Pith. sign in

REVIEW 3 major objections 6 minor 36 references

LLM-based Discriminative Reasoning for Knowledge Graph Question Answering

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Reformulating knowledge graph QA into three discriminative subtasks (search, prune, answer) where an LLM picks from option pools instead of generating free text yields state-of-the-art Hits@1/F1 of 0.840/0.845 on WebQSP and 0.802/0.820 on…

desk verdict READS is a genuinely different discriminative decomposition for KGQA, and the ablations hold up; the only thing that keeps me from trusting the SOTA numbers is the ambiguous SPARQL-expansion step in Appendix B. read the letter →

arxiv 2412.12643 v2 pith:QB34DTFM submitted 2024-12-17 cs.CL

Mufan Xu , Kehai Chen , Xuefeng Bai , Muyun Yang , Tiejun Zhao , Min Zhang This is my paper · ORCID
classification cs.CL
keywords knowledgegraphquestionansweringdiscriminativereasoninglargelanguagemodelshallucinationmitigationsubgraphretrievalanswerinferenceWebQSPCWQ
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

This paper tries to establish that the hallucination and ungrounded reasoning of large language models on knowledge graph question answering can be reduced by converting the task into three discriminative subtasks rather than an open-ended generation problem. Given a question, the LLM first chooses which neighboring relation to expand in the question-related subgraph, then selects which constraints in the question apply where, then picks the node position that holds the answer. The claim is that constraining the model to choose from option pools at each step grounds its reasoning in the retrieved graph. If correct, this gives a recipe for making 7-billion-parameter LLMs reliable on knowledge graph QA without relying on their internal knowledge or toolboxes.

What carries the argument

The central object is the abstract subgraph structure that groups entities in the retrieved subgraph into semantic nodes by structural position while keeping relation names, optionally annotated with entity types (entity, topic, num, date). READS performs all three decisions on this skeleton: choose the next (node, relation) pair to expand, choose constraints as (position, operator, target) triples, and choose the answer position. This lets a 7-billion-parameter LLM reason over a small typed structure instead of raw triples, and the discriminative option-pool formulation with logit-constrained beam search is what carries the argument.

What would settle it

Take WebQSP and CWQ, and for a random subset of test questions either corrupt or delete the pre-linked starting entity before READS runs; compare Hits@1 and F1 to the reported 0.840/0.845 and 0.802/0.820. If scores do not fall well below those numbers, the entity-linking assumption is not doing the work; if they fall sharply, the headline results depend on the assumed linker rather than on the discriminative subtasks alone.

Watch

Extended reading notes

Core claim

READS decomposes knowledge graph question answering into graph searching, graph pruning, and answer inference, and casts each as a multiple-choice selection over an option pool. The next triple to expand is chosen by constrained beam search over the LLM's output logits; constraints are mapped to positions with one of seven operators (=, <, ≤, >, ≥, min, max); and the answer is located as a semantic node in the abstract structure, with every entity grouped at that node returned as an answer. On WebQSP and CWQ the method reports Hits@1/F1 of 0.840/0.845 and 0.802/0.820, exceeding the reproduced generative baseline and the interactive baselines, with the largest margin on the harder CWQ benchmark. The authors interpret this as evidence that discriminative subtask decomposition alleviates ungrounded subgraph planning and answer hallucination.

Load-bearing premise

The load-bearing premise is that question entities have already been correctly linked to Freebase entities by rule-based recognition before READS runs (with gold SPARQL queries available to build the training data), because a wrong starting entity or missing gold query cannot be repaired by any later subtask.

Editorial extensions

If this is right

  • On WebQSP and CWQ, the method attains state-of-the-art Hits@1 and F1 of 0.840/0.845 and 0.802/0.820, surpassing generative and interactive baselines.
  • All three subtasks contribute to the performance: replacing any one with a generative counterpart lowers Hits@1, and the search subtask is the largest single source of error.
  • Entity-type labels on semantic nodes are load-bearing, since removing them drops Hits@1 on both benchmarks.
  • The decomposition cuts average input tokens per request roughly in half relative to the generative baseline while keeping a similar number of model calls.
  • On the GrailQA development set, the method improves Hits@1 over the single-task discriminative baseline on i.i.d. and compositional splits, though not on the zero-shot split.

