Pith. sign in

REVIEW 3 major objections 6 minor 57 references

mRAKL: Multilingual Retrieval-Augmented Knowledge Graph Construction for Low-Resourced Languages

T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that knowledge-graph completion for low-resourced languages can be reframed as retrieval-augmented question answering, so missing tail entities are read from Wikipedia context rather than recalled from scarce structured…

desk verdict Useful new datasets and a sensible RAG pipeline for low-resource KGC, but the headline gains come from an oracle retriever and the real BM25 wins are small. read the letter →

arxiv 2507.16011 v1 pith:RXDRCWE7 submitted 2025-07-21 cs.CL

Hellina Hailu Nigatu , Min Li , Maartje ter Hoeve , Saloni Potdar , Sarah Chasins This is my paper
classification cs.CL
keywords retrieval-augmentedgenerationmultilingualknowledgegraphconstructionlow-resourcedlanguagesTigrinyaAmhariccross-linguallinkpredictionquestionansweringtransfer
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

mRAKL argues that knowledge-graph construction (KGC) for low-resourced languages should be reframed as question answering: each (head, relation, tail) triple becomes a template question, and a generator predicts the tail using Wikipedia context retrieved at inference time. The paper contributes new KGs and QA datasets for Tigrinya (3.5k triples) and Amharic (34k triples), with Arabic and English as transfer languages, and motivates the design by showing that the structured context prior methods rely on — entity descriptions and one-hop neighbors — is rarely available for these languages and contains the answer less than 2% of the time. Its headline result is that the context-augmented setup beats the KGT5 sequence-to-sequence baselines by 4.92 percentage points on Tigrinya and 8.79 on Amharic, measured as the share of test triples whose single top prediction equals the true tail. Those numbers, however, come from an idealized '(im)perfect' retriever that finds the gold tail inside the head entity's Wikipedia article; the deployed BM25 retriever improves over the no-context setting by only about 1.2 and 2.2 points, so how much of the gain survives real retrieval remains an open question.

What carries the argument

The load-bearing mechanism is the language-tagged QA input sequence. Every triple is converted into a template question Q (head entity and relation) with the tail as answer, the retriever supplies up to two context sentences C, and the generator input is [C-LANt]C | [Q-LANt]Q? [A-LANt], where LANt is the ISO language code, '?' ends the question, '|' ends the context, and the final [A-LAN] token selects the output language. This one format lets a LoRA-finetuned AfriTeVa generator train across all four languages at once and makes the answer language a decodable choice, which is what converts cross-lingual retrieval into implicit entity alignment: an Arabic context paired with a Tigrinya question and answer forces the model to treat the Arabic phrase and the Tigrinya label as the same entity. The '(im)perfect' retriever — a heuristic that searches the head entity's Wikipedia lead for the gold tail and keeps at most two sentences containing it — generates the training contexts and doubles as an upper bound on what a deployed retriever could achieve.

What would settle it

Run mRAKL's trained generator on the test split using contexts produced by a retriever that never sees the gold tail — BM25 as in the paper, or a dense retriever that covers Tigrinya — and score each retrieved context for whether it mentions the true tail. If H@1 conditioned on the tail actually appearing in context stays far below the 59-78% the (im)perfect retriever reaches, and if overall gains remain near the observed 1-2 point BM25 level, the central accuracy claim would hold only for the oracle setup rather than the deployed pipeline.

Watch

Extended reading notes

Core claim

The paper's central claim is that the binding constraint for multilingual KGC in low-resourced languages is context, not model capacity: for Tigrinya and Amharic, structured contextual data covers at most half of the triples (49.8% description coverage for Tigrinya, 6.3% for Amharic) and contains the tail entity under 2% of the time, whereas unstructured Wikipedia text in the target or transfer languages frequently does contain the answer. mRAKL exploits this by converting each triple to a question-answer pair, tagging context and question with language tokens, and fine-tuning one AfriTeVa sequence-to-sequence model so a single generator serves Tigrinya, Amharic, Arabic, and English and can be told which language to answer in. The experiments show that context beats no context, multilingual context beats monolingual (a 4.69-point gain for Tigrinya), and cross-lingual retrieval acts as implicit entity alignment: when only the Arabic context contains the tail, the model still returns the correct Amharic or Tigrinya label. Against prior work, adding context flips the comparison — KGT5 beats mRAKL by 1.78 and 3.43 points in the no-context setting, but mRAKL beats KGT5 and its structured-context variants by 4.92 and 8.79 points once Wikipedia context is supplied.

