整合檢索增強生成與大型語言模型應用於精準運動科學平台：架構與實現

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：137

、訪客IP：3.136.236.3

姓名

丁郁庭(Yu-Ting Ting) 查詢紙本館藏

畢業系所

資訊工程學系

論文名稱

整合檢索增強生成與大型語言模型應用於精準運動科學平台：架構與實現
(Integrating Retrieval-Augmented Generation and Large Language Models in a Precision Sports Science System: Architecture and Implementation)

相關論文

★ 在有干擾的虛擬教室環境下大人小孩的行為表現與腦神經反應的異同	★ 使用映射模型和跨資料集遷移式學習的輕量化居家衰弱症訓練系統
★ 心率生理回饋放鬆訓練對於海洛因使用疾患(HUD)生理資訊之影響分析	★ 基於深度學習模型的3D心理旋轉對認知障礙的診斷與評估
★ 評估注意力偵測之穿戴式腦電電極放置有效性	★ 建立數位地球：基於Omniverse平台的東南亞衛星雲圖與雷達圖可視化
★ 基於多維度的臺灣天氣類型機器學習臨近預報與分類系統	★ 透過生理數據分析的VR戰車訓練系統，評估壓力對認知專注力與穩定性的影響及通過多次訓練表現驗證系統有效性
★ 基於機器學習分析ADHD亞型利用VR認知測驗同步的神經生理數據	★ FrAIlti：利用人工智慧和3D攝影技術提升老年照護的自動化衰弱評估系統
★ 自閉症譜系障礙虛擬實境訓練系統的開發與驗證	★ 智慧醫療物聯網平台之裝置管理與應用
★ 智慧醫療物聯網平台之多租戶應用	★ XRCURE：基於實證醫學的AIOT、XR和可穿戴感測器在AWS上的數位療法
★ 重複性經顱磁刺激同步虛擬實境與生理監測用於失語症創新治療與評估	★ 將網路威脅情報與多視角分析和雙聚類結合：一種多維視覺化方法

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 (2026-7-29以後開放)

摘要(中)

精準運動科學平台收集和管理運動員的數據，為所有運動員提供高度機密和安全的數據存取空間。它支援高效搜尋和即時回饋，基於角色的存取控制和 18 種資料格式的導入，允許個性化訓練和調整。該平台也採用增強的檢索生成以及大型語言模型來增加資料相關性並產生更完整的句子。這確保了最合適的個人化健康建議和最合適的飲食調整得到完整的實施，同時資料的完整性和可用性受到加密和保護。跨領域整合、視覺化介面、預測分析、即時監控和人性化設計使教練和運動員能夠更好地理解和實施收集到的數據。這些數據和後續資訊可以有助於體重和身高、體脂率的觀察和調整，以及避免不健康的體重和體脂率。各領域運動員數據的收集和管理分為兩部分。外在負荷：透過收集運動員日常訓練、友誼賽、模擬比賽、正式比賽的數據來獲得負荷，包括體能測驗結果和學業成績。在個人特徵方面，每月收集運動員的體重、身高、體脂等生理數據，每週的食品補充劑數據和每月的運動數據。內部負荷：這部分自動收集運動員骨骼監測、生理測試、心血管測試和生物力學評估的數據。它審查和評估運動員的內部負荷。在健康問題部分，系統會自動收集運動員的健康數據。它在此收集的數據包括 LEAF 問卷和OSTRC 傷害記錄，都是每月一次。還收集有關運動員心理測試的數據，這些數據是由Garmin 設備收集的。將數據儲存在雲端共享資料庫和雲端共享平台中，研究人員可以隨時存取它們。精準運動科學平台可以分析和管理大量的運動數據，為運動員提供最佳化的訓練方案

摘要(英)

Athlete’s data is collected and managed using the Precision Sports Science Platform, which provides all athletes with a highly confidential and secure data access space. It supports efficient search and instant feedback, being based on role-based access control and the import of 18 data formats, which allow personalized training and adjustment. The platform employs enhanced retrieval generation as well as large language models to increase data relevance and generate fuller sentences. This ensures that the most suitable personalized health advice and the most appropriate dietary adjustment are implemented, while the data are confidential and protected in terms of their integrity and availability. The cross-domain integration, a visual interface, predictive analysis, real-time monitoring, and user-friendly design enable coaches and athletes to better understand and implement the collected data. The data and the following information can then contribute to the weight and height, body fat-watching and adjustment, as well as avoiding an unhealthy weight and body fat percentage. The collection and management of athlete data in various fields is divided into two parts. External Load: To obtain the load, data is collected for athletes’ daily training, friendly games, simulations competitions, and official competitions, including physical fitness test results and academic performances. In personal characteristics, data are collected monthly on the athletes’ physiologies such as weight, height body fat, weekly data on food supplements and monthly data on exercise. Internal Load: This part of the section automatically collects data on athletes’ bone monitors, physiological tests, cardiovascular tests, and biomechanical evaluations. It reviews and evaluates the athlete’s internal load. In the health problems part, the system automatically collects data on the health of the athlete. The data it collects under this include the LEAF questionnaire and OSTRC injury records, both monthly. We also gather data on the athlete’s psychological tests under this are the data collected by the Garmin devices.These data are stored in a cloud-shared database and in a cloud-shared platform, and researchers can always get access to them. In such a way, the Precision Sports Science Platform can analyze and manage a large number of sports data and can provide sportsmen with the optimal training programs.

