Pith. sign in

REVIEW 3 major objections 4 minor 56 references

Multi-level Matching Network for Multimodal Entity Linking

T0 review · 3 major / 4 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 M3EL links mentions to entities by adding same-modality contrastive learning and two-direction cross-modal matching, reporting new state-of-the-art scores on WikiMEL, RichpediaMEL, and WikiDiverse.

desk verdict Plausible incremental MEL model with a well-ablated attribute variant, but the headline description-based gains are confounded by richer text inputs and the paper overstates some of its results. read the letter →

arxiv 2412.10440 v1 pith:WVKRYQ3K submitted 2024-12-11 cs.CV cs.AI

Zhiwei Hu , Víctor Gutiérrez-Basulto , Ru Li , Jeff Z. Pan This is my paper · ORCID
classification cs.CVcs.AI
keywords multimodalentitylinkingintra-modalcontrastivelearningcross-modalmatchingdisambiguationCLIPvision-languageknowledgebase
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

Multimodal entity linking decides which knowledge-base entity an ambiguous mention refers to when both text and an image are available. M3EL, the network proposed here, adds two ingredients that prior models miss: negative examples taken from inside the same modality, and cross-modal matching that runs in both directions (text-to-image and image-to-text). On the three standard benchmarks, WikiMEL, RichpediaMEL, and WikiDiverse, M3EL reports the highest MRR and Hits@1 scores, with the biggest relative jump on WikiDiverse, the dataset with the shortest entity text. The paper argues these gains show that same-modality negatives and bidirectional matching should be part of any MEL system.

What carries the argument

The argument is carried by a composite matching score $M_U=(M_T+M_V+M_C)/3$, where $M_T$ and $M_V$ are intra-modal (text-text, image-image) scores combining global-to-global dot products and global-to-local attention-weighted scores, and $M_C$ averages the textual-to-visual and visual-to-textual cross-modal scores obtained via bidirectional attention with multi-head fusion and a gating operation. The joint loss sums the intra-modal contrastive loss $\mathcal{L}_{cl}$, a unit-consistent loss on $M_U$, and separate unit-consistent losses on each component score. The intra-modal contrastive loss treats matching mention-entity pairs as positives and both same-source and inter-source embeddings within the modality as negatives, with separate weight coefficients $\beta$ and $\gamma$.

What would settle it

Run the MIMIC baseline with exactly the Wikidata description text used for M3ELdesc on all three datasets; if MIMIC then matches or exceeds M3ELdesc's MRR and Hits@1, the claim that the new modules drive the improvement is falsified. A complementary check is to feed M3ELattr the same descriptions and measure how much of the gain over MIMIC comes from the text change alone.

Watch

Extended reading notes

Core claim

On the paper's own numbers, M3ELdesc reaches MRR/Hits@1 of 92.30/88.84 on WikiMEL, 88.26/82.82 on RichpediaMEL, and 81.29/74.06 on WikiDiverse, beating the strongest baseline MIMIC (91.82/87.98, 86.95/81.02, and 73.44/63.51, respectively) on every dataset and metric. The authors attribute the improvement to three components working together: an intra-modal contrastive loss that distinguishes an entity's textual (or visual) embedding from other entities and mentions in the same modality; an intra-modal matching network that scores both global-to-global and global-to-local similarity; and a cross-modal matching network that computes both textual-to-visual and visual-to-textual matching scores. The paper shows through ablations that removing any of these components, or replacing the contrastive loss with InfoNCE or MCLET, lowers accuracy.

Load-bearing premise

The headline comparison assumes that M3EL's gains over MIMIC come from its new matching and contrastive modules, but the winning variant uses richer entity descriptions than MIMIC had, and no baseline was run on the same text.

Editorial extensions

If this is right

  • MEL systems should mine negative pairs within each modality, not only between modalities, to make embeddings more discriminative.
  • Bidirectional text-image matching outperforms either direction alone, as the T2V/V2T ablation in the appendix shows.
  • The modules transfer to low-resource settings: M3EL improves over MIMIC on 10% and 20% training subsets, with gains growing as training data increases.
  • Entity text richness matters: switching from attribute knowledge to longer Wikidata descriptions improves results, most on WikiDiverse.