Load-bearing premise

The headline accuracy gains come from an '(im)perfect' retriever that already knows the gold tail: it searches the head entity's own Wikipedia article for the tail and returns sentences containing it, so the context leaks the answer, while the deployed BM25 retriever, which must find the answer without being told it, produces far smaller gains.

Editorial extensions

If this is right

  • Knowledge-graph completion for low-resourced languages can be driven by unlabeled Wikipedia text: retrieval supplies the fact the model's parameters do not contain, replacing the large structured training sets prior mKGC work requires.
  • Cross-lingual entity alignment emerges implicitly from multilingual context: the generator answers in the target language even when only a transfer-language context carries the tail, so explicit alignment supervision is not needed.
  • The released resources — the 3.5k-triple Tigrinya KG, the 34k-triple Amharic KG, and question templates in four languages — give future work a benchmark for mKGC on these languages.
  • Transfer languages contribute complementary coverage: Arabic context supports region-specific and Middle East/Asia queries that English context misses and vice versa, so transfer-language choice should reflect cultural coverage, not just resource abundance.
  • An idealized retriever lifts H@1 to roughly 59-78% on queries where it finds context, indicating that the deployed retrievers, not the generator, are the limiting component of the pipeline.

Reading between the lines

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

  • The gap between the (im)perfect retriever's 59-78% H@1 and BM25's 14.15% (Tigrinya) and 35.27% (Amharic) suggests retrieval quality is the highest-leverage investment for these languages; the paper's upper-bound framing implies this, though the paper never states it as a conclusion.
  • A diagnostic the paper does not run: compute the tail-mention rate of each retriever's contexts and plot H@1 conditional on whether the context contains the tail. If tail-mention rate drives accuracy, retrieval recall becomes a cheap proxy metric for mKGC in any low-resourced language where end-to-end evaluation is expensive.
  • The zero-shot results — at most 5.83% H@1 on Amharic and 2.23% on Tigrinya even for GPT-4 — imply that parametric memory is not a viable channel for these languages; by extension, grounding through retrieval may be the default strategy for any fact-oriented task in low-resourced languages, not just knowledge-graph completion.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes mRAKL, a retrieval-augmented generation system that reformulates multilingual knowledge graph construction (mKGC) as question answering. The authors construct KGs for Tigrinya (3.5k triples) and Amharic (34k triples) from Wikidata, verbalize triples into question–answer pairs using manually prepared templates in four languages, and fine-tune AfriTeVa with LoRA as the generator. They evaluate BM25 and LaBSE retrievers against a no-context setting and also an '(im)perfect retriever' that selects context by searching the head entity's Wikipedia article for the gold tail entity. The paper reports that with this idealized retriever, mRAKL improves H@1 by 4.92 and 8.79 percentage points over prior baselines for Tigrinya and Amharic, and that BM25 retrieval improves over no-context. It also presents cross-lingual link prediction results and a bias-focused analysis of the constructed data.

Significance. The paper contributes new resources (Tigrinya and Amharic KGs, question templates, and a trained retriever/generator pipeline) and a sensible reformulation of mKGC as QA, which could be useful for other low-resourced languages. The result that BM25-based RAG improves over no-context—though modest—plus the analysis of cross-lingual context transfer are valuable. The method is described in enough detail to be reproduced in principle, with hyperparameters and retrieval settings given, though code and data are not yet released. However, the central accuracy claims are tied to an oracle retriever that leaks the answer into the retrieved context; the real deployed pipeline's gains are much smaller. These claims need to be reframed and the arithmetic corrected before publication.

