pith. machine review for the scientific record. sign in

arxiv: 2604.25683 · v1 · submitted 2026-04-28 · 💻 cs.IR

Recognition: unknown

K-CARE: Knowledge-driven Symmetrical Contextual Anchoring and Analogical Prototype Reasoning for E-commerce Relevance

Chen Yifei , Tian Zhixing , Wang Chenyang , Cheng Ziguang
Authors on Pith no claims yet

Pith reviewed 2026-05-07 15:22 UTC · model grok-4.3

classification 💻 cs.IR
keywords e-commerce search relevanceLLM knowledge groundingcontextual anchoringanalogical reasoninginformation retrievalknowledge boundariesprototype-based calibration
0
0 comments X

The pith

K-CARE fills LLM knowledge gaps in e-commerce by anchoring queries with behavior data and expert prototypes.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper claims that LLMs hit persistent performance limits on e-commerce relevance tasks because their built-in knowledge leaves a contextual void for unusual queries or niche products. This void cannot be closed just by refining reasoning steps. K-CARE therefore adds two external-knowledge components: symmetrical contextual anchoring that pulls in implicit knowledge from user behavior logs to connect queries and products, and analogical prototype reasoning that uses curated expert examples to set clearer decision boundaries inside the model's context window. If these additions work, models move from generic reasoning to domain-grounded judgments, producing more accurate relevance scores on real industrial search traffic.

Core claim

K-CARE extends the model's cognitive reach by grounding reasoning in external knowledge. It does so through symmetrical contextual anchoring, which fills the contextual void by anchoring queries and products with behavior-derived implicit knowledge, and analogical prototype reasoning, which leverages expert-curated prototypical knowledge to calibrate decision boundaries through in-context analogy.

What carries the argument

Symmetrical Contextual Anchoring (SCA) paired with Analogical Prototype Reasoning (APR), which together supply external knowledge to close parametric memory gaps that reasoning-path optimization alone cannot fix.

If this is right

  • LLM-based relevance systems can reach higher accuracy on corner-case queries and niche products without additional reasoning-trajectory training.
  • Behavior logs and expert prototypes become direct inputs that calibrate search decisions in real time.
  • Commercial platforms see measurable lifts in click-through and conversion rates from more precise relevance judgments.
  • The same two-component structure can be reused across other knowledge-intensive ranking tasks on the platform.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The method suggests that retrieval-augmented or in-context knowledge injection may be more efficient than full fine-tuning for industrial search domains.
  • Combining implicit signals from user logs with explicit expert prototypes could reduce reliance on proprietary training data while still meeting domain accuracy needs.
  • Similar anchoring-plus-analogy designs might transfer to other areas where LLMs lack specialized context, such as product recommendation or customer support routing.

Load-bearing premise

The main bottleneck for LLMs on e-commerce relevance is missing domain knowledge in parametric memory rather than limits on how well the model can optimize its reasoning steps.

What would settle it

An experiment that applies only reasoning-path optimization, without the SCA or APR knowledge components, and still matches or exceeds K-CARE's offline and online relevance gains on the same e-commerce dataset would falsify the claim that knowledge boundaries are the primary limit.

Figures

Figures reproduced from arXiv: 2604.25683 by Cheng Ziguang, Chen Yifei, Tian Zhixing, Wang Chenyang.

Figure 1
Figure 1. Figure 1: K-CARE: (1) Symmetrical Contextual Anchoring (SCA) grounds reasoning in behavior-derived implicit knowledge view at source ↗
read the original abstract