Reading between the lines

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

  • A matched-text comparison is the missing experiment: because M3ELdesc uses longer Wikidata descriptions while the MIMIC numbers were produced with attribute knowledge, part of the reported gap may be due to the richer text rather than the new modules; running MIMIC on the same descriptions would separate the two effects.
  • The paper's own appendix suggests the framework is encoder-agnostic; if the modules transfer to other vision-language encoders (for example, BLIP or BLIP-2 at full precision), the design is a general recipe for MEL rather than a CLIP-specific fix.
  • The contrastive loss is applied only at the global-feature level because local-level contrastive learning hurt performance; a hierarchical contrastive scheme that weights local negatives differently might recover that signal.
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 / 4 minor

Summary. The paper proposes M3EL, a multimodal entity linking model with three modules: multimodal feature extraction with intra-modal contrastive learning (ICL), an intra-modal matching network with global-to-global and global-to-local matching, and a bidirectional cross-modal matching network with textual-to-visual and visual-to-textual matching. The model is trained with a joint loss combining the contrastive loss, a union matching loss, and per-module losses. Experiments are reported on WikiMEL, RichpediaMEL, and WikiDiverse in full and low-resource settings, with two variants: M3ELattr, which uses entity attribute knowledge as in MIMIC, and M3ELdesc, which uses longer Wikidata descriptions. The authors report that M3ELdesc outperforms the MIMIC baseline on all three datasets, and ablation studies are used to attribute gains to the proposed losses and modules. Code and data links are provided in the paper.

Significance. If the reported gains are reliable, the paper makes a useful contribution to multimodal entity linking by combining intra-modal negative sampling and bidirectional cross-modal matching on top of a CLIP backbone. The paper includes a thorough ablation suite covering losses, modules, pooling operations, contrastive losses, parameter sensitivity, and resource consumption, and it releases code and data. The strongest part of the evidence is the M3ELattr variant, which follows MIMIC's attribute-text input and still outperforms MIMIC on the three datasets, together with ablations showing that removing modules generally degrades performance. However, the headline M3ELdesc variant changes the entity text source simultaneously with the architecture, so the main comparison against MIMIC is confounded. The reported margins on WikiMEL for the fairer M3ELattr comparison are small (0.48 MRR and 0.51 Hits@1), and no significance testing or error bars are provided, so the central claim of state-of-the-art performance is not yet fully supported.

major comments (3)
  1. [§4.2, Tables 2 and 3] The headline comparison between M3ELdesc and MIMIC is confounded by the entity text source. Section 4.2 states that M3ELdesc uses Wikidata descriptions instead of MIMIC's attribute knowledge, and it explicitly reports that on WikiDiverse the average entity text length grows from 1.24 to 4.50 words. The large WikiDiverse gains of M3ELdesc over MIMIC (7.85 MRR and 10.55 Hits@1) therefore cannot be attributed to the proposed modules, because no MIMIC baseline was run with the same Wikidata description input. The paper also states that 'M3EL appearing in subsequent parts will uniformly refer to M3ELdesc,' so the main claim is built on this confounded comparison. The authors should either add a MIMIC+description baseline, or make M3ELattr (which uses MIMIC's attribute text) the primary variant for the state-of-the-art claim and treat M3ELdesc as an additional input-change study.
  2. [§4.1–4.2, Tables 2 and 3] All baseline numbers in Tables 2 and 3 are copied from the MIMIC paper, and no error bars, repeated runs, or significance tests are reported. This matters because the fairer M3ELattr comparison on WikiMEL is very small: 0.48 MRR and 0.51 Hits@1 over MIMIC (92.30 vs. 91.82 MRR and 88.49 vs. 87.98 Hits@1). Without variance or a paired test, these margins are within the range that could arise from optimization randomness or evaluation differences. The authors should report means and standard deviations over multiple seeds, ideally with a significance test, and should state whether the MIMIC results were obtained with the same candidate sets, text preprocessing, and evaluation code.
  3. [§4.3, Table 4] The ablation table does not consistently support the claim that every loss component contributes. On RichpediaMEL, removing the union loss L_U gives MRR 88.34 versus 88.26 for the full model, and removing the contrastive loss L_cl gives Hits@1 83.63 versus 82.82; on WikiDiverse, removing L_U also gives a higher Hits@3 (87.15 vs. 86.57). The text acknowledges 'singularies' but still concludes that 'the overall performance is the best when all losses and modules are simultaneously used.' This is not strictly true for all metrics on all datasets, and the discrepancy should be analyzed or the claims softened. The module-level ablations (w/o L_T+M_T, w/o L_V+M_V, w/o L_C+M_C) are more supportive, but the loss-level analysis needs a clearer interpretation.
minor comments (4)
  1. [§3.4, Equation 8] The formula for the cross-modal matching score writes M_C = (M_T2V^m + M_V2T^m)/2, but the superscript m is inconsistent with the definitions in Equations 7 and 8, where M_T2V and M_V2T are already the combined entity- and mention-level scores. This should be M_C = (M_T2V + M_V2T)/2.
  2. [§3.2.2, Equation 2] The text says 'the nominator is symmetric,' but the correct English term is 'numerator.' Also, the asymmetry is in the denominator, which includes inner-source and inter-source negative sums; this could be stated more explicitly.
  3. [Figure 4 and Table 4] There are several typos: Figure 4 axis labels read 'Inner-souce' and 'Inter-souce' instead of 'source,' and the ablation section uses 'singularies' instead of 'singularities.'
  4. [§4.2, Low Resource Setting] The text says 'the performance of M3ELl' in the low-resource paragraph; the subscript l is unexplained and should be removed.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the M3ELdesc-vs-MIMIC comparison is confounded by a change of entity text, but no prediction or derivation reduces to its own inputs.