Reading between the lines

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

  • A testable extension is to run READS with a learned entity linker instead of the assumed rule-based linking; the reported scores probably bound what is achievable when linking is imperfect.
  • Because the framework does not use the model's internal knowledge and transfers across several backbones, it may generalize to other deterministic graph tasks such as database query selection or document-graph retrieval, though the paper does not test them.
  • The paper's data-efficiency curves suggest about 25% of the training data is enough to reach near-best performance, so the gold-SPARQL dependency could be relaxed with a smaller set of pseudo-labels from a strong model without losing much accuracy.
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.

Referee Report

3 major / 6 minor

Summary. The paper proposes READS, a KGQA framework that replaces open-ended LLM generation for subgraph planning and answer reasoning with three discriminative subtasks: (i) iterative subgraph searching in which the LLM selects the next (entity, relation) triple from an option pool built from the current abstract subgraph structure; (ii) subgraph pruning in which constraint entities are mapped to positions in the abstract structure with comparison operators; and (iii) answer inference in which the LLM selects a semantic node whose entity group forms the answer. The framework is trained on samples derived from gold SPARQL queries in the WebQSP and CWQ training sets and evaluated on those benchmarks (plus a GrailQA dev-set experiment), reporting Hits@1/F1 of 0.840/0.845 on WebQSP and 0.802/0.820 on CWQ, above the listed generative and interactive baselines. The paper also reports search-quality metrics (relation recall and minimum graph edit distance), ablations, error-type frequencies, data-efficiency curves, and a base-model universality table.

Significance. Conditional on a clean evaluation, the paper makes a valuable contribution: it demonstrates that reformulating KGQA as constrained discriminative selections can outperform generative and interactive KGQA methods with a 7B-scale model, and it offers a concrete mechanism for the improvement (reduced ungrounded subgraph and reasoning generation). The ablations in Table 3 and the error-type analysis in Table 11 are consistent with that mechanism, and Table 2 provides a useful decomposition of search quality. The data-efficiency and model-universality analyses are also informative. The main caveat is that the headline state-of-the-art numbers depend on the exact construction of the test-time search space, which is not currently described unambiguously in Appendix B; the entity-linking assumption also needs quantification. The paper does not yet release code or data, so its reproducibility strength cannot currently be verified.

major comments (3)
  1. [Appendix B, Freebase preprocessing] The manuscript does not state whether the benchmark-SPARQL subgraph expansion in Appendix B is applied only to training questions or also to test questions. The passage says the authors 'extract subgraphs from Freebase using breadth-first search for each question' and then 'expand these subgraphs using the SPARQL queries provided in the benchmarks to ensure the presence of constraint branches,' adding these to 'the original dataset.' If test-time subgraphs are also expanded from gold SPARQL, the 'searching' subtask in Equation (6) is not searching an unbiased graph; the candidate option pool already contains branches derived from the gold reasoning structure. That would inflate the relation-recall and graph-edit-distance numbers in Table 2 and the end-to-end Hits@1/F1 in Table 1, and it would make the comparison to RoG, ToG, and KG-Agent unfair. Please state explicitly that the SPARQL-query expansion is train-only, or, if it is not, rerun the test evaluation without test-time expansion and report both settings.
  2. [Section 3.2; Limitations item 1] Section 3.2 assumes the question entities are already linked to Freebase entities through rule-based recognition, and Limitations item 1 acknowledges this, but the paper never reports which linker or rules are used or how often they succeed on WebQSP and CWQ. Because every later subtree is rooted at the linked starting entity, a wrong linker output cannot be repaired by any subsequent subtask. Please name the linker, report its coverage on the two benchmarks, and quantify the fraction of test questions whose starting node is incorrect; without this, the comparison against methods that do not assume such a root is not fully interpretable.
  3. [Table 1 and Appendix C] The WebQSP state-of-the-art claim relies on discounting the original RoG Hits@1 of 0.857 as miscalculated, but the manuscript only points to an external author response. Please provide the corrected calculation in-paper using the Appendix C formula, and confirm that the same evaluation script, answer normalization, and entity linking were used for every row in Table 1. At a minimum, report variance or exact reproducibility of the reproduced RoG run (0.795), since a 4-point gap relative to the original paper is too large to leave undocumented.
minor comments (6)
  1. [Appendix C] The text calls the F1 score 'Macro-F1,' but the described procedure (averaging per-sample precision/recall and then taking the harmonic mean of the aggregate precision and recall) is not the usual macro-F1; please rename it or clarify the aggregation.
  2. [Appendix G] The text says 'see Figure 12' but the statistic is presented in Table 12; please fix the cross-reference.
  3. [Table 4] The 'READS' row for GPT-4 is described as not fine-tuned or constrained in the same way as the Llama2-7b READS model; please label it 'READS (zero-shot selection)' to avoid the impression that the same READS procedure is used.
  4. [Table 1 and Section 5.2] No confidence intervals or repeated-seed statistics are reported for the headline numbers; given the small differences to some baselines, a brief statement of run-to-run variability (or that the pipeline is deterministic) would improve confidence.
  5. [Figure 4] The caption does not specify what the two curves represent (generative reasoning with ToG-style in-context prompting versus READS discriminative selection) or how the averages are computed over test questions; please expand the caption.
  6. [Table 11 and Appendix F] The column headers 'Total' and 'Seperate' should be corrected, and the interpretation of those columns (per-dataset totals versus per-error-type counts) should be stated explicitly.