major comments (3)
  1. [Abstract; §3.1; §4.3; Table 5] The headline improvements of 4.92 and 8.79 percentage points are computed with the '(im)perfect retriever' of §3.1, which selects context by searching the head entity's Wikipedia article for the gold tail entity and keeping up to two sentences that contain it. This leaks the answer into the retrieved context, so these numbers describe an oracle upper bound rather than the deployed system. With the actual BM25 retriever, Table 5 shows gains of only 1.18 points for Tigrinya (H@1 from 12.97 to 14.15) and 2.15 points for Amharic (H@1 from 33.12 to 35.27). The Abstract states these gains as the headline result (even with the qualifier 'with an idealized retrieval system'), and the Conclusion restates them without any qualifier. I recommend reporting the BM25-based gains as the primary accuracy result and presenting the (im)perfect-retriever numbers strictly as an upper-bound ablation in all summaries, including the Abstract and Conclusion.
  2. [§4.3, Table 4, footnote 11] The 8.79-point Amharic gain is computed as 41.37 − 32.58, i.e., against KGT5-No-Context, not against the best prior baseline. The best KGT5 variant for Amharic is KGT5-Description with H@1=32.91 (Table 4), so under the paper's own definition in footnote 11 ('best-performing methods for each work'), the correct gain is 8.46 points. This 0.33-point overstatement propagates to the Abstract and Conclusion and should be corrected.
  3. [§4.1, §4.4, Table 5] All results are reported from a single inference run (§4), while the key BM25-versus-no-context differences in Table 5 are small (1.18 and 2.15 points for Tigrinya and Amharic H@1). Without multiple seeds, confidence intervals, or significance tests, these differences may not be stable. At minimum the paper should acknowledge this uncertainty, and ideally report variance across runs or a significance test (e.g., a paired bootstrap) to support the claim that BM25 retrieval improves over no-context.
minor comments (6)
  1. [Appendix A.2] The sentence 'In this section, we provide figures and graphs to support the results reported in Sec.' contains an empty cross-reference; it should point to the relevant results section (e.g., §4.4 or §4.5).
  2. [Table 8 caption; §B.1] The caption of Table 8 spells 'Hyperparameters' as 'Hyperparamters', and §B.1 contains the typo 'teh training data point' for 'the training data point'.
  3. [§3.2 input representation] The set notation '∀ tϵ{T igrinya, Amharic, English, Arabic}' has a stray space in 'T igrinya' and should use the proper membership symbol '∈'; the same issue appears in the second occurrence of this formula.
  4. [Table 3 and throughout] The model name is spelled 'AfriTeVa' in most of the paper but 'AfriTeva' in Table 3 and in the surrounding text of §4.2; please unify the spelling.
  5. [References] The reference to Vrandečić and Krötzsch (2014) appears twice with slightly different formatting; the duplicate should be removed.
  6. [§1 Languages of Study] The term 'Afro-Semitic languages' is not standard; consider 'Afro-Asiatic (Semitic) languages' or simply 'Semitic languages'.

Circularity Check

1 steps flagged · score 6.0 of 10

The headline accuracy gains (4.92/8.79 points) are computed with an oracle retriever that searches for the gold tail entity and injects it into the context; the deployable BM25 retriever yields only about 1.2/2.2 point gains.

  1. self definitional [Section 3.1 (Extracting Context), with numbers used in Abstract, Section 4.3 (Table 4), and Conclusion]
    "We extracted the context for each question by searching for the tail entity in the first paragraph of the Tigrinya Wikipedia article associated with the head entity. We then kept a maximum of two sentences that had the tail entity as context."

    The (im)perfect retriever constructs its retrieved context by searching for the gold tail entity and keeping the sentences that contain it, so the answer is part of the retrieval input by construction. The 'Self-Context' results in Table 4 — the 4.92 and 8.79 percentage point improvements repeated in the Abstract, Section 4, and Conclusion — are computed under this setup. The improvement over the no-context setting therefore measures how well the generator copies an answer that was inserted into the context using the test labels, not how well a deployable retriever supports knowledge graph construction. The real BM25 retriever, which does not see the tail at query time, gives only +1.18 (Tigrinya) and +2.15 (Amharic) points over no-context in Table 5.

