Rank, Chunk and Expand: Lineage-Oriented Reasoning for Taxonomy Expansion

Kumar Arjun; Sahil Mishra; Tanmoy Chakraborty

REVIEW 3 major objections 4 minor 44 references

Rank, Chunk and Expand: Lineage-Oriented Reasoning for Taxonomy Expansion

T0 review · 3 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Taxonomy expansion can be done without fine-tuning by ranking candidate parents into small chunks and prompting an LLM to reason over verbalized lineage paths; the paper reports 12% higher accuracy and 5% higher Wu & Palmer similarity…

desk verdict Solid retrieve-then-reason pipeline for taxonomy expansion, but the headline accuracy gain is overclaimed and doesn't match the paper's own Table 2. read the letter →

arxiv 2505.13282 v4 pith:CCVCPN7B submitted 2025-05-19 cs.CL

Sahil Mishra , Kumar Arjun , Tanmoy Chakraborty This is my paper · ORCID

classification cs.CL

keywords taxonomyexpansionlargelanguagemodelshypernymyhierarchicalreasoningdiscriminativerankingpromptengineeringEulertourpathverbalizationSemEval2016Task13

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Taxonomy expansion asks where a new concept should hang in an existing is-a hierarchy. LORex answers by splitting the problem into a cheap ranking step and an expensive reasoning step: a discriminative ranker orders every candidate parent by how well its verbalized lineage path matches the query, and an LLM then works through those candidates in small chunks. The generative stage filters irrelevant chunks, retrieves the most granular plausible parent, and verifies that choice against candidate paths by their average token log-probabilities. On four benchmark taxonomies the paper reports roughly 12% higher accuracy and 5% higher Wu & Palmer similarity, a structural closeness measure based on the least common ancestor, than the strongest fine-tuned baseline. The central claim is that generative hierarchy reasoning becomes reliable only after discriminative ranking has narrowed the candidate window.

What carries the argument

The load-bearing mechanism is TEMPORA's verbalized lineage path: a modified Euler tour flattens each candidate's local taxonomy into a string that visits the anchor, climbs to the root, returns through siblings, and passes through children, with every edge written as a relational phrase. Feeding that path to a pretrained encoder with a dual-path margin loss produces a ranking score for every candidate; ranking is what lets LORex chunk candidates into batches of five without losing the true parent in most cases. The second mechanism is the iterative LLM loop over chunks, consisting of a semantic filter, a parent retriever, and a path verifier that scores candidate paths by mean token log-probability, which together accept, reject, or refine a parent decision.

What would settle it

On the Food benchmark, where TEMPORA's Hit@15 is below 0.9, split the test queries into those whose true parent falls inside versus outside the top-15 window and compute LORex's accuracy separately on each group. The paper's limitation predicts near-zero accuracy on the outside group; if accuracy there is not near zero, the chunking and verification stages are recovering from ranker misses, contradicting the claim that the method is only as good as its ranker.

Watch

Extended reading notes

Core claim

The paper's discovery is a decomposition: taxonomy expansion need not be either pure retrieval or pure text generation. LORex uses TEMPORA to turn local hierarchy into a verbalized Euler-tour path, covering the anchor, ancestors to the root, siblings, and children with edges labeled is parent of and is child of, and scores those paths with a dual-path margin loss to rank all candidates. The ranked list is chunked, and an instruction-tuned LLM filters chunks, retrieves the best parent, and verifies it path by path; a rejected parent is removed and retrieval repeats on the same chunk. With three open LLMs and no fine-tuning, the system beats all comparison baselines on four benchmarks, and ablations show the ranker, the paths, and the verifier each contribute a large share of the final accuracy.

Load-bearing premise

LORex's central assumption is that the true parent appears in the top-15 ranked candidates, because the LLM only ever sees three chunks of five; the paper's limitation section concedes that the method is only as good as its ranker.