Circularity Check

1 steps flagged · score 6.0 of 10

Gold-SPARQL expansion of search subgraphs makes the central comparison partially circular; the derivations themselves are otherwise non-circular.

  1. self definitional [Appendix B, 'Freebase preprocessing'; Section 5.3, Eq. (9); Limitations item 2]
    "Additionally, we expand these subgraphs using the SPARQL queries provided in the benchmarks to ensure the presence of constraint branches. ... Therefore, the paths included in SPARQL effectively represent the correct subgraph structure required to answer the current question."

    The graph-searching subtask (Eq. 6) selects the next triple from an option pool. Appendix B constructs that pool by BFS plus expansion from the benchmark gold SPARQL queries, and the same SPARQL queries define Ggold used in Eq. (9) for the relation-recall metric and provide the answer labels.

full rationale

The claimed derivation chain is mostly self-contained: Equations (6)-(8) are argmax rules over explicit option pools; the model is fine-tuned on labels derived from gold SPARQL, which is standard supervised learning and not circular. The decomposition into search/prune/answer is an architectural choice, not a renamed known result, and the main baselines (RoG, ToG, KG-Agent, PANGU) are external and independently published. The one serious circular element is the Freebase preprocessing in Appendix B: the candidate subgraphs used for the search process are expanded using the benchmark gold SPARQL queries, while the same queries define the golden subgraph used to compute relation recall (Eq. 9) and contain the answer labels. Because the paper never limits this expansion to training data, the reported Hits@1/F1 and search-quality improvements can be partially attributed to gold-query-informed search inputs rather than to the discriminative strategy itself. This is a partial, evaluation-level circularity; it does not make the subtask equations themselves circular. Score 6 reflects that one central 'prediction' (the retrieved subgraph) is constructed with its own target.

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

The central claim rests on supervised training data derived from gold SPARQL queries and on a pre-linked entity assumption. No numeric free parameters are fitted to the test sets; the reported gains come from the choice of subtask decomposition and training procedure. Semantic nodes are borrowed from UniKGQA. The main uncharged input is the gold-query supervision, which is strong but limits the method to benchmarks with query annotations.

assumptions (4)
  • domain assumption Question entities are already linked to Freebase entities via rule-based recognition before READS begins.
    Stated in Section 3.2 and Limitations item 1; the search root and constraint mapping both depend on these links. READS does not perform or evaluate entity linking.
  • domain assumption Gold SPARQL queries for WebQSP and CWQ are available and are used to construct training data for all three subtasks.
    Appendix B builds searching, pruning, and answer inference training samples from SPARQL queries; Limitations item 2 acknowledges the dependence on a gold query. This restricts applicability to benchmarks with query annotations.
  • domain assumption Freebase subgraphs extracted with breadth-first search plus SPARQL-based expansion contain all branches needed for correct answers.
    Appendix B describes this preprocessing and refers to open-source code for implementation; if the extracted candidate subgraph omits a needed relation, search cannot retrieve it.
  • domain assumption Semantic node grouping from UniKGQA preserves the information needed for answer position inference.
    Section 3.1 adopts semantic nodes to group entities by structural position; the answer strategy then selects positions rather than entities, so grouping errors would propagate to answers.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LLM-based Discriminative Reasoning for Knowledge Graph Question Answering." pith.science (2026). https://pith.science/paper/QB34DTFM