full rationale

The paper's central accuracy claim partially reduces to answer leakage. The (im)perfect retriever described in Section 3.1 is defined by searching for the gold tail entity in the head entity's Wikipedia article and returning at most two sentences containing that tail. When the resulting context is fed to the generator, the answer is already present in the context by construction, so the reported gains over the no-context setting in Table 4 and the Abstract (4.92 Tigrinya, 8.79 Amharic) are upper-bound numbers rather than evidence for the deployed retrieval pipeline. The paper is transparent about this: it calls the setup '(im)perfect' and says it 'provides an upper bound of performance.' However, the Abstract and especially the Conclusion present the 4.92/8.79 numbers as the headline result without consistently carrying the 'idealized retriever' qualifier, which makes the central claim partially circular. The real, non-leaky RAG pipeline using BM25 is evaluated separately in Table 5 and shows modest but genuine gains (about 1.18 and 2.15 points in H@1), so the derivation is not wholly circular. There are no load-bearing self-citations or imported uniqueness theorems; the LaBSE retriever is also trained on the oracle contexts, as described in Appendix B.1, but that is not the headline result. Overall, the score reflects one central 'prediction' that reduces by construction to the gold answer being placed in the context, while the actually deployable system's evaluation remains independent.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests on a small number of hand-chosen parameters: the context sentence limit and the relation set size. The assumptions are domain-level facts about Wikipedia co-occurrence, Wikidata label alignment, and the validity of the QA reformulation. No new physical or abstract entities are postulated.

free parameters (2)
  • context sentence limit = max two sentences
    The (im)perfect retriever keeps at most two sentences containing the tail entity; this cutoff is chosen by hand and directly affects the oracle upper-bound performance reported in the headline results (Section 3.1).
  • relations set size = 120 relations
    The final relation set (96 Tigrinya relations plus 24 translated from Amharic) determines the KG structure and the question templates; this size is a resource-driven choice, not derived from a principled criterion (Section 3.1).
assumptions (4)
  • domain assumption The tail entity is frequently mentioned in the head entity's Wikipedia article in at least one language.
    The (im)perfect retriever and, more generally, the RAG benefit rely on the co-occurrence of head and tail entities in the same article. This is stated in Section 3.1 where context is extracted by searching for the tail in the head's article.
  • domain assumption Wikidata labels in Arabic, English, Amharic, and Tigrinya accurately represent the same entities and relations.
    KG extraction and cross-lingual transfer assume that the labels in different languages refer to the same underlying entity or property. This is used throughout Section 3.1 for extracting triples and constructing questions.
  • domain assumption Template-based question reformulation preserves the semantics of the original (head, relation, tail) triple.
    The QA reformulation assumes that a question built from the head and relation is answerable by the tail. This is the basis of the whole method (Section 3.1, Template-Based Reformulation).
  • domain assumption Exact or substring matching between the predicted text and the tail label is a valid measure of correctness.
    The evaluation counts a prediction as correct if it contains the tail entity as a substring (Section 4.2 and Appendix D). This lenient matching rule is not standard exact-match evaluation and can inflate scores.

how reviews work

0 comments
Cite this review

Pith. "Pith review of mRAKL: Multilingual Retrieval-Augmented Knowledge Graph Construction for Low-Resourced Languages." pith.science (2026). https://pith.science/paper/RXDRCWE7

@misc{pith2026250716011,
  author       = {Pith},
  title        = {Pith review of: mRAKL: Multilingual Retrieval-Augmented Knowledge Graph Construction for Low-Resourced Languages},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RXDRCWE7}},
  note         = {Machine review of arXiv:2507.16011}
}
read the original abstract

Knowledge Graphs represent real-world entities and the relationships between them. Multilingual Knowledge Graph Construction (mKGC) refers to the task of automatically constructing or predicting missing entities and links for knowledge graphs in a multilingual setting. In this work, we reformulate the mKGC task as a Question Answering (QA) task and introduce mRAKL: a Retrieval-Augmented Generation (RAG) based system to perform mKGC. We achieve this by using the head entity and linking relation in a question, and having our model predict the tail entity as an answer. Our experiments focus primarily on two low-resourced languages: Tigrinya and Amharic. We experiment with using higher-resourced languages Arabic and English for cross-lingual transfer. With a BM25 retriever, we find that the RAG-based approach improves performance over a no-context setting. Further, our ablation studies show that with an idealized retrieval system, mRAKL improves accuracy by 4.92 and 8.79 percentage points for Tigrinya and Amharic, respectively.

Figures

Figures reproduced from arXiv: 2507.16011 by the authors.