Editorial extensions

If this is right

A taxonomy can be expanded by prompting an 8B instruction-tuned model, so teams without access to paid or fine-tuned large models can still update real-world hierarchies.
Path-based verification, not bigger models, supplies most of the generative gain: switching from the path verifier to a candidate-only verifier drops accuracy by roughly 47% in the paper's ablations.
Semantic filtering makes the loop practical, cutting the number of retrieval-verification rounds per query from 3.1 to 1.7 while almost never skipping the batch that contains the true parent.
Because the main runs examine only the top-15 candidates, any improvement in the ranker's Hit@15 should translate directly into higher LORex accuracy; the paper's random-shuffle ablation shows accuracy collapsing from roughly 67% to 6% on the Environment benchmark when ranking is removed.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The same rank-chunk-verify loop is a candidate recipe for other structure-placement tasks, such as taxonomy completion or knowledge-graph completion, where one correct position must be chosen from a large candidate set; the paper only frames taxonomy expansion as the testbed.
The mean token-log-probability path scorer could be reused as a general stabilizer for small LLMs on multiple-choice reasoning, since it replaces unstable free-form path generation with a ranking over the model's own continuations.
A direct stress test the paper does not run is to place the true parent at positions 16 through 20 for the Food queries that fall outside TEMPORA's top-15 window; LORex's accuracy on exactly that subset would settle how much of the reported gain is inherited from the ranker.
In practice the limitation suggests monitoring TEMPORA's Hit@15 per taxonomy before deploying LORex, and replacing or retraining the ranker whenever that rate drops; the framework itself is agnostic to which ranker supplies the chunks.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things to know about this one. First, the architecture is genuinely interesting: a BERT-based ranker (TEMPORA) feeds chunks of candidates to an LLM that filters, retrieves, and verifies parent terms using verbalized taxonomy paths. That pipeline is new as a combination, and it avoids fine-tuning LLMs, which matters for cost. Second, the paper's headline numbers—12% accuracy and 5% Wu&P improvement over FLAME—do not survive contact with Table 2. The best relative accuracy gain on any single dataset is about 9.5% (WordNet), Food accuracy actually drops below FLAME, and the mean relative gain across the four datasets is roughly 3%. No aggregation I tried produces 12%. That is a real problem, because the paper's conclusion rests on it.

Credit where it's due. The ranker is a sensible extension of TEMP—adding sibling and child context and verbalizing edges—and it shows consistent Hit@1/Hit@10 gains over TEMP on all four benchmarks. The ablations are reasonably thorough: they test chunk sizes, filter impact, verifier types, and the ranker's contribution, and the appendix includes prompts, pseudocode, and failure cases. Code is released. This is not a toy evaluation.

Soft spots, in proportion. The 'without discarding relevant candidates' claim is contradicted by the top-15 truncation disclosed in Appendix F and by Table 5, where the semantic filter skips 45 batches that contain the true parent. That is an overstatement, not a fatal flaw. More concerning is that the 12% claim is simply not reproducible from the reported tables, and the paper never discloses how it was computed. Also, lambda_1 and lambda_2 in Eq. 6 are never reported in the hyperparameter table, which is sloppy for reproducibility. The Limitations section does acknowledge the method is only as good as its ranker—good—but it does not acknowledge the headline mismatch.

Verdict: the central pipeline is plausible and probably useful; the weakness is in the claims, not the architecture. With the aggregation disclosed and the headline recalibrated to what Table 2 actually shows, this would be a solid practical contribution. It does not reshape the field, but it gives the community a cheaper alternative to fine-tuning LLMs for taxonomy expansion.

Who it is for: practitioners building taxonomy expansion systems, and people working on retrieve-then-reason LLM pipelines. It deserves a serious referee, but the referee should ask for major revisions—specifically, correct the headline numbers, report the missing hyperparameters, and soften the 'no discarding' language. I would send it out, expecting a revise-and-resubmit.

Referee Report

3 major / 4 minor

Summary. The paper proposes LORex, a taxonomy-expansion framework that combines a discriminative candidate ranker (TEMPORA) with an LLM pipeline of semantic filtering, parent retrieval, and path-based verification. Candidates are ranked, chunked into batches, and iteratively processed; the claimed contribution is a plug-and-play method that avoids LLM fine-tuning and does not discard relevant candidates. The method is evaluated on four benchmarks (SemEval-2016 Env, Sci, Food, and WordNet) against twelve baselines, and the abstract and introduction claim a 12% accuracy improvement and a 5% Wu & Palmer (Wu&P) gain over the best baseline. The paper also includes ablations on chunk size, verifier type, ranker contribution, and semantic filtering, plus case studies and a limitations section.