This paper targets e-commerce search relevance. While Large Language Models (LLMs) have demonstrated significant potential in this field, they often encounter performance bottlenecks in persistent 'corner cases' within complex industrial scenarios. Existing research primarily focuses on optimizing reasoning trajectories via Reinforcement Learning. However, real-world observations suggest that the primary bottleneck stems from knowledge boundaries, where the absence of domain-specific intelligence in the model's parametric memory creates a contextual void. This void persists when interpreting idiosyncratic queries or niche products and cannot be resolved solely through reasoning-path optimization. To bridge this gap, we propose K-CARE, a framework that extends the model's cognitive reach by grounding reasoning in external knowledge. K-CARE comprises two synergistic components: (1) Symmetrical Contextual Anchoring (SCA), which fills the contextual void by anchoring queries and products with behavior-derived implicit knowledge; and (2) Analogical Prototype Reasoning (APR), which leverages expert-curated prototypical knowledge to calibrate decision boundaries through in-context analogy. Extensive offline evaluations and online A/B tests on a leading e-commerce platform demonstrate that K-CARE significantly outperforms state-of-the-art baselines, delivering substantial commercial impact by resolving knowledge-intensive relevance challenges.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes the K-CARE framework for e-commerce search relevance to address LLM knowledge boundaries that create a persistent contextual void in corner cases and niche products. It introduces two components: Symmetrical Contextual Anchoring (SCA) to anchor queries/products with behavior-derived implicit knowledge, and Analogical Prototype Reasoning (APR) to calibrate decisions via expert-curated prototypical knowledge and in-context analogy. The central claim is that K-CARE significantly outperforms state-of-the-art baselines in offline evaluations and online A/B tests on a leading platform, delivering substantial commercial impact by resolving knowledge-intensive challenges that reasoning-path optimization alone cannot fix.

Significance. If the empirical claims hold with proper isolation of contributions, the work could advance LLM applications in industrial IR by demonstrating the value of explicit knowledge grounding for real-world e-commerce corner cases. The emphasis on online A/B testing provides a strength in showing potential deployability and commercial relevance beyond academic benchmarks.

major comments (2)
  1. [Abstract] Abstract: the claim that the primary bottleneck is a knowledge-boundary contextual void 'that cannot be resolved solely through reasoning-path optimization' is load-bearing for the motivation and for attributing gains to SCA/APR, yet no direct comparison (e.g., applying multi-step CoT, ReAct, or self-consistency to the identical base LLM on the same corner-case/niche queries) is described to isolate this premise from general inference improvements.
  2. [Evaluation section] Evaluation section: the abstract asserts 'extensive offline evaluations and online A/B tests' with 'significant outperformance' and 'substantial commercial impact,' but supplies no details on datasets, baselines, metrics, statistical tests, or error analysis; without these, the central empirical claim cannot be assessed.
minor comments (2)
  1. The acronyms SCA and APR should be defined at first use in the main text and consistently expanded in figure captions and tables for clarity.
  2. Consider adding a short reproducibility note on how the expert-curated prototypical knowledge for APR is collected, validated, and updated.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback, which helps strengthen the empirical grounding of our claims. We address each major comment below and will revise the manuscript accordingly to improve clarity and rigor.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that the primary bottleneck is a knowledge-boundary contextual void 'that cannot be resolved solely through reasoning-path optimization' is load-bearing for the motivation and for attributing gains to SCA/APR, yet no direct comparison (e.g., applying multi-step CoT, ReAct, or self-consistency to the identical base LLM on the same corner-case/niche queries) is described to isolate this premise from general inference improvements.

    Authors: We agree that isolating the knowledge-boundary premise from general inference gains is important for attributing improvements specifically to SCA and APR. The claim originates from our internal analysis of production logs, where advanced reasoning methods still left persistent gaps on niche products. To address this directly, the revised manuscript will add a new subsection in the Experiments section with controlled comparisons: we will apply multi-step CoT, ReAct, and self-consistency to the identical base LLM on the exact corner-case and niche query sets used in our main evaluations. We will report performance deltas and include qualitative examples showing where reasoning-path optimization alone fails to fill the contextual void. This will provide the requested isolation while preserving the original motivation. revision: yes

  2. Referee: [Evaluation section] Evaluation section: the abstract asserts 'extensive offline evaluations and online A/B tests' with 'significant outperformance' and 'substantial commercial impact,' but supplies no details on datasets, baselines, metrics, statistical tests, or error analysis; without these, the central empirical claim cannot be assessed.