關鍵字(中)

★ 運動科學
★ 生成式 AI
★ 大型語言模型(LLM)
★ 檢索增強生成(RAG)
★ 語意匹配
★ LangChain

關鍵字(英)

★ Sports Science
★ Generative AI
★ Large Language models(LLM)
★ Retrieval Enhancement Generation(RAG)
★ Semantic Matching
★ LangChain

論文目次

摘要 i
Abstract ii
致謝 iv
Table of Contents v
List of Figures vi
List of Tables viii
I. Introduction 1
II. Related Works 7
III. Method 22
IV. Results 47
V. Discussions 82
VI. Conclusion and Future Works 86
Reference 90

參考文獻

[1] Amir Torkashvand, "The Impact of Sports Science on Athletic Performance," Journal of Sports Science and Medicine, vol. 15, no. 3, pp. 456-467, 2016.
[2] Mark Connor and Michael O′Neill, "Large Language Models in Sport Science & Medicine: Opportunities, Risks and Considerations," DeepAI, 2023
[3] Rebelo, A., Martinho, D. V., Valente-Dos-Santos, J., Coelho-E-Silva, M. J., & Teixeira, D. S. (2023). From data to action: a scoping review of wearable technologies and biomechanical assessments informing injury prevention strategies in sport. BMC Sports Science, Medicine and Rehabilitation, 15(1), 169.
[4] Torres-Ronda, L., Beanland, E., Whitehead, S., Sweeting, A., & Clubb, J. (2022). Tracking systems in team sports: A narrative review of applications of the data and sport specific analysis. Sports Medicine - Open, 8(1), 15.
[5] Sandhu, R. S., Coyne, E. J., Feinstein, H. L., and Youman, C. E., "Role-Based Access Control Models," IEEE Computer, vol. 29, no. 2, pp. 38-47, Feb. 1996.
[6] D. F. Ferraiolo, R. Sandhu, S. Gavrila, D. R. Kuhn, and R. Chandramouli, "A Proposed Standard for Role-Based Access Control," National Institute of Standards and Technology, December 18, 2000.
[7] Y. Liu, M. Ott, N. Goyal, J. Du, M. Joshi, D. Chen, O. Levy, M. Lewis, L. Zettlemoyer, and V. Stoyanov, "Improving Text Embeddings with Large Language Models," arXiv preprint arXiv:2401.00368, 2024.
[8] Jinhyuk Lee et al, "Gecko: Versatile Text Embeddings Distilled from Large Language Models," arXiv preprint arXiv:2403.20327, 2024.
[9] Haoran Luo, Haihong E, Yuhao Yang, Gengxian Zhou, Yikai Guo, Tianyu Yao, Zichen Tang, Xueyuan Lin, Kaiyang Wan, "NQE: N-ary Query Embedding for Complex Query Answering over Hyper-Relational Knowledge Graphs," arXiv preprint arXiv:2303.20327, 2023.
[10] Dimitrios Alivanistos, Max Berrendorf, Michael Cochez, Mikhail Galkin, "Query Embedding on Hyper-relational Knowledge Graphs," in Proceedings of the International Conference on Learning Representations (ICLR), 2022.
[11] Nicola Tonellotto, Craig Macdonald, "Query Embedding Pruning for Dense Retrieval," arXiv preprint arXiv:2108.09268, 2021.
[12] Demiao LIN, "Revolutionizing Retrieval-Augmented Generation with Enhanced PDF Structure Recognition," arXiv preprint arXiv:2401.12599, 2024.
[13] Antonio Jimeno Yepes, "Financial Report Chunking for Effective Retrieval Augmented Generation," arXiv preprint arXiv:2402.05131, 2024.
[14] Y. Han et al., "A Comprehensive Survey on Vector Database: Storage and Retrieval Technique, Challenge," 2023. [Online]. Available: https://arxiv.org/abs/2310.11703
[15] J. J. Pan et al., "Survey of Vector Database Management Systems," 2023.
[16] T. Taipalus, "Vector Database Management Systems: Fundamental Concepts, Use-Cases, and Future Directions," Cognitive Systems Research, vol. 85, pp. 101216, 2024.
[17] Zhi Jing et al., "When Large Language Models Meet Vector Databases: A Survey," 2023.
[18] B. De Silva, K.-W. Huang, G. Lee, K. Hovsepian, Y. Xu, and M. Shen, "Semantic matching for text classification with complex class descriptions," in Proc. 2023 Conf. Empirical Methods Natural Language Process., Singapore: Association for Computational Linguistics, 2023.
[19] De Silva, B., Huang, K.-W., Lee, G., Hovsepian, K., Xu, Y., & Shen, M. (2023). Semantic matching for text classification with complex class descriptions. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics.