Figure 1
Figure 1. Inference with mRAKL In this example our triple is (Surafel Dagnachew, place of birth, Ethiopia). Taking the reformulated question “What is Surafel Dagnachew’s place of birth?”, the retriever encodes the query and fetches sentences with the highest similarity from the passages available from the Wikipedia articles in each of the languages. These sentences are then passed to the generator as context along with the qu… view at source ↗
Figure 2
Figure 2. Percentage of triples with the given relation [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 6
Figure 6. Context from Arabic contains the tail entity [PITH_FULL_IMAGE:figures/full_fig_p014_6.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Effects of context from different languages on للة [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: Effects of context from different languages on [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 7
Figure 7. Figure 7: Example where the BM25 and Heuristic re [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]
Figure 8
Figure 8. Figure 8: Reformulating triples into question-answer pairs. In each step depicted above, we highlight the head entity in red, the relation in blue, and the tail entity in orange. In Step 1, we start with a triple that has entity IDs and Property ID extracted from Wikidata. In St…
Figure 9
Figure 9. Figure 9: Figure showing top 10 head and tail entities in Amharic KG. English translations for entities are provided [PITH_FULL_IMAGE:figures/full_fig_p017_9.png]
Figure 10
Figure 10. Figure 10: Figure showing top 10 head and tail entities in Tigrinya KG. English translations for entities are provided [PITH_FULL_IMAGE:figures/full_fig_p017_10.png]
Figure 11
Figure 11. Figure 11: Tail entities that are spelled the same in [PITH_FULL_IMAGE:figures/full_fig_p018_11.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

57 extracted references · 31 canonical work pages

  1. [1]

    Akari Asai, Sewon Min, Zexuan Zhong, and Danqi Chen. 2023. Acl 2023 tutorial: Retrieval-based language models and applications. ACL 2023

    work page 2023

  2. [2]

    Shaik Johny Basha, Duggineni Veeraiah, Boddu Venkat Charan, Wiltrud Sahithi Joyce Yeddu, and Devalla Ganesh Babu. 2023. https://doi.org/10.1109/ICICT57646.2023.10134103 Detection and comparative analysis of handwritten words of amharic language to english using cnn-based frameworks . In 2023 International Conference on Inventive Computation Technologies (...

    work page arXiv 2023

  3. [3]

    Antoine Bosselut, Hannah Rashkin, Maarten Sap, Chaitanya Malaviya, Asli Celikyilmaz, and Yejin Choi. 2019. Comet: Commonsense transformers for automatic knowledge graph construction. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4762--4779

    work page 2019

  4. [4]

    Andrew Carlson, Justin Betteridge, Bryan Kisiel, Burr Settles, Estevam Hruschka, and Tom Mitchell. 2010. Toward an architecture for never-ending language learning. In Proceedings of the AAAI conference on artificial intelligence, volume 24, pages 1306--1313

    work page 2010

  5. [5]

    Soumen Chakrabarti, Harkanwar Singh, Shubham Lohiya, Prachi Jain, et al. 2022. Joint completion and alignment of multilingual knowledge graphs. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, page 11922—11938

    work page 2022

  6. [6]

    Muhao Chen, Weijia Shi, Ben Zhou, and Dan Roth. 2021. Cross-lingual entity alignment with incidental supervision. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 645--658

    work page 2021

  7. [7]

    Muhao Chen, Yingtao Tian, Mohan Yang, and Carlo Zaniolo. 2017. Multilingual knowledge graph embeddings for cross-lingual knowledge alignment. In Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence. International Joint Conferences on Artificial Intelligence Organization

    work page 2017

  8. [8]

    Xuelu Chen, Muhao Chen, Changjun Fan, Ankith Uppunda, Yizhou Sun, and Carlo Zaniolo. 2020. https://doi.org/10.18653/v1/2020.findings-emnlp.290 Multilingual knowledge graph completion via ensemble knowledge transfer . In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 3227--3238, Online. Association for Computational Linguistics

    doi:10.18653/v1/2020.findings-emnlp.290 2020

Show all 57 references
  1. [9]

    Xin Dong, Evgeniy Gabrilovich, Geremy Heitz, Wilko Horn, Ni Lao, Kevin Murphy, Thomas Strohmann, Shaohua Sun, and Wei Zhang. 2014. Knowledge vault: A web-scale approach to probabilistic knowledge fusion. In Proceedings of the 20th ACM SIGKDD international conference on Knowled...

    2014

  2. [10]

    Fangxiaoyu Feng, Yinfei Yang, Daniel Cer, Naveen Arivazhagan, and Wei Wang. 2022. https://doi.org/10.18653/v1/2022.acl-long.62 Language-agnostic BERT sentence embedding . In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long...