    Authors: We acknowledge that the current manuscript does not provide sufficient detail in the Evaluation section for full assessment of the empirical claims. In the revision, we will substantially expand this section to include: (i) precise descriptions of the offline datasets (query counts, product categories, labeling protocol, and train/test splits); (ii) full specifications of all baselines and their hyperparameter settings; (iii) complete metric definitions and reporting (e.g., NDCG@K, MAP, and business metrics); (iv) statistical significance testing with p-values from appropriate tests (paired t-test or bootstrap); and (v) error analysis with categorized failure cases and representative examples. For the online A/B tests, we will add setup details including traffic allocation, test duration, and quantified commercial lifts with confidence intervals. These additions will make the results fully reproducible and assessable. revision: yes

Circularity Check

0 steps flagged

No circularity: framework proposal rests on external observations and evaluations, not self-referential derivation

full rationale

The paper presents K-CARE as a response to an observed bottleneck in LLMs for e-commerce relevance, attributing it to knowledge boundaries creating a contextual void that reasoning optimization alone cannot resolve. This premise is stated as arising from real-world observations rather than derived from any equations or prior self-citations within the text. The two components (SCA and APR) are introduced as new mechanisms to address the gap, with performance claims tied to offline evaluations and online A/B tests. No load-bearing steps reduce by construction to fitted inputs, self-definitions, or author-unique theorems; the derivation chain is self-contained as an empirical framework proposal without mathematical equivalence to its inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 2 invented entities

Review limited to abstract; ledger populated only from explicit statements therein. No free parameters, equations, or invented physical entities appear.

axioms (1)
  • domain assumption Primary bottleneck in LLM e-commerce relevance is knowledge boundaries rather than reasoning trajectories
    Stated directly in the abstract as the motivation for the work.
invented entities (2)
  • Symmetrical Contextual Anchoring (SCA) no independent evidence
    purpose: Fills contextual void by anchoring queries and products with behavior-derived implicit knowledge
    New component introduced in the abstract with no prior reference or independent evidence provided.
  • Analogical Prototype Reasoning (APR) no independent evidence
    purpose: Leverages expert-curated prototypical knowledge to calibrate decision boundaries through in-context analogy
    New component introduced in the abstract with no prior reference or independent evidence provided.

pith-pipeline@v0.9.0 · 5519 in / 1272 out tokens · 38349 ms · 2026-05-07T15:22:16.368426+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

33 extracted references · 21 canonical work pages · 4 internal anchors

  1. [1]

    Shuai Bai, Yuxuan Cai, Ruizhe Chen, Keqin Chen, Xionghui Chen, Zesen Cheng, Lianghao Deng, Wei Ding, Chang Gao, Chunjiang Ge, Wenbin Ge, Zhifang Guo, Qidong Huang, Jie Huang, Fei Huang, Binyuan Hui, Shutong Jiang, Zhaohai Li, Mingsheng Li, Mei Li, Kaixin Li, Zicheng Lin, Junyang Lin, Xuejing Liu, Jiawei Liu, Chenglong Liu, Yang Liu, Dayiheng Liu, Shixuan ...

    work page Pith review doi:10.48550/arxiv.2511.21631 2025

  2. [2]

    Rae, Erich Elsen, and Laurent Sifre

    Sebastian Borgeaud, Arthur Mensch, Jordan Hoffmann, Trevor Cai, Eliza Ruther- ford, Katie Millican, George van den Driessche, Jean-Baptiste Lespiau, Bogdan Damoc, Aidan Clark, Diego de Las Casas, Aurelia Guy, Jacob Menick, Roman Ring, Tom Hennigan, Saffron Huang, Loren Maggiore, Chris Jones, Albin Cassirer, Andy Brock, Michela Paganini, Geoffrey Irving, O...