Significance. If the headline results were reproducible, LORex would be a practically valuable contribution: it offers a training-light integration of discriminative ranking with LLM reasoning, and the TEMPORA ranker appears to improve over prior path-based rankers such as TEMP. The manuscript has several concrete strengths: it publicly releases code, evaluates on four benchmarks, reports ablations for each module, and provides failure-case analyses. However, the central quantitative claim is not supported by the paper's own Table 2, and the Appendix F top-15 truncation materially qualifies the 'no discarding relevant candidates' claim. These issues affect the paper's main empirical conclusion and require correction before the stated state-of-the-art claim is defensible.

major comments (3)

[Abstract, Section 1, Section 5.2, Table 2] The repeated claim that LORex improves accuracy by 12% over FLAME is not reproducible from Table 2. For LORex8B-3.1I versus FLAME, the accuracy differences are: Env 67.3 vs 63.4 (+3.9 pp), Sci 64.7 vs 63.2 (+1.5 pp), Food 55.3 vs 58.7 (-3.4 pp), WordNet 49.5 vs 45.2 (+4.3 pp). The mean absolute gain is 1.6 pp, the mean relative gain is about 3.1%, and the largest per-dataset relative gain is 9.5% (WordNet); no standard aggregation yields 12%. The Wu&P claim is also mixed: Environment decreases from 85.1 to 82.9, while the other datasets improve. Please state the exact aggregation rule used for the 12% and 5% figures and revise the abstract, introduction, and Section 5.2 accordingly.
[Appendix F, Fig. 12, Table 5] The claim that LORex expands taxonomies 'without discarding relevant candidates' is contradicted by the inference protocol in Appendix F. Inference is performed only on three chunks of five candidates, i.e., the top 15 ranked candidates, because Hit@15 exceeds 0.9 for most benchmarks; Fig. 12 shows Hit@15 is below 0.9 for Food, so a nontrivial fraction of Food queries have the true parent outside the examined window. Since the semantic filter, retriever, and verifier operate only on these chunks, the top-15 truncation is a hard upper bound on accuracy that is not disclosed in the main text. Moreover, Table 5 shows the semantic filter itself skips 45 batches containing the true parent (3 Env, 8 Sci, 13 Food, 21 WordNet). Please disclose the truncation prominently and either justify it as a design choice or remove the 'no discarding' claim.
[Section 5.1, Table 2, Section 2] CodeTaxo (Zeng et al., 2024b), which the paper cites as a third-generation LLM-based taxonomy expansion method and whose evaluation protocol it adopts in Section 5.1 and Appendix I, is not included among the baselines in Table 2. Without a comparison to CodeTaxo, the claim that LORex surpasses state-of-the-art third-generation prompting methods is not fully supported. Please add CodeTaxo to the comparison or clearly justify its exclusion (e.g., API cost) and qualify the state-of-the-art claim accordingly.

minor comments (4)

[Section 1] The introduction says LORex 'surpasses eight baselines across four benchmarks,' but Table 2 reports twelve baselines; please align these numbers.
[Section 4.1, Eqs. (3)-(7)] The notation for the losses, especially L⇕ and the summation over P and P′, is underspecified. Please define the positive and negative path sets and the scope of the summation explicitly for reproducibility.
[Figure 2] The diagram has several small labels that are difficult to read, including the text 'See d T a xon o m y' and the Euler-tour notation. Please clean up the figure so that the pipeline stages are visually clear.
[Appendix F, Table 7] The claim in Section 4.1 that TEMPORA needs only a 'minimal number of training epochs' is not fully supported by Table 7, where Food uses 50 epochs while other datasets use 20. Please explain the choice or soften the claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: LORex's evaluation is empirical and self-contained, with no claim derived from its own inputs.