    2022 doi

  3. [11]

    https://dumps.wikimedia.org Wikimedia downloads

    Wikimedia Foundation. https://dumps.wikimedia.org Wikimedia downloads

  4. [12]

    Fitsum Gaim, Wonsuk Yang, Hancheol Park, and Jong Park. 2023. https://doi.org/10.18653/v1/2023.acl-long.661 Question-answering in a low-resourced language: Benchmark dataset and models for T igrinya . In Proceedings of the 61st Annual Meeting of the Association for Computation...

    2023 doi

  5. [13]

    Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yi Dai, Jiawei Sun, and Haofen Wang. 2023. Retrieval-augmented generation for large language models: A survey. arXiv preprint arXiv:2312.10997

    2023 arXiv

  6. [14]

    Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat, and Ming-Wei Chang. 2020. Realm: retrieval-augmented language model pre-training. In Proceedings of the 37th International Conference on Machine Learning, pages 3929--3938

    2020

  7. [15]

    Johannes Hoffart, Fabian M Suchanek, Klaus Berberich, and Gerhard Weikum. 2013. Yago2: A spatially and temporally enhanced knowledge base from wikipedia. Artificial intelligence, 194:28--61

    2013

  8. [16]

    Edward J Hu, yelong shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. https://openreview.net/forum?id=nZeVKeeFYf9 Lo RA : Low-rank adaptation of large language models . In International Conference on Learning Representations

    2022

  9. [17]

    Xiao Huang, Jingyuan Zhang, Dingcheng Li, and Ping Li. 2019. https://doi.org/10.1145/3289600.3290956 Knowledge graph embedding based question answering . In Proceedings of the Twelfth ACM International Conference on Web Search and Data Mining, WSDM '19, page 105–113, New York,...

    2019

  10. [18]

    Jinhao Jiang, Kun Zhou, Xin Zhao, and Ji-Rong Wen. 2023. Unikgqa: Unified retrieval and reasoning for solving multi-hop question answering over knowledge graph. In The Eleventh International Conference on Learning Representations

    2023

  11. [19]

    Zhengbao Jiang, Antonios Anastasopoulos, Jun Araki, Haibo Ding, and Graham Neubig. 2020. X-factr: Multilingual factual knowledge retrieval from pretrained language models. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages ...

    2020

  12. [20]

    Pratik Joshi, Sebastin Santy, Amar Budhiraja, Kalika Bali, and Monojit Choudhury. 2020. https://doi.org/10.18653/v1/2020.acl-main.560 The state and fate of linguistic diversity and inclusion in the NLP world . In Proceedings of the 58th Annual Meeting of the Association for Co...

    2020 doi

  13. [21]

    Odunayo Jude Ogundepo, Akintunde Oladipo, Mofetoluwa Adeyemi, Kelechi Ogueji, and Jimmy Lin. 2022. https://doi.org/10.18653/v1/2022.deeplo-1.14 A fri T e VA : Extending ?small data? pretraining approaches to sequence-to-sequence models . In Proceedings of the Third Workshop on...

    2022 doi

  14. [22]

    Nora Kassner, Philipp Dufter, and Hinrich Sch \"u tze. 2021. Multilingual lama: Investigating knowledge in multilingual pretrained language models. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, page...

    2021

  15. [23]

    Adrian Kochsiek, Apoorv Saxena, Inderjeet Nair, and Rainer Gemulla. 2023. https://doi.org/10.18653/v1/2023.repl4nlp-1.11 Friendly neighbors: Contextualized sequence-to-sequence link prediction . In Proceedings of the 8th Workshop on Representation Learning for NLP (RepL4NLP 20...

    2023 doi

  16. [24]

    Jens Lehmann, Robert Isele, Max Jakob, Anja Jentzsch, Dimitris Kontokostas, Pablo N Mendes, Sebastian Hellmann, Mohamed Morsey, Patrick Van Kleef, S \"o ren Auer, et al. 2015. Dbpedia--a large-scale, multilingual knowledge base extracted from wikipedia. Semantic web, 6(2):167--195

    2015

  17. [25]

    u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt \

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich K \"u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt \"a schel, et al. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Proc...