    2022

  3. [4]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. InProceedings of the 2019 Conference of the North American Chapter of the Associ- ation for Computational Linguistics: Human Language Technologies, NAACL-HLT 2019, Minneapolis, MN, USA, June 2-7, 2019, V...

    work page doi:10.18653/v1/n19-1423 2019

  4. [5]

    Chenhe Dong, Shaowei Yao, Pengkun Jiao, Jianhui Yang, Yiming Jin, Zerui Huang, Xiaojiang Zhou, Dan Ou, and Haihong Tang. 2025. TaoSR1: The Thinking Model for E-commerce Relevance Search.CoRRabs/2508.12365 (2025). arXiv:2508.12365 doi:10.48550/ARXIV.2508.12365

    work page doi:10.48550/arxiv.2508.12365 2025

  5. [6]

    Kawin Ethayarajh, Winnie Xu, Niklas Muennighoff, Dan Jurafsky, and Douwe Kiela. 2024. KTO: Model Alignment as Prospect Theoretic Optimization.CoRR abs/2402.01306 (2024). arXiv:2402.01306 doi:10.48550/ARXIV.2402.01306

    work page internal anchor Pith review doi:10.48550/arxiv.2402.01306 2024

  6. [7]

    Tianyu Gao, Howard Yen, Jiatong Yu, and Danqi Chen. 2023. Enabling Large Language Models to Generate Text with Citations. InProceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, EMNLP 2023, Singapore, December 6-10, 2023, Houda Bouamor, Juan Pino, and Kalika Bali (Eds.). Association for Computational Linguistics, 6465–648...

    work page doi:10.18653/v1/2023 2023

  7. [8]

    Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yi Dai, Jiawei Sun, Qianyu Guo, Meng Wang, and Haofen Wang. 2023. Retrieval-Augmented Generation for Large Language Models: A Survey.CoRR abs/2312.10997 (2023). arXiv:2312.10997 doi:10.48550/ARXIV.2312.10997

    work page Pith review doi:10.48550/arxiv.2312.10997 2023

  8. [9]

    Shivam Garg, Dimitris Tsipras, Percy Liang, and Gregory Valiant. 2022. What Can Transformers Learn In-Context? A Case Study of Simple Function Classes. InAdvances in Neural Information Processing Systems 35: Annual Conference on Neural Information Processing Systems 2022, NeurIPS 2022, New Orleans, LA, USA, November 28 - December 9, 2022, Sanmi Koyejo, S....

    2022

  9. [10]

    Pengcheng He, Xiaodong Liu, Jianfeng Gao, and Weizhu Chen. 2021. Deberta: decoding-Enhanced Bert with Disentangled Attention. In9th International Con- ference on Learning Representations, ICLR 2021, Virtual Event, Austria, May 3-7,

    2021

  10. [11]

    https://openreview.net/forum?id=XPZIaotutsD Yifei Chen, Zhixing Tian, Chenyang Wang, and Ziguang Cheng

    OpenReview.net. https://openreview.net/forum?id=XPZIaotutsD Yifei Chen, Zhixing Tian, Chenyang Wang, and Ziguang Cheng

  11. [12]

    Po-Sen Huang, Xiaodong He, Jianfeng Gao, Li Deng, Alex Acero, and Larry P. Heck. 2013. Learning deep structured semantic models for web search using clickthrough data. In22nd ACM International Conference on Information and Knowledge Management, CIKM’13, San Francisco, CA, USA, October 27 - November 1, 2013, Qi He, Arun Iyengar, Wolfgang Nejdl, Jian Pei, a...

    work page doi:10.1145/2505515.2505665 2013

  12. [14]

    Karen Spärck Jones. 2004. A statistical interpretation of term specificity and its application in retrieval.J. Documentation60, 5 (2004), 493–502. doi:10.1108/ 00220410410560573

    2004

  13. [15]

    Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, and Radu Soricut. 2020. ALBERT: A Lite BERT for Self-supervised Learning of Language Representations. In8th International Conference on Learn- ing Representations, ICLR 2020, Addis Ababa, Ethiopia, April 26-30, 2020. OpenRe- view.net. https://openreview.net/forum?id=H1eA7AEtvS

    2020

  14. [16]

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela. 2020. Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks. InAdvances in Neural In- formation Processing Systems 33: Annual Conference on Neural Informati...