full rationale

LORex is an empirical systems paper: the claimed contributions are benchmark results obtained by running a trained ranker (TEMPORA) and prompted LLMs, not by deriving a conclusion from an assumption that already contains it. TEMPORA is trained with the paper's own margin and path losses and evaluated on held-out leaves using a standard 20% split, so its Hit@k numbers are not fitted-parameters-renamed-as-predictions. The LORex pipeline prompts LLMs to filter, retrieve, and verify parents; no equation in the paper defines the output metric in terms of the input candidates by construction. The self-citations to FLAME and Quantaxo are used as a baseline and as related work, respectively, and neither is load-bearing for the central claim. The Limitations statement that 'the method is only as good as its ranker' is an acknowledged dependency, not a circular step. The abstract's 12% accuracy and 5% Wu&P gains are not reproducible from Table 2, but that is a numerical aggregation or reporting concern, which belongs under correctness risk rather than circularity. No circular step can be exhibited; therefore the appropriate finding is no significant circularity.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The reported gains rest on several tuned hyperparameters, at least two of which (lambda_1, lambda_2) are not disclosed. The pipeline assumes tree-like taxonomies, reliable LLM prompt-following, and that path log-probabilities rank validity.

free parameters (6)

chunk_size = 5 (three chunks = 15 candidates)
Chunk size is chosen by the authors; Appendix F fixes inference to 15 candidates, and Fig. 5 shows performance varies with chunk size.
margin_d = 0.2
Set per dataset in Table 7 as the margin adjustment in the dynamic margin loss (Eq. 7).
lambda_1 = not reported
Used in joint loss Eq. 6 to weight the path-difference loss; value is never stated, hurting reproducibility.
lambda_2 = not reported
Used in joint loss Eq. 6; value is never stated.
training_epochs = 20-50 per dataset
Table 7 sets epochs per benchmark; more than a 'few', despite Section 4.1 saying minimal training should suffice.
top_k_chunks = 3 chunks (15 candidates)
Inference limit set in Appendix F; determines which candidates are ever considered.

assumptions (4)

domain assumption Benchmark taxonomies are tree-structured and ground truth parents are unique most-granular hypernyms.
The Euler-tour path construction and the parent retriever prompt assume a single root path and a unique best parent; Definition 1 says DAG, but the algorithm treats taxonomies as trees.
domain assumption Average token log-probability of a generated path is a valid ranking signal for path correctness.
The parent verifier (Eq. 8) ranks candidate paths by mean token log-probability from an instruction-tuned LLM; this correlation is assumed, not proven.
domain assumption LLM responses to YES/NO and NOT FOUND prompts are reliable enough for filtering and retrieval.
The semantic filter discards batches on a Boolean LLM answer, and the retriever is trusted unless the verifier disagrees; the paper does not measure per-prompt reliability.
domain assumption GPT-4o-corrected definitions preserve the original benchmark semantics.
Appendix G says definitions were corrected with GPT-4o; if corrections alter meaning, ranking and retrieval are affected.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Rank, Chunk and Expand: Lineage-Oriented Reasoning for Taxonomy Expansion." pith.science (2026). https://pith.science/paper/CCVCPN7B

@misc{pith2026250513282,
  author       = {Pith},
  title        = {Pith review of: Rank, Chunk and Expand: Lineage-Oriented Reasoning for Taxonomy Expansion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CCVCPN7B}},
  note         = {Machine review of arXiv:2505.13282}
}

read the original abstract

Taxonomies are hierarchical knowledge graphs crucial for recommendation systems, and web applications. As data grows, expanding taxonomies is essential, but existing methods face key challenges: (1) discriminative models struggle with representation limits and generalization, while (2) generative methods either process all candidates at once, introducing noise and exceeding context limits, or discard relevant entities by selecting noisy candidates. We propose LORex (Lineage-Oriented Reasoning for Taxonomy Expansion), a plug-and-play framework that combines discriminative ranking and generative reasoning for efficient taxonomy expansion. Unlike prior methods, LORex ranks and chunks candidate terms into batches, filtering noise and iteratively refining selections by reasoning candidates' hierarchy to ensure contextual efficiency. Extensive experiments across four benchmarks and twelve baselines show that LORex improves accuracy by 12% and Wu & Palmer similarity by 5% over state-of-the-art methods.