    2020

  18. [26]

    Xing Han Lù. 2024. http://arxiv.org/abs/2407.03618 Bm25s: Orders of magnitude faster lexical search via eager sparse scoring

    2024 arXiv

  19. [27]

    Hellina Hailu Nigatu, John Canny, and Sarah E. Chasins. 2024. https://doi.org/10.1145/3613904.3642605 Low-resourced languages and online knowledge repositories: A need-finding study. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems, CHI '24, New ...

    2024

  20. [28]

    Tolulope Ogunremi, Dan Jurafsky, and Christopher Manning. 2023. https://doi.org/10.18653/v1/2023.findings-eacl.93 Mini but mighty: Efficient multilingual pretraining with linguistically-informed data selection . In Findings of the Association for Computational Linguistics: EAC...

    2023 doi

  21. [29]

    Jessica Ojo, Kelechi Ogueji, Pontus Stenetorp, and David Ifeoluwa Adelani. 2024. http://arxiv.org/abs/2311.07978 How good are large language models on african languages?

    2024 arXiv

  22. [30]

    Akintunde Oladipo, Mofetoluwa Adeyemi, Orevaoghene Ahia, Abraham Toluwase Owodunni, Odunayo Ogundepo, David Ifeoluwa Adelani, and Jimmy Lin. 2023. https://openreview.net/forum?id=ybc9V6Cbq2 Better quality pre-training data and t5 models for african languages . In The 2023 Conf...

    2023

  23. [31]

    OpenAI, Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, Red Avila, Igor Babuschkin, Suchir Balaji, Valerie Balcom, Paul Baltescu, Haiming Bao, Mohammad Bavarian, Jeff ...

    2024 arXiv

  24. [32]

    Shirui Pan, Linhao Luo, Yufei Wang, Chen Chen, Jiapu Wang, and Xindong Wu. 2024. Unifying large language models and knowledge graphs: A roadmap. IEEE Transactions on Knowledge and Data Engineering (TKDE)

    2024

  25. [33]

    Heiko Paulheim. 2018. https://ceur-ws.org/Vol-2180/ISWC_2018_Outrageous_Ideas_paper_10.pdf How much is a triple? estimating the cost of knowledge graph creation

    2018

  26. [34]

    Fabio Petroni, Tim Rockt \"a schel, Sebastian Riedel, Patrick Lewis, Anton Bakhtin, Yuxiang Wu, and Alexander Miller. 2019. https://doi.org/10.18653/v1/D19-1250 Language models as knowledge bases? In Proceedings of the 2019 Conference on Empirical Methods in Natural Language P...

    2019 doi

  27. [35]

    Nils Reimers and Iryna Gurevych. 2019. https://arxiv.org/abs/1908.10084 Sentence-bert: Sentence embeddings using siamese bert-networks . In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics

    2019 arXiv

  28. [36]

    Ridho Reinanda, Edgar Meij, and Maarten de Rijke. 2020. https://doi.org/10.1561/1500000063 Knowledge graphs: An information retrieval perspective . Found. Trends Inf. Retr., 14(4):289–444

    2020 doi

  29. [37]

    Stephen Robertson and Hugo Zaragoza. 2009. https://doi.org/10.1561/1500000019 The probabilistic relevance framework: Bm25 and beyond . Foundations and Trends® in Information Retrieval, 3(4):333--389

    2009 doi

  30. [38]

    Farhan Samir, Chan Young Park, Anjalie Field, Vered Shwartz, and Yulia Tsvetkov. 2024. https://doi.org/10.18653/v1/2024.emnlp-main.384 Locating information gaps and narrative inconsistencies across languages: A case study of LGBT people portrayals on W ikipedia . In Proceeding...

    2024 doi

  31. [39]

    Apoorv Saxena, Adrian Kochsiek, and Rainer Gemulla. 2022 a . Sequence-to-sequence knowledge graph completion and question answering. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2814--2828

    2022

  32. [40]

    Apoorv Saxena, Adrian Kochsiek, and Rainer Gemulla. 2022 b . https://doi.org/10.18653/v1/2022.acl-long.201 Sequence-to-sequence knowledge graph completion and question answering . In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volum...