    2020

  15. [17]

    Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. RoBERTa: A Robustly Optimized BERT Pretraining Approach.CoRRabs/1907.11692 (2019). arXiv:1907.11692 http://arxiv.org/abs/1907.11692

    work page internal anchor Pith review arXiv 2019

  16. [18]

    Chenji Lu, Zhuo Chen, Hui Zhao, Zhiyuan Zeng, Gang Zhao, Junjie Ren, Ruicong Xu, Haoran Li, Songyan Liu, Pengjie Wang, Jian Xu, and Bo Zheng. 2025. LORE: A Large Generative Model for Search Relevance.CoRRabs/2512.03025 (2025). arXiv:2512.03025 doi:10.48550/ARXIV.2512.03025

    work page doi:10.48550/arxiv.2512.03025 2025

  17. [19]

    Navid Mehrdad, Hrushikesh Mohapatra, Mossaab Bagdouri, Prijith Chandran, Alessandro Magnani, Xunfan Cai, Ajit Puthenputhussery, Sachin Yadav, Tony Lee, ChengXiang Zhai, and Ciya Liao. 2024. Large Language Models for Relevance Judgment in Product Search.CoRRabs/2406.00247 (2024). arXiv:2406.00247 doi:10.48550/ARXIV.2406.00247

    work page doi:10.48550/arxiv.2406.00247 2024

  18. [20]

    Baopu Qiu, Hao Chen, Yuanrong Wu, Changtong Zan, Chao Wei, Weiru Zhang, and Xiaoyi Zeng. 2026. Thinking Broad, Acting Fast: Latent Reasoning Distil- lation from Multi-Perspective Chain-of-Thought for E-Commerce Relevance. arXiv:2601.21611 [cs.IR] https://arxiv.org/abs/2601.21611

    work page arXiv 2026

  19. [21]

    Manning, Ste- fano Ermon, and Chelsea Finn

    Rafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D. Manning, Ste- fano Ermon, and Chelsea Finn. 2023. Direct Preference Optimization: Your Language Model is Secretly a Reward Model. InAdvances in Neural Infor- mation Processing Systems 36: Annual Conference on Neural Information Pro- cessing Systems 2023, NeurIPS 2023, New Orleans, LA, USA, Dece...

    2023

  20. [22]

    Robertson, Steve Walker, Susan Jones, Micheline Hancock-Beaulieu, and Mike Gatford

    Stephen E. Robertson, Steve Walker, Susan Jones, Micheline Hancock-Beaulieu, and Mike Gatford. 1994. Okapi at TREC-3. InProceedings of The Third Text REtrieval Conference, TREC 1994, Gaithersburg, Maryland, USA, November 2-4, 1994 (NIST Special Publication, Vol. 500-225), Donna K. Harman (Ed.). National Institute of Standards and Technology (NIST), 109–12...

    1994

  21. [23]

    Gerard Salton and Michael E. Lesk. 1965. The SMART automatic document retrieval systems - an illustration.Commun. ACM8, 6 (1965), 391–398. doi:10. 1145/364955.364990

    work page arXiv 1965

  22. [24]

    Zhihong Shao, Peiyi Wang, Qihao Zhu, Runxin Xu, Junxiao Song, Mingchuan Zhang, Y. K. Li, Y. Wu, and Daya Guo. 2024. DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models.CoRRabs/2402.03300 (2024). arXiv:2402.03300 doi:10.48550/ARXIV.2402.03300

    work page Pith review doi:10.48550/arxiv.2402.03300 2024

  23. [25]

    Tian Tang, Zhixing Tian, Zhenyu Zhu, Chenyang Wang, Haiqing Hu, Guoyu Tang, Lin Liu, and Sulong Xu. 2025. LREF: A Novel LLM-based Relevance Framework for E-commerce Search. InCompanion Proceedings of the ACM on Web Conference 2025, WWW 2025, Sydney, NSW, Australia, 28 April 2025 - 2 May 2025, Guodong Long, Michale Blumestein, Yi Chang, Liane Lewin-Eytan, ...

    work page doi:10.1145/3701716.3715246 2025

  24. [26]

    Paul Thomas, Seth Spielman, Nick Craswell, and Bhaskar Mitra. 2023. Large lan- guage models can accurately predict searcher preferences.CoRRabs/2309.10621 (2023). arXiv:2309.10621 doi:10.48550/ARXIV.2309.10621

    work page doi:10.48550/arxiv.2309.10621 2023

  25. [27]

    Gomez, Lukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. Attention is All you Need. InAdvances in Neural Information Processing Systems 30: An- nual Conference on Neural Information Processing Systems 2017, December 4- 9, 2017, Long Beach, CA, USA, Isabelle Guyon, Ulrike von Luxbur...

