{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,16]],"date-time":"2026-05-16T16:17:04Z","timestamp":1778948224367,"version":"3.51.4"},"reference-count":71,"publisher":"Association for Computing Machinery (ACM)","issue":"4","content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Inf. Syst."],"published-print":{"date-parts":[[2026,5,31]]},"abstract":"<jats:p>\n                    Large Language Models (LLMs) have been extensively applied in various recommendation scenarios, including bundle generation, thanks to their exceptional reasoning capabilities and comprehensive knowledge. However, exploiting large-scale LLMs for bundle generation introduces significant efficiency challenges\u2014primarily high computational costs during fine-tuning and inference due to their massive parameterization. Knowledge Distillation (KD) offers a promising solution by transferring expertise from large teacher models to more compact student models. This study systematically investigates KD approaches for bundle generation with the goal of minimizing computational demands while preserving performance. Specifically, we explore three critical research questions: (1) how does the\n                    <jats:italic toggle=\"yes\">format of distilled knowledge<\/jats:italic>\n                    impact bundle generation performance? (2) to what extent does the\n                    <jats:italic toggle=\"yes\">quantity of distilled knowledge<\/jats:italic>\n                    influence the performance? and (3) how do different\n                    <jats:italic toggle=\"yes\">ways of utilizing the distilled knowledge<\/jats:italic>\n                    affect the performance? To support this investigation, we propose a comprehensive KD framework that (i) progressively extracts knowledge from raw data in increasingly complex forms, i.e., frequent patterns\n                    <jats:inline-formula content-type=\"math\/tex\">\n                      <jats:tex-math notation=\"LaTeX\" version=\"MathJax\">\\(\\rightarrow\\)<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    formalized rules\n                    <jats:inline-formula content-type=\"math\/tex\">\n                      <jats:tex-math notation=\"LaTeX\" version=\"MathJax\">\\(\\rightarrow\\)<\/jats:tex-math>\n                    <\/jats:inline-formula>\n                    deep thoughts; (ii) captures varying quantities of distilled knowledge through different sampling strategies, multi-domain accumulation, and multi-format aggregation; and (iii) exploits complementary LLM adaptation techniques\u2014in-context learning, supervised fine-tuning, and their combination\u2014to leverage the distilled knowledge for domain-specific adaptation and enhanced efficiency in small student models. Through extensive experiments on multiple real-world datasets, we provide valuable insights into how knowledge format, quantity, and utilization methods collectively shape the performance of LLM-based bundle generation, which exhibits the significant potential of KD for more efficient yet effective LLM-based bundle generation.\n                  <\/jats:p>","DOI":"10.1145\/3808223","type":"journal-article","created":{"date-parts":[[2026,4,21]],"date-time":"2026-04-21T14:29:20Z","timestamp":1776781760000},"page":"1-40","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":2,"title":["Does Knowledge Distillation Matter for Large Language Model-Based Bundle Generation?"],"prefix":"10.1145","volume":"44","author":[{"ORCID":"https:\/\/orcid.org\/0009-0002-2601-5537","authenticated-orcid":false,"given":"Kaidong","family":"Feng","sequence":"first","affiliation":[{"name":"Yanshan University, Qinhuangdao, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3350-7022","authenticated-orcid":false,"given":"Zhu","family":"Sun","sequence":"additional","affiliation":[{"name":"Information Systems Technology and Design, Singapore University of Technology and