Figures

Figures reproduced from arXiv: 2505.13282 by the authors.

**Figure 2.** An illustration of our proposed framework, [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Performance comparison of hierarchical and [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Figures from the paper (9 more)

**Figure 4.** Figure 4: Performance comparison of different types [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

**Figure 5.** Figure 5: Effect of different chunk sizes on LORex’s performance across the SemEval-2016 benchmarks. Setting Avg. Runs per Query Queries Resolved in ≤3 Runs for "Env" Semantic Filter 1.7 41 / 52 W/O Semantic Filter 3.1 18 / 52 [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

**Figure 6.** Figure 6: A hierarchical path of "water pollution" from [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]

**Figure 7.** Figure 7: An illustration of loss functions defined in [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]

**Figure 8.** Figure 8: Prompt 1 – Verbalized Path Examples of P and Pr. For a query term n i c , the discriminator TEMPORA first ranks and chunks all candidate terms N o as B. Then, we select a chunk B and apply the parent retriever to get the most appropriate parent nˆp, followed by verifi…

**Figure 9.** Figure 9: Prompt 2 – Semantic Filter Prompt Example [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]

**Figure 10.** Figure 10: Prompt 3 – Parent Retriever Prompt Example [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]

**Figure 11.** Figure 11: Prompt 4 – Parent Verifier Prompt Example [PITH_FULL_IMAGE:figures/full_fig_p017_11.png]

**Figure 12.** Figure 12: Comparison of Hit@K performance for TEMPORA across different benchmarks (Environment, Science, Food, and WordNet) [PITH_FULL_IMAGE:figures/full_fig_p017_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 28 canonical work pages

[1]

Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774

arXiv 2023
[2]

Ines Arous, Ljiljana Dolamic, and Philippe Cudr\' e -Mauroux. 2023. https://doi.org/10.1145/3543507.3583342 Taxocomplete: Self-supervised taxonomy completion leveraging position-enhanced semantic matching . In Proceedings of the ACM Web Conference 2023, WWW '23, page 2509–2518, New York, NY, USA. Association for Computing Machinery

arXiv 2023
[3]

Mohit Bansal, David Burkett, Gerard de Melo, and Dan Klein. 2014. Structured learning for taxonomy induction with belief propagation. In Proceedings of the 52nd ACL, pages 1041--1051, Baltimore, Maryland. Association for Computational Linguistics

work page 2014
[4]

Zhen Bi, Jing Chen, Yinuo Jiang, Feiyu Xiong, Wei Guo, Huajun Chen, and Ningyu Zhang. 2024. Codekgc: Code language model for generative knowledge graph construction. ACM Transactions on Asian and Low-Resource Language Information Processing, 23(3):1--16

2024
[5]

Georgeta Bordea, Els Lefever, and Paul Buitelaar. 2016 a . Semeval-2016 task 13: Taxonomy extraction evaluation (texeval-2). In Proceedings of the 10th international workshop on semantic evaluation (semeval-2016), pages 1081--1091

work page 2016
[6]

Georgeta Bordea, Els Lefever, and Paul Buitelaar. 2016 b . S em E val-2016 task 13: Taxonomy extraction evaluation ( TE x E val-2). In SemEval, pages 1081--1091, San Diego, California. Association for Computational Linguistics

work page 2016
[7]

Haw-Shiuan Chang, Ziyun Wang, Luke Vilnis, and Andrew McCallum. 2018. Distributional inclusion vector embedding for unsupervised hypernymy detection. In NAACL, pages 485--495, New Orleans, Louisiana. Association for Computational Linguistics

work page 2018
[8]

Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. https://doi.org/10.18653/v1/N19-1423 BERT : Pre-training of deep bidirectional transformers for language understanding . In Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long a...

doi:10.18653/v1/n19-1423 2019

Show all 44 references

[9]

Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Amy Yang, Angela Fan, et al. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783

2024 arXiv
[10]

Gon c alves, Matthew Horridge, Rui Li, Yu Liu, Mark A

Rafael S. Gon c alves, Matthew Horridge, Rui Li, Yu Liu, Mark A. Musen, Csongor I. Nyulas, Evelyn Obamos, Dhananjay Shrouty, and David Temple. 2019. Use of owl and semantic web technologies at pinterest. In The Semantic Web – 18th ISWC 2019, Auckland, New Zealand, October 26–3...