    2022 doi

  33. [41]

    Ran Song, Shizhu He, Shengxiang Gao, Li Cai, Kang Liu, YU Zhengtao, and Jun Zhao. 2023. Multilingual knowledge graph completion from pretrained language models with knowledge constraints. In The 61st Annual Meeting Of The Association For Computational Linguistics

    2023

  34. [42]

    Zequn Sun, Chengming Wang, Wei Hu, Muhao Chen, Jian Dai, Wei Zhang, and Yuzhong Qu. 2020. Knowledge graph alignment network with gated multi-hop neighborhood aggregation. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pages 222--229

    2020

  35. [43]

    Shiyu Tian, Yangyang Luo, Tianze Xu, Caixia Yuan, Huixing Jiang, Chen Wei, and Xiaojie Wang. 2024. Kg-adapter: Enabling knowledge graph integration in large language models through parameter-efficient fine-tuning. In Findings of the Association for Computational Linguistics AC...

    2024

  36. [44]

    Ahmet \"U st \"u n, Viraat Aryabumi, Zheng Yong, Wei-Yin Ko, Daniel D ' souza, Gbemileke Onilude, Neel Bhandari, Shivalika Singh, Hui-Lee Ooi, Amr Kayid, Freddie Vargus, Phil Blunsom, Shayne Longpre, Niklas Muennighoff, Marzieh Fadaee, Julia Kreutzer, and Sara Hooker. 2024. ht...

    2024 doi

  37. [45]

    Denny Vrande c i \'c and Markus Kr \"o tzsch. 2014. Wikidata: a free collaborative knowledgebase. Communications of the ACM, 57(10):78--85

    2014

  38. [46]

    Denny Vrande c i\' c and Markus Kr\" o tzsch. 2014. https://doi.org/10.1145/2629489 Wikidata: a free collaborative knowledgebase . Commun. ACM, 57(10):78–85

    2014 doi

  39. [47]

    Xiaozhi Wang, Tianyu Gao, Zhaocheng Zhu, Zhengyan Zhang, Zhiyuan Liu, Juanzi Li, and Jian Tang. 2021. https://doi.org/10.1162/tacl_a_00360 KEPLER : A unified model for knowledge embedding and pre-trained language representation . Transactions of the Association for Computation...

    2021 doi

  40. [48]

    Yuxiang Wu, Yu Zhao, Baotian Hu, Pasquale Minervini, Pontus Stenetorp, and Sebastian Riedel. 2022. An efficient memory-augmented transformer for knowledge-intensive nlp tasks. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 5184--5196

    2022

  41. [49]

    Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, and Colin Raffel. 2021. https://doi.org/10.18653/v1/2021.naacl-main.41 m T 5: A massively multilingual pre-trained text-to-text transformer . In Proceedings of the 2021 Conferenc...

    2021 doi

  42. [50]

    Linyao Yang, Hongyang Chen, Zhao Li, Xiao Ding, and Xindong Wu. 2024. Give us the facts: Enhancing large language models with knowledge graphs for fact-aware language modeling. IEEE Transactions on Knowledge and Data Engineering

    2024

  43. [51]

    Liang Yao, Chengsheng Mao, and Yuan Luo. 2019. http://arxiv.org/abs/1909.03193 Kg-bert: Bert for knowledge graph completion

    2019 arXiv

  44. [52]

    Seid Muhie Yimam, Hizkiel Mitiku Alemayehu, Abinew Ayele, and Chris Biemann. 2020. https://doi.org/10.18653/v1/2020.coling-main.91 Exploring A mharic sentiment analysis from social media texts: Building annotation tools and classification models . In Proceedings of the 28th In...

    2020 doi

  45. [53]

    Haeun Yu, Pepa Atanasova, and Isabelle Augenstein. 2024. https://doi.org/10.18653/v1/2024.acl-long.444 Revealing the parametric knowledge of language models: A unified framework for attribution methods . In Proceedings of the 62nd Annual Meeting of the Association for Computat...

    2024 doi

  46. [54]

    Lingfeng Zhong, Jia Wu, Qian Li, Hao Peng, and Xindong Wu. 2023. A comprehensive survey on automatic knowledge graph construction. ACM Computing Surveys, 56(4):1--62

    2023

  47. [55]

    Wenxuan Zhou, Fangyu Liu, Ivan Vuli \'c , Nigel Collier, and Muhao Chen. 2022. https://doi.org/10.18653/v1/2022.acl-long.371 Prix- LM : Pretraining for multilingual knowledge base construction . In Proceedings of the 60th Annual Meeting of the Association for Computational Lin...

    2022 doi

  48. [56]

    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...

  49. [57]

    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 6, 2026 · model on record in the stance chip above.