Design, Singapore, Singapore"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0350-0313","authenticated-orcid":false,"given":"Jie","family":"Yang","sequence":"additional","affiliation":[{"name":"Delft University of Technology, Delft, Netherlands"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9788-6634","authenticated-orcid":false,"given":"Hui","family":"Fang","sequence":"additional","affiliation":[{"name":"Key Laboratory of Interdisciplinary Research of Computation and Economics and School of Computing and Artificial Intelligence, Shanghai University of Finance and Economics, Shanghai, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8072-2019","authenticated-orcid":false,"given":"Xinghua","family":"Qu","sequence":"additional","affiliation":[{"name":"Bytedance (Seed), Singapore, Singapore"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2186-4153","authenticated-orcid":false,"given":"Wenyuan","family":"Liu","sequence":"additional","affiliation":[{"name":"Yanshan University, Qinhuangdao, China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"320","published-online":{"date-parts":[[2026,5,11]]},"reference":[{"key":"e_1_3_2_2_2","first-page":"487","volume-title":"Proceedings of 20th International Conference on Very Large Scale Data Bases (VLDB)","author":"Agrawal Rakesh","year":"1994","unstructured":"Rakesh Agrawal and Ramakrishnan Srikant. 1994. Fast algorithms for mining association rules. In Proceedings of 20th International Conference on Very Large Scale Data Bases (VLDB), 487\u2013499."},{"key":"e_1_3_2_3_2","doi-asserted-by":"publisher","DOI":"10.1145\/3523227.3546754"},{"key":"e_1_3_2_4_2","doi-asserted-by":"publisher","DOI":"10.1145\/3308558.3313568"},{"key":"e_1_3_2_5_2","doi-asserted-by":"publisher","DOI":"10.1145\/3604915.3608857"},{"key":"e_1_3_2_6_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.knosys.2016.08.013"},{"key":"e_1_3_2_7_2","doi-asserted-by":"publisher","DOI":"10.1145\/3077136.3080779"},{"key":"e_1_3_2_8_2","doi-asserted-by":"publisher","DOI":"10.1145\/3397271.3401198"},{"issue":"3","key":"e_1_3_2_9_2","first-page":"2326","article-title":"Bundle recommendation and generation with graph neural networks","volume":"35","author":"Chang Jianxin","year":"2021","unstructured":"Jianxin Chang, Chen Gao, Xiangnan He, Depeng Jin, and Yong Li. 2021. Bundle recommendation and generation with graph neural networks. IEEE Transactions on Knowledge and Data Engineering 35, 3 (2021), 2326\u20132340.","journal-title":"IEEE Transactions on Knowledge and Data Engineering"},{"key":"e_1_3_2_10_2","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2019\/290"},{"key":"e_1_3_2_11_2","doi-asserted-by":"publisher","DOI":"10.1145\/3292500.3330652"},{"key":"e_1_3_2_12_2","doi-asserted-by":"publisher","DOI":"10.1145\/3640457.3688118"},{"key":"e_1_3_2_13_2","doi-asserted-by":"publisher","DOI":"10.1145\/3604915.3610646"},{"key":"e_1_3_2_14_2","doi-asserted-by":"publisher","DOI":"10.1145\/3474085.3475440"},{"key":"e_1_3_2_15_2","doi-asserted-by":"publisher","DOI":"10.1145\/3340531.3412734"},{"key":"e_1_3_2_16_2","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0441"},{"key":"e_1_3_2_17_2","doi-asserted-by":"publisher","DOI":"10.1145\/3240323.3240348"},{"key":"e_1_3_2_18_2","doi-asserted-by":"publisher","DOI":"10.1109\/DSC.2018.00076"},{"key":"e_1_3_2_19_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.dss.2006.05.005"},{"key":"e_1_3_2_20_2","doi-asserted-by":"publisher","DOI":"10.1145\/2559169"},{"key":"e_1_3_2_21_2","doi-asserted-by":"publisher","DOI":"10.1177\/0092070306288405"},{"key":"e_1_3_2_22_2","doi-asserted-by":"publisher","DOI":"10.1145\/3604915.3610639"},{"key":"e_1_3_2_23_2","doi-asserted-by":"publisher","DOI":"10.1145\/2872427.2883037"},{"key":"e_1_3_2_24_2","doi-asserted-by":"publisher","DOI":"10.1145\/3357384.3358030"},{"key":"e_1_3_2_25_2","doi-asserted-by":"publisher","DOI":"10.1145\/3336191.3371776"},{"key":"e_1_3_2_26_2","doi-asserted-by":"publisher","DOI":"10.1145\/3583780.3614949"},{"key":"e_1_3_2_27_2","doi-asserted-by":"publisher","DOI":"10.1145\/3523227.3546755"},{"key":"e_1_3_2_28_2","unstructured":"Geoffrey