@misc{pith2026241212643,
  author       = {Pith},
  title        = {Pith review of: LLM-based Discriminative Reasoning for Knowledge Graph Question Answering},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QB34DTFM}},
  note         = {Machine review of arXiv:2412.12643}
}
read the original abstract

Large language models (LLMs) based on generative pre-trained Transformer have achieved remarkable performance on knowledge graph question-answering (KGQA) tasks. However, LLMs often produce ungrounded subgraph planning or reasoning results in KGQA due to the hallucinatory behavior brought by the generative paradigm. To tackle this issue, we propose READS to reformulate the KGQA process into discriminative subtasks, which simplifies the search space for each subtasks. Based on the subtasks, we design a new corresponding discriminative inference strategy to conduct the reasoning for KGQA, thereby alleviating hallucination and ungrounded reasoning issues in LLMs. Experimental results show that the proposed approach outperforms multiple strong comparison methods, along with achieving state-of-the-art performance on widely used benchmarks WebQSP and CWQ.

Figures

Figures reproduced from arXiv: 2412.12643 by the authors.

Figure 1
Figure 1. The generation-based methods tend to generate unsupported or redundant subgraphs and reasoning results (left), while the proposed method address the issue by establishing proper searching space for each of the KGQA subtasks (right). Despite their success, the generative reasoning methods often produce ungrounded planning or reasoning results due to the hallucinatory behavior (Zhang et al., 2023; Sun et al., 2024b; P… view at source ↗
Figure 2
Figure 2. The proposed READS for KGQA. Start from the question (bottom left) with a given starting entity [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The number of cases with the size (number of [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: The trend of average Hits@1 as the size (number of triplets) of retrieved subgraph increases. We analyze the impact of the size of the retrieved subgraph (i.e., the number of triples included) on the overall performance of the strategy (the result is shown in [PITH_FU…
Figure 5
Figure 5. Figure 5: Case frequency of different types of errors, [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: The Hits@1 performance using different proportion of finetuning data. As shown in [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 10 canonical work pages

  1. [1]

    Kurt Bollacker, Colin Evans, Praveen Paritosh, Tim Sturge, and Jamie Taylor. 2008. Freebase: a collaboratively created graph database for structuring human knowledge. In Proceedings of the 2008 ACM SIGMOD international conference on Management of data, pages 1247--1250

    2008

  2. [2]

    Shulin Cao, Jiaxin Shi, Liangming Pan, Lunyiu Nie, Yutong Xiang, Lei Hou, Juanzi Li, Bin He, and Hanwang Zhang. 2022. Kqa pro: A dataset with explicit compositional programs for complex question answering over knowledge base. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 6101--6119

    work page 2022

  3. [3]

    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

    arXiv 2024

  4. [4]

    o ldiak, Pascal Hitzler, Thomas Icard, Kai-Uwe K \

    Artur d'Avila Garcez, Tarek R Besold, Luc De Raedt, Peter F \"o ldiak, Pascal Hitzler, Thomas Icard, Kai-Uwe K \"u hnberger, Luis C Lamb, Risto Miikkulainen, and Daniel L Silver. 2015. Neural-symbolic learning and reasoning: contributions and challenges. In 2015 AAAI Spring Symposium Series

    work page 2015

  5. [5]

    Yu Gu, Xiang Deng, and Yu Su. 2023. Don’t generate, discriminate: A proposal for grounding language models to real-world environments. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 4928--4949

    work page 2023

  6. [6]

    Yu Gu and Yu Su. 2022. https://aclanthology.org/2022.coling-1.148 A rcane QA : Dynamic program induction and contextualized encoding for knowledge base question answering . In Proceedings of the 29th International Conference on Computational Linguistics, pages 1718--1731, Gyeongju, Republic of Korea. International Committee on Computational Linguistics

    work page 2022

  7. [7]

