, 27730-27744.
Pang, B., & Lee, L. (2008). Opinion mining and sentiment analysis. Foundations and Trends® in information retrieval, 2(1–2), 1-135.
Radford, A., Narasimhan, K., et al. (2018). Improving language understanding by generative pre-training.
Raffel, C., Shazeer, N., et al. (2020a). Exploring the limits of transfer learning with a unified text-to-text transformer. The Journal of Machine Learning Research, 21(1), 5485-5551.
Raffel, C., Shazeer, N., et al. (2020b). Exploring the limits of transfer learning with a unified text-to-text transformer. J. Mach. Learn. Res., 21(140), 1-67.
Rush, A. M., Chopra, S., et al. (2015). A neural attention model for abstractive sentence summarization. arXiv preprint arXiv:1509.00685.
Shen, Y., Heacock, L., et al. (2023). ChatGPT and other large language models are double-edged swords. In (Vol. 307, pp. e230163): Radiological Society of North America.
Singh, M. (2020). Summarize Reddit Comments using T5, BART, GPT-2, XLNet Models. Retrieved from Towards Data Science: https://towardsdatascience.com/summarize-reddit-comments-using-t5-bart-gpt-2-xlnet-models-a3e78a5ab944
Stiennon, N., Ouyang, L., et al. (2020). Learning to summarize with human feedback. Advances in neural information processing systems, 33, 3008-3021.
Sun, J., & Hou, X. (2022). A Roberta-Seq2Seq based model for Chinese text abstractive summarization. Paper presented at the International Conference on Advanced Algorithms and Neural Networks (AANN 2022).
Vaswani, A., Shazeer, N., et al. (2017). Attention is all you need. Advances in neural information processing systems, 30.
Wang, J., Liang, Y., et al. (2023). Cross-lingual summarization via chatgpt. arXiv preprint arXiv:2302.14229.
46
Yadav, A., & Vishwakarma, D. K. (2020). Sentiment analysis using deep learning architectures: a review. Artificial Intelligence Review, 53(6), 4335-4385.
Yang, Z., Dai, Z., et al. (2019). Xlnet: Generalized autoregressive pretraining for language understanding. Advances in neural information processing systems, 32.
Zhang, J., Zhao, Y., et al. (2020). Pegasus: Pre-training with extracted gap-sentences for abstractive summarization. Paper presented at the International Conference on Machine Learning.
Zhao, B. (2017). Web scraping. Encyclopedia of big data, 1-3.
Zolotareva, E., Tashu, T. M., et al. (2020). Abstractive Text Summarization using Transfer Learning. Paper presented at the ITAT.
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
無相關期刊