Hinton Oriol Vinyals and Jeff Dean. 2015. Distilling the knowledge in a neural network. arXiv:1503.02531. Retrieved from https:\/\/arxiv.org\/abs\/1503.02531"},{"key":"e_1_3_2_29_2","doi-asserted-by":"publisher","DOI":"10.1145\/3589334.3645377"},{"key":"e_1_3_2_30_2","doi-asserted-by":"publisher","DOI":"10.1145\/3298689.3347003"},{"key":"e_1_3_2_31_2","first-page":"9459","article-title":"Retrieval-augmented generation for knowledge-intensive nlp tasks","volume":"33","author":"Lewis Patrick","year":"2020","unstructured":"Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich K\u00fcttler, Mike Lewis, Wen-Tau Yih, Tim Rockt\u00e4schel, et al. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Processing Systems 33 (2020), 9459\u20139474.","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_3_2_32_2","doi-asserted-by":"publisher","DOI":"10.1145\/3589334.3645467"},{"key":"e_1_3_2_33_2","doi-asserted-by":"publisher","DOI":"10.1145\/3022185"},{"key":"e_1_3_2_34_2","doi-asserted-by":"publisher","DOI":"10.1109\/ICDM.2011.118"},{"key":"e_1_3_2_35_2","doi-asserted-by":"publisher","DOI":"10.1145\/3690624.3709255"},{"key":"e_1_3_2_36_2","doi-asserted-by":"publisher","DOI":"10.1145\/2645710.2645750"},{"key":"e_1_3_2_37_2","doi-asserted-by":"publisher","DOI":"10.1145\/3640810"},{"key":"e_1_3_2_38_2","doi-asserted-by":"publisher","DOI":"10.1145\/3534678.3539229"},{"key":"e_1_3_2_39_2","doi-asserted-by":"publisher","DOI":"10.1145\/3664647.3681349"},{"key":"e_1_3_2_40_2","doi-asserted-by":"publisher","DOI":"10.1145\/3616855.3635854"},{"key":"e_1_3_2_41_2","first-page":"46534","volume-title":"Advances in Neural Information Processing Systems (NeurIPS)","author":"Madaan Aman","year":"2023","unstructured":"Aman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan, Luyu Gao, Sarah Wiegreffe, Uri Alon, Nouha Dziri, Shrimai Prabhumoye, Yiming Yang, et al. 2023. Self-refine: Iterative refinement with self-feedback. In Advances in Neural Information Processing Systems (NeurIPS), 46534\u201346594."},{"key":"e_1_3_2_42_2","doi-asserted-by":"publisher","DOI":"10.1145\/3077136.3080724"},{"key":"e_1_3_2_43_2","doi-asserted-by":"publisher","DOI":"10.1145\/3539618.3591666"},{"key":"e_1_3_2_44_2","first-page":"452","volume-title":"Proceedings of the 25th Conference on Uncertainty in Artificial Intelligence (UAI)","author":"Rendle Steffen","year":"2009","unstructured":"Steffen Rendle, Christoph Freudenthaler, Zeno Gantner, and Lars Schmidt-Thieme. 2009. BPR: Bayesian personalized ranking from implicit feedback. In Proceedings of the 25th Conference on Uncertainty in Artificial Intelligence (UAI), 452\u2013461."},{"key":"e_1_3_2_45_2","doi-asserted-by":"publisher","DOI":"10.1145\/3626772.3657783"},{"key":"e_1_3_2_46_2","doi-asserted-by":"publisher","DOI":"10.1145\/2872427.2883057"},{"key":"e_1_3_2_47_2","unstructured":"Wenqi Sun Ruobing Xie Junjie Zhang Wayne Xin Zhao Leyu Lin and Ji-Rong Wen. 2024. Distillation is all you need for practically using different pre-trained recommendation models. arXiv:2401.00797v1. Retrieved from https:\/\/arxiv.org\/html\/2401.00797v1"},{"key":"e_1_3_2_48_2","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3652865","article-title":"Revisiting bundle recommendation for intent-aware product bundling","volume":"24","author":"Sun Zhu","year":"2024","unstructured":"Zhu Sun, Kaidong Feng, Jie Yang, Hui Fang, Xinghua Qu, Yew-Soon Ong, and Wenyuan Liu. 2024. Revisiting bundle recommendation for intent-aware product bundling. ACM Transactions on Recommender Systems 2, 3, Article 24 (2024), 24:1\u201324:34.","journal-title":"ACM Transactions on Recommender Systems"},{"key":"e_1_3_2_49_2","doi-asserted-by":"publisher","DOI":"10.1145\/3626772.3657808"},{"key":"e_1_3_2_50_2","doi-asserted-by":"publisher","DOI":"10.1145\/3626772.3657688"},{"key":"e_1_3_2_51_2","doi-asserted-by":"publisher","DOI":"10.1145\/3477495.3531904"},{"key":"e_1_3_2_52_2","unstructured":"Wenjie Wang Xinyu Lin Fuli Feng Xiangnan He and Tat-Seng Chua. 2023. Generative