    Gaole He, Yunshi Lan, Jing Jiang, Wayne Xin Zhao, and Ji-Rong Wen. 2021. https://doi.org/10.1145/3437963.3441753 Improving multi-hop knowledge base question answering by learning intermediate supervision signals . In Proceedings of the 14th ACM International Conference on Web Search and Data Mining, WSDM ’21. ACM

    arXiv 2021

  8. [8]

    Ruixin Hong, Hongming Zhang, Hong Zhao, Dong Yu, and Changshui Zhang. 2023. Faithful question answering with monte-carlo planning. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 3944--3965

    work page 2023

Show all 36 references
  1. [9]

    Jie Huang and Kevin Chen-Chuan Chang. 2023. Towards reasoning in large language models: A survey. In Findings of the Association for Computational Linguistics: ACL 2023, pages 1049--1065

    2023

  2. [10]

    Jinhao Jiang, Kun Zhou, Wayne Xin Zhao, Yang Song, Chen Zhu, Hengshu Zhu, and Ji-Rong Wen. 2024. Kg-agent: An efficient autonomous agent framework for complex reasoning over knowledge graph. arXiv preprint arXiv:2402.11163

    2024 arXiv

  3. [11]

    Jinhao Jiang, Kun Zhou, Wayne Xin Zhao, and Ji-Rong Wen. 2022. Unikgqa: Unified retrieval and reasoning for solving multi-hop question answering over knowledge graph. arXiv preprint arXiv:2212.00959

    2022 arXiv

  4. [12]

    Yunshi Lan and Jing Jiang. 2020. https://doi.org/10.18653/v1/2020.acl-main.91 Query graph generation for answering multi-hop complex questions from knowledge bases . In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 969--974, Onl...

    2020 doi

  5. [13]

    LINHAO LUO, Yuan-Fang Li, Reza Haf, and Shirui Pan. 2024. Reasoning on graphs: Faithful and interpretable large language model reasoning. In The Twelfth International Conference on Learning Representations

    2024

  6. [14]

    Alexander Miller, Adam Fisch, Jesse Dodge, Amir-Hossein Karimi, Antoine Bordes, and Jason Weston. 2016. https://doi.org/10.18653/v1/D16-1147 Key-value memory networks for directly reading documents . In Proceedings of the 2016 Conference on Empirical Methods in Natural Languag...

    2016 doi

  7. [15]

    Philipp Mondorf and Barbara Plank. 2024. Beyond accuracy: Evaluating the reasoning behavior of large language models--a survey. arXiv preprint arXiv:2404.01869

    2024 arXiv

  8. [16]

    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 & Data Engineering, (01):1--20

    2024

  9. [17]

    Haitian Sun, Bhuwan Dhingra, Manzil Zaheer, Kathryn Mazaitis, Ruslan Salakhutdinov, and William Cohen. 2018. https://doi.org/10.18653/v1/D18-1455 Open domain question answering using early fusion of knowledge bases and text . In Proceedings of the 2018 Conference on Empirical ...

    2018 doi

  10. [18]

    Jiashuo Sun, Chengjin Xu, Lumingyuan Tang, Saizhuo Wang, Chen Lin, Yeyun Gong, Lionel Ni, Heung-Yeung Shum, and Jian Guo. 2024 a . Think-on-graph: Deep and responsible reasoning of large language model on knowledge graph. In The Twelfth International Conference on Learning Rep...

    2024

  11. [19]

    Kai Sun, Yifan Xu, Hanwen Zha, Yue Liu, and Xin Luna Dong. 2024 b . Head-to-tail: How knowledgeable are large language models (llms)? aka will llms replace knowledge graphs? In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computationa...