[20] Wang, B., & Li, H. (2023). Relational Sentence Embedding for Flexible Semantic Matching. In Proceedings of the 8th Workshop on Representation Learning for NLP (RepL4NLP 2023), pp. 238-252. Association for Computational Linguistics.
[21] Wu, L., Hu, J., & Du, S. (2023). Text semantic matching with an enhanced sample building method based on contrastive learning. International Journal of Machine Learning and Cybernetics, 14(9), 3105-3112.\r [22] Z. Hei, W. Liu, W. Ou, J. Qiao, J. Jiao, G. Song, T. Tian, and Y. Lin, "DR-RAG: Applying Dynamic Document Relevance to Retrieval-Augmented Generation for Question-Answering," arXiv preprint arXiv:2406.07348, 2024.
[23] Y. Hoshi, D. Miyashita, Y. Ng, K. Tatsuno, Y. Morioka and O. Torii, "Retrieval-Augmented Generation for Large Language Models: A Survey," arXiv preprint arXiv:2312.10997, 2023.
[24] Y. Gao, Y. Xiong, X. Gao, K. Jia, J. Pan, Y. Bi, Y. Dai, J. Sun, and H. Wang, "Benchmarking Large Language Models in Retrieval-Augmented Generation," arXiv preprint arXiv:2309.01431, 2023.
[25] W. Yu, M. L. Pacheco, D. Chen and N. Xue, "Retrieval-Augmented Generation across Heterogeneous Knowledge," in Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies: Student Research Workshop, Seattle, WA, USA, 2022, pp. 52-58.
[26] K. Pandya and M. Holia, "Automating Customer Service using LangChain: Building custom open-source GPT Chatbot for organizations," arXiv preprint arXiv:2310.05421, 2023.
[27] Z. Duan, "Application Development Exploration and Practice Based on LangChain+ChatGLM+Rasa," in Proc. 2023 2nd Int. Conf. Cloud Comput., Big Data Appl. Softw. Eng. (CBASE), 2023.
[28] O. Topsakal, "Creating Large Language Model Applications Utilizing LangChain: A Primer on Developing LLM Apps Fast," in Proc. Int. Conf. Appl. Eng. Nat. Sci., July 2023, doi: 10.59287/icaens.1127.
[29] Max Dean, Raymond R. Bond, Michael F. McTear, Maurice D. Mulvenna, "ChatPapers: An AI Chatbot for Interacting with Academic Research," in Proc. Int. Conf. Appl. Eng. Nat. Sci., July 2023, pp. 52-58. doi: 10.59287/icaens.1127.
[30] L. Gao, X. Ma, J. Lin, M. Zhang, and C. L. Liu, "Dense X Retrieval: What Retrieval Granularity Should We Use?" arXiv preprint arXiv:2312.06648, 2023.
[31] A. Rebelo, D. V. Martinho, J. Valente-dos-Santos, M. J. Coelho-e-Silva, and D. S. Teixeira, "From data to action: a scoping review of wearable technologies and biomechanical assessments informing injury prevention strategies in sport," BMC Sports Science, Medicine and Rehabilitation, vol. 15, Art. no. 169, Dec. 2023
[32] H. Van Eetvelde, L. D. Mendonça, C. Ley, et al., "Machine learning methods in sport injury prediction and prevention: a systematic review," Journal of Experimental Orthopaedics, vol. 8, no. 1, pp. 1-15, 2021
[33] E. Navarro, A. Navandar, S. Veiga, and A. F. San Juan Ferrer, "Applied Biomechanics: Sport Performance and Injury Prevention," Applied Sciences, vol. 11, no. 9, Art. no. 4230, Dec. 2021
[34] D. J. Bentley, "Artificial Intelligence in Sports Injury and Injury Prevention," Sports, vol. 10, no. 1, 2022, doi: 10.3390/sports10010064.
[35] M. S. Dasa, O. Friborg, M. Kristoffersen, G. Pettersen, J. V. Sagen, J. Sundgot-Borgen, and J. H. Rosenvinge, "Evaluating the suitability of the Low Energy Availability in Females Questionnaire (LEAF-Q) for female football players," Sports Medicine - Open, vol. 9, no. 1, Art. no. 54, 2023
[36] A. Melin, Å. B. Tornberg, S. Skouby, J. Faber, C. Ritz, A. Sjödin, and J. Sundgot-Borgen, "The LEAF questionnaire: a screening tool for the identification of female athletes at risk for the female athlete triad," British Journal of Sports Medicine, vol. 48, no. 7, pp. 540-545, Apr. 2014

指導教授

葉士青吳曉光(Shih-Ching Yeh Hsiao-Kuang Wu)

審核日期

2024-7-30

推文