recommendation: Towards next-generation recommender paradigm. arXiv:2304.03516. Retrieved from https:\/\/arxiv.org\/abs\/2304.03516"},{"key":"e_1_3_2_53_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-short.6"},{"key":"e_1_3_2_54_2","first-page":"12648","volume-title":"Proceedings of the 11th International Conference on Learning Representations (ICLR)","author":"Wang Xuezhi","year":"2023","unstructured":"Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc Le, Ed Chi, Sharan Narang, Aakanksha Chowdhery, and Denny Zhou. 2023. Self-consistency improves chain of thought reasoning in language models. In Proceedings of the 11th International Conference on Learning Representations (ICLR), 12648\u201312671."},{"key":"e_1_3_2_55_2","unstructured":"Yan Wang Zhixuan Chu Xin Ouyang Simeng Wang Hongyan Hao Yue Shen Jinjie Gu Siqiao Xue James Y. Zhang Qing Cui et al. 2023. Enhancing recommender systems with large language model reasoning graphs. arXiv:2308.10835. Retrieved from https:\/\/arxiv.org\/abs\/2308.10835"},{"key":"e_1_3_2_56_2","doi-asserted-by":"publisher","DOI":"10.1145\/3589334.3645671"},{"key":"e_1_3_2_57_2","doi-asserted-by":"publisher","DOI":"10.1145\/3477495.3531909"},{"key":"e_1_3_2_58_2","doi-asserted-by":"publisher","DOI":"10.1145\/3539618.3591771"},{"key":"e_1_3_2_59_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.findings-emnlp.280"},{"key":"e_1_3_2_60_2","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v38i8.28769"},{"key":"e_1_3_2_61_2","unstructured":"Likang Wu Zhi Zheng Zhaopeng Qiu Hao Wang Hongchao Gu Tingjia Shen Chuan Qin Chen Zhu Hengshu Zhu Qi Liu et al. 2023. A survey on large language models for recommendation. arXiv:2305.19860. Retrieved from https:\/\/arxiv.org\/abs\/2305.19860"},{"key":"e_1_3_2_62_2","doi-asserted-by":"publisher","DOI":"10.1145\/1864708.1864739"},{"key":"e_1_3_2_63_2","doi-asserted-by":"publisher","DOI":"10.14778\/2733085.2733099"},{"key":"e_1_3_2_64_2","doi-asserted-by":"publisher","DOI":"10.1145\/3626772.3657719"},{"key":"e_1_3_2_65_2","doi-asserted-by":"publisher","DOI":"10.1145\/2396761.2398608"},{"key":"e_1_3_2_66_2","unstructured":"Yang Yu Fangzhao Wu Chuhan Wu Jingwei Yi and Qi Liu. 2021. Tiny-newsrec: Effective and efficient plm-based news recommendation. arXiv:2112.00944. Retrieved from https:\/\/arxiv.org\/abs\/2112.00944"},{"key":"e_1_3_2_67_2","doi-asserted-by":"crossref","unstructured":"Xu Yuan Hongshen Chen Yonghao Song Xiaofang Zhao Zhuoye Ding Zhen He and Bo Long. 2021. Improving sequential recommendation consistency with self-supervised imitation. arXiv:2106.14031. Retrieved from https:\/\/arxiv.org\/abs\/2106.14031","DOI":"10.24963\/ijcai.2021\/457"},{"key":"e_1_3_2_68_2","doi-asserted-by":"publisher","DOI":"10.1145\/3581783.3612252"},{"key":"e_1_3_2_69_2","doi-asserted-by":"publisher","DOI":"10.1145\/3604915.3608860"},{"key":"e_1_3_2_70_2","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v36i4.20359"},{"key":"e_1_3_2_71_2","doi-asserted-by":"publisher","DOI":"10.1145\/2600428.2609603"},{"key":"e_1_3_2_72_2","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2023.3329175"}],"container-title":["ACM Transactions on Information Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3808223","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,16]],"date-time":"2026-05-16T15:43:41Z","timestamp":1778946221000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3808223"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,5,11]]},"references-count":71,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2026,5,31]]}},"alternative-id":["10.1145\/3808223"],"URL":"https:\/\/doi.org\/10.1145\/3808223","relation":{},"ISSN":["1046-8188","1558-2868"],"issn-type":[{"value":"1046-8188","type":"print"},{"value":"1558-2868","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,5,11]]},"assertion":[{"value":"2025-04-23","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-04-04","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-05-11","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}