    2024

  12. [20]

    Alon Talmor and Jonathan Berant. 2018. The web as a knowledge-base for answering complex questions. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 641--651

    2018

  13. [21]

    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

  14. [22]

    Keheng Wang, Feiyu Duan, Sirui Wang, Peiguang Li, Yunsen Xian, Chuantao Yin, Wenge Rong, and Zhang Xiong. 2023 a . https://arxiv.org/abs/2308.13259 Knowledge-driven cot: Exploring faithful reasoning in llms for knowledge-intensive question answering . Preprint, arXiv:2308.13259

    2023 arXiv

  15. [23]

    Lei Wang, Chen Ma, Xueyang Feng, Zeyu Zhang, Hao Yang, Jingsen Zhang, Zhiyuan Chen, Jiakai Tang, Xu Chen, Yankai Lin, et al. 2024. A survey on large language model based autonomous agents. Frontiers of Computer Science, 18(6):186345

    2024

  16. [24]

    Lei Wang, Wanyu Xu, Yihuai Lan, Zhiqiang Hu, Yunshi Lan, Roy Ka-Wei Lee, and Ee-Peng Lim. 2023 b . Plan-and-solve prompting: Improving zero-shot chain-of-thought reasoning by large language models. In Proceedings of the 61st Annual Meeting of the Association for Computational ...

    2023

  17. [25]

    Haoyi Xiong, Zhiyuan Wang, Xuhong Li, Jiang Bian, Zeke Xie, Shahid Mumtaz, Anwer Al-Dulaimi, and Laura E Barnes. 2024. Converging paradigms: The synergy of symbolic and connectionist ai in llm-empowered autonomous agents. arXiv preprint arXiv:2407.08516

    2024 arXiv

  18. [26]

    An Yang, Baosong Yang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Zhou, Chengpeng Li, Chengyuan Li, Dayiheng Liu, Fei Huang, et al. 2024. Qwen2 technical report. arXiv preprint arXiv:2407.10671

    2024 arXiv

  19. [27]

    Michihiro Yasunaga, Hongyu Ren, Antoine Bosselut, Percy Liang, and Jure Leskovec. 2021. https://doi.org/10.18653/v1/2021.naacl-main.45 QA - GNN : Reasoning with language models and knowledge graphs for question answering . In Proceedings of the 2021 Conference of the North Ame...

    2021 doi

  20. [28]

    Xi Ye, Semih Yavuz, Kazuma Hashimoto, Yingbo Zhou, and Caiming Xiong. 2022. https://doi.org/10.18653/v1/2022.acl-long.417 RNG - KBQA : Generation augmented iterative ranking for knowledge base question answering . In Proceedings of the 60th Annual Meeting of the Association fo...

    2022 doi

  21. [29]

    Wen-tau Yih, Matthew Richardson, Christopher Meek, Ming-Wei Chang, and Jina Suh. 2016. The value of semantic parse labeling for knowledge base question answering. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers...

    2016

  22. [30]

    Donghan Yu, Sheng Zhang, Patrick Ng, Henghui Zhu, Alexander Hanbo Li, Jun Wang, Yiqun Hu, William Yang Wang, Zhiguo Wang, and Bing Xiang. 2022. Decaf: Joint decoding of answers and logical forms for question answering over knowledge bases. In The Eleventh International Confere...

    2022

  23. [31]

    Jing Zhang, Xiaokang Zhang, Jifan Yu, Jian Tang, Jie Tang, Cuiping Li, and Hong Chen. 2022. https://doi.org/10.18653/v1/2022.acl-long.396 Subgraph retrieval enhanced model for multi-hop knowledge base question answering . In Proceedings of the 60th Annual Meeting of the Associ...

    2022 doi

  24. [32]

    Yue Zhang, Yafu Li, Leyang Cui, Deng Cai, Lemao Liu, Tingchen Fu, Xinting Huang, Enbo Zhao, Yu Zhang, Yulong Chen, et al. 2023. Siren's song in the ai ocean: a survey on hallucination in large language models. arXiv preprint arXiv:2309.01219

    2023 arXiv

  25. [33]

    Wanjun Zhong, Lianghong Guo, Qiqi Gao, He Ye, and Yanlin Wang. 2024. Memorybank: Enhancing large language models with long-term memory. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 19724--19731

    2024

  26. [34]

    Yuqi Zhu, Xiaohan Wang, Jing Chen, Shuofei Qiao, Yixin Ou, Yunzhi Yao, Shumin Deng, Huajun Chen, and Ningyu Zhang. 2024. Llms for knowledge graph construction and reasoning: Recent capabilities and future opportunities. World Wide Web, 27(5):58

    2024

  27. [35]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list...

  28. [36]

    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 11, 2026 · model on record in the stance chip above.