    2017

  26. [28]

    Chenglong Wang, Canjia Li, Xingzhao Zhu, Yifu Huo, Huiyu Wang, Weixiong Lin, Yun Yang, Qiaozhi He, Tianhua Zhou, Xiaojia Chang, Jingbo Zhu, and Tong Xiao. 2026. SERM: Self-Evolving Relevance Model with Agent-Driven Learning from Massive Query Streams.CoRRabs/2601.09515 (2026). arXiv:2601.09515 doi:10.48550/ARXIV.2601.09515

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2601.09515 2026

  27. [29]

    Han Wang, Mukuntha Narayanan Sundararaman, Onur Gungor, Yu Xu, Kr- ishna Kamath, Rakesh Chalasani, Kurchi Subhra Hazra, and Jinfeng Rao. 2024. Improving Pinterest Search Relevance Using Large Language Models.CoRR abs/2410.17152 (2024). arXiv:2410.17152 doi:10.48550/ARXIV.2410.17152

    work page doi:10.48550/arxiv.2410.17152 2024

  28. [30]

    Runze Xia, Yupeng Ji, Yuxi Zhou, Haodong Liu, Teng Zhang, and Piji Li. 2025. From Reasoning LLMs to BERT: A Two-Stage Distillation Framework for Search Relevance.CoRRabs/2510.11056 (2025). arXiv:2510.11056 doi:10.48550/ARXIV. 2510.11056

    work page internal anchor Pith review doi:10.48550/arxiv 2025

  29. [31]

    Jianhui Yang, Yiming Jin, Pengkun Jiao, Chenhe Dong, Zerui Huang, Shaowei Yao, Xiaojiang Zhou, Dan Ou, and Haihong Tang. 2025. TaoSR-AGRL: Adaptive Guided Reinforcement Learning Framework for E-commerce Search Relevance. CoRRabs/2510.08048 (2025). arXiv:2510.08048 doi:10.48550/ARXIV.2510.08048

    work page doi:10.48550/arxiv.2510.08048 2025

  30. [32]

    Shaowei Yao, Jiwei Tan, Xi Chen, Juhao Zhang, Xiaoyi Zeng, and Keping Yang

  31. [33]

    KDD ’22, Association for Computing Machinery, New York, NY, USA, 3336–3346

    ReprBERT: Distilling BERT to an Efficient Representation-Based Rel- evance Model for E-Commerce. InKDD ’22: The 28th ACM SIGKDD Confer- ence on Knowledge Discovery and Data Mining, Washington, DC, USA, August 14 - 18, 2022, Aidong Zhang and Huzefa Rangwala (Eds.). ACM, 4363–4371. doi:10.1145/3534678.3539090

    work page doi:10.1145/3534678.3539090 2022

  32. [34]

    Chi, and Denny Zhou

    Michihiro Yasunaga, Xinyun Chen, Yujia Li, Panupong Pasupat, Jure Leskovec, Percy Liang, Ed H. Chi, and Denny Zhou. 2024. Large Language Models as Analogical Reasoners. InThe Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024. OpenReview.net. https://openreview.net/forum?id=AgDICX1h50

    2024

  33. [35]

    Dezhi Ye, Jie Liu, Junwei Hu, Jiabin Fan, Bowen Tian, Haijin Liang, and Jin Ma. 2025. Applying Large Language Model For Relevance Search In Tencent. In Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining, V.2, KDD 2025, Toronto ON, Canada, August 3-7, 2025, Luiza Antonie, Jian Pei, Xiaohui Yu, Flavio Chierichetti, Hady W. ...

    work page doi:10.1145/3711896.3737193 2025