2019
[11]

Rik DT Janssen and Henderik A Proper. 2021. A functionality taxonomy for document search engines. arXiv preprint arXiv:2105.12989

2021 arXiv
[12]

Minhao Jiang, Xiangchen Song, Jieyu Zhang, and Jiawei Han. 2022. Taxoenrich: Self-supervised taxonomy completion via structure-semantic representations. In WWW, page 925–934, New York, NY, USA. ACM

2022
[13]

Song Jiang, Qiyue Yao, Qifan Wang, and Yizhou Sun. 2023. A single vector is not enough: Taxonomy expansion via box embeddings. In WWW, page 2467–2476, New York, NY, USA. Association for Computing Machinery

2023
[14]

Bowen Jin, Chulin Xie, Jiawei Zhang, Kashob Kumar Roy, Yu Zhang, Zheng Li, Ruirui Li, Xianfeng Tang, Suhang Wang, Yu Meng, et al. 2024. Graph chain-of-thought: Augmenting large language models by reasoning on graphs. arXiv preprint arXiv:2404.07103

2024 arXiv
[15]

David Jurgens and Mohammad Taher Pilehvar. 2015. Reserating the awesometastic: An automatic extension of the W ord N et taxonomy for novel terms. In NAACL, pages 1459--1465, Denver, Colorado. Association for Computational Linguistics

2015
[16]

David Jurgens and Mohammad Taher Pilehvar. 2016. Semeval-2016 task 14: Semantic taxonomy enrichment. In Proceedings of the 10th international workshop on semantic evaluation (SemEval-2016), pages 1092--1102

2016
[17]

Giannis Karamanolakis, Jun Ma, and Xin Luna Dong. 2020. TXtract : Taxonomy - Aware Knowledge Extraction for Thousands of Product Categories . Journal Name, pages 8489--8502

2020
[18]

Ben Liu, Jihai Zhang, Fangquan Lin, Cheng Yang, and Min Peng. 2025. https://aclanthology.org/2025.coling-main.740/ Filter-then-generate: Large language models with structure-text adapter for knowledge graph completion . In Proceedings of the 31st International Conference on Co...

2025
[19]

Zichen Liu, Hongyuan Xu, Yanlong Wen, Ning Jiang, HaiYing Wu, and Xiaojie Yuan. 2021. TEMP : Taxonomy expansion with dynamic margin loss through taxonomy-paths. In EMNLP, pages 3854--3863, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics

2021
[20]

Emaad Manzoor, Rui Li, Dhananjay Shrouty, and Jure Leskovec. 2020. Expanding taxonomies with implicit edge semantics. In WWW, page 2044–2054, New York, NY, USA. Association for Computing Machinery

2020
[21]

Yuning Mao, Tong Zhao, Andrey Kan, Chenwei Zhang, Xin Luna Dong, Christos Faloutsos, and Jiawei Han. 2020. Octet: Online catalog taxonomy enrichment with self-supervision. In SIGKDD, page 2247–2257, New York, NY, USA

2020
[22]

Sahil Mishra, Avi Patni, Niladri Chatterjee, and Tanmoy Chakraborty. 2025. Quantaxo: A quantum approach to self-supervised taxonomy expansion. arXiv preprint arXiv:2501.14011

2025
[23]

Sahil Mishra, Ujjwal Sudev, and Tanmoy Chakraborty. 2024. https://doi.org/10.1145/3709007 Flame: Self-supervised low-resource taxonomy expansion using large language models . ACM Trans. Intell. Syst. Technol. Just Accepted

2024 doi
[24]

Alexander Panchenko, Stefano Faralli, Eugen Ruppert, Steffen Remus, Hubert Naets, C \'e drick Fairon, Simone Paolo Ponzetto, and Chris Biemann. 2016. TAXI at S em E val-2016 task 13: a taxonomy induction method based on lexico-syntactic patterns, substrings and focused crawlin...

2016
[25]

David Schellenberger Costa, Gerhard Boehnisch, Martin Freiberg, Rafaël Govaerts, Matthias Grenié, Michael Hassler, Jens Kattge, Alexandra Muellner-Riehl, Blanca Rojas Andrés, Marten Winter, Mark Watson, Alexander Zizka, and Christian Wirth. 2023. https://doi.org/10.1111/nph.18...

2023 doi
[26]

Conrad L Schoch, Stacy Ciufo, Mikhail Domrachev, Carol L Hotton, Sivakumar Kannan, Rogneda Khovanskaya, Detlef Leipe, Richard Mcveigh, Kathleen O’Neill, Barbara Robbertse, Shobha Sharma, Vladimir Soussov, John P Sullivan, Lu Sun, Seán Turner, and Ilene Karsch-Mizrachi. 2020. h...

2020 doi
[27]

Jiaming Shen, Zhihong Shen, Chenyan Xiong, Chi Wang, Kuansan Wang, and Jiawei Han. 2020. Taxoexpan: Self-supervised taxonomy expansion with position-enhanced graph neural network. In WWW, page 486–497, New York, NY, USA. Association for Computing Machinery

2020
[28]

Artur Strzelecki and Paulina Rutecka. 2019. The snippets taxonomy in web search engines. In Perspectives in Business Informatics Research: 18th International Conference, BIR 2019, Katowice, Poland, September 23--25, 2019, Proceedings 18, pages 177--188. Springer

2019
[29]

Jiashuo Sun, Chengjin Xu, Lumingyuan Tang, Saizhuo Wang, Chen Lin, Yeyun Gong, Lionel Ni, Heung-Yeung Shum, and Jian Guo. 2024. https://openreview.net/forum?id=nnVO1PvbTv Think-on-graph: Deep and responsible reasoning of large language model on knowledge graph . In The Twelfth...

2024
[30]

Kunihiro Takeoka, Kosuke Akimoto, and Masafumi Oyamada. 2021. Low-resource taxonomy enrichment with pretrained language models. In EMNLP, pages 2747--2758, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics

2021
[31]

Xingyu Tan, Xiaoyang Wang, Qing Liu, Xiwei Xu, Xin Yuan, and Wenjie Zhang. 2024. Paths-over-graph: Knowledge graph empowered large language model reasoning. arXiv preprint arXiv:2410.14211

2024 arXiv
[32]

Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. 2023. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288

2023 arXiv
[33]

Suyuchen Wang, Ruihui Zhao, Xi Chen, Yefeng Zheng, and Bang Liu. 2021. Enquire one’s parent and child before decision: Fully exploit hierarchical structure for self-supervised taxonomy expansion. In WWW, page 3291–3304, New York, NY, USA. Association for Computing Machinery

2021
[34]

Xingyao Wang, Sha Li, and Heng Ji. 2022. Code4struct: Code generation for few-shot event structure prediction. arXiv preprint arXiv:2210.12810

2022 arXiv
[35]

Yanbin Wei, Qiushi Huang, Yu Zhang, and James Kwok. 2023. https://doi.org/10.18653/v1/2023.findings-emnlp.580 KICGPT : Large language model with knowledge in context for knowledge graph completion . In Findings of the Association for Computational Linguistics: EMNLP 2023, page...

2023 doi
[36]

Fei Xia, Yixuan Weng, Shizhu He, Kang Liu, and Jun Zhao. 2023. https://doi.org/10.18653/v1/2023.eacl-main.73 Find parent then label children: A two-stage taxonomy completion method with pre-trained language model . In Proceedings of the 17th Conference of the European Chapter ...

2023 doi
[37]

Hongyuan Xu, Yunong Chen, Zichen Liu, Yanlong Wen, and Xiaojie Yuan. 2022. Taxoprompt: A prompt-based generation method with taxonomic context for self-supervised taxonomy expansion. In IJCAI-22, pages 4432--4438. IJCAI. Main Track

2022
[38]

Hongyuan Xu, Ciyi Liu, Yuhang Niu, Yunong Chen, Xiangrui Cai, Yanlong Wen, and Xiaojie Yuan. 2023. T aco P rompt: A collaborative multi-task prompt learning method for self-supervised taxonomy completion. In EMNLP, pages 15804--15817, Singapore. Association for Computational L...

2023
[39]

Hongyuan Xu, Yuhang Niu, Yanlong Wen, and Xiaojie Yuan. 2025. https://openreview.net/forum?id=vrqvcVZL3Y Compress and mix: Advancing efficient taxonomy completion with large language models . In THE WEB CONFERENCE 2025

2025
[40]

Yue Yu, Yinghao Li, Jiaming Shen, Hao Feng, Jimeng Sun, and Chao Zhang. 2020. Steam: Self-supervised taxonomy expansion with mini-paths. In SIGKDD, page 1026–1035, New York, NY, USA. Association for Computing Machinery

2020
[41]

Qingkai Zeng, Yuyang Bai, Zhaoxuan Tan, Shangbin Feng, Zhenwen Liang, Zhihan Zhang, and Meng Jiang. 2024 a . Chain-of-layer: Iteratively prompting large language models for taxonomy induction from limited examples. In Proceedings of the 33rd ACM International Conference on Inf...

2024
[42]

Qingkai Zeng, Yuyang Bai, Zhaoxuan Tan, Zhenyu Wu, Shangbin Feng, and Meng Jiang. 2024 b . Codetaxo: Enhancing taxonomy expansion with limited examples via code language prompts. arXiv preprint arXiv:2408.09070

2024 arXiv
[43]

URL: " 'urlintro :=

ENTRY address author booktitle chapter edition editor howpublished institution journal key month note number organization pages publisher school series title type volume year eprint doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before...
[44]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774

arXiv 2023

[2] [2]

Ines Arous, Ljiljana Dolamic, and Philippe Cudr\' e -Mauroux. 2023. https://doi.org/10.1145/3543507.3583342 Taxocomplete: Self-supervised taxonomy completion leveraging position-enhanced semantic matching . In Proceedings of the ACM Web Conference 2023, WWW '23, page 2509–2518, New York, NY, USA. Association for Computing Machinery

arXiv 2023

[3] [3]

Mohit Bansal, David Burkett, Gerard de Melo, and Dan Klein. 2014. Structured learning for taxonomy induction with belief propagation. In Proceedings of the 52nd ACL, pages 1041--1051, Baltimore, Maryland. Association for Computational Linguistics

work page 2014

[4] [4]

Zhen Bi, Jing Chen, Yinuo Jiang, Feiyu Xiong, Wei Guo, Huajun Chen, and Ningyu Zhang. 2024. Codekgc: Code language model for generative knowledge graph construction. ACM Transactions on Asian and Low-Resource Language Information Processing, 23(3):1--16

2024

[5] [5]

Georgeta Bordea, Els Lefever, and Paul Buitelaar. 2016 a . Semeval-2016 task 13: Taxonomy extraction evaluation (texeval-2). In Proceedings of the 10th international workshop on semantic evaluation (semeval-2016), pages 1081--1091

work page 2016

[6] [6]

Georgeta Bordea, Els Lefever, and Paul Buitelaar. 2016 b . S em E val-2016 task 13: Taxonomy extraction evaluation ( TE x E val-2). In SemEval, pages 1081--1091, San Diego, California. Association for Computational Linguistics

work page 2016

[7] [7]

Haw-Shiuan Chang, Ziyun Wang, Luke Vilnis, and Andrew McCallum. 2018. Distributional inclusion vector embedding for unsupervised hypernymy detection. In NAACL, pages 485--495, New Orleans, Louisiana. Association for Computational Linguistics

work page 2018

[8] [8]

Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. https://doi.org/10.18653/v1/N19-1423 BERT : Pre-training of deep bidirectional transformers for language understanding . In Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long a...

doi:10.18653/v1/n19-1423 2019