full rationale

The paper's claims are benchmark comparisons against external datasets and independent baselines, not derivations from first principles. Each component of M3EL is defined in its own equations (ICL in Eq. 2, G2G/G2L in Eqs. 4-5, T2V/V2T in Eqs. 6-8, joint loss in Eq. 9) and is tested by ablations that remove it. No equation is defined in terms of the metric it is supposed to predict, and no grid-searched hyperparameter is relabeled as a prediction. The load-bearing evidence is empirical and external: Table 2 states 'all the baselines results are from the MIMIC [28] paper,' and the modules' contributions are probed by the paper's own ablation tables rather than by self-citation. The one serious methodological weakness is in Section 4.2: the advertised M3ELdesc variant changes both the architecture and the entity text source, using Wikidata descriptions instead of MIMIC's attribute knowledge, with the paper noting that word length grows from 1.24 to 4.50 on WikiDiverse and concluding that 'the availability of textual knowledge is one of the important factors affecting the model performance.' That makes the headline gain over MIMIC confounded, but confounding is not circularity: the comparison numbers are independent external results, and M3ELattr, which keeps MIMIC's attribute text, still beats MIMIC by smaller margins. Self-citations to MCLET [16] and MIMIC [28] supply comparison or baseline losses, but the argument does not reduce to those citations. No uniqueness theorem is imported, and no ansatz is smuggled in as an externally derived law. Therefore no circular step can be exhibited under the required standard.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The model is an engineering combination of CLIP, MIMIC's unit-consistent objective, a contrastive loss close to the authors' earlier MCLET, and attention modules. The central claim relies on grid-selected hyperparameters and on the assumption that CLIP embeddings and the copied loss function are appropriate for MEL. No new conceptual entities are introduced.

free parameters (5)
  • beta (inner-source alignment weight in Eq. 2) = Grid searched over {0.2, 0.4, 0.6, 0.8, 1.0}; no per-dataset best value reported
    Controls the weight of same-view negative samples in the intra-modal contrastive loss; Appendix A shows performance fluctuates as beta increases.
  • gamma (inter-source alignment weight in Eq. 2) = Grid searched over {0.2, 0.4, 0.6, 0.8, 1.0}; larger values generally better
    Controls the weight of cross-view negative samples in the intra-modal contrastive loss; selected by grid search per dataset.
  • tau (temperature in Eq. 2) = 0.03 after grid search over {0.03, 0.10, 0.25, 0.5, 0.75}
    Temperature for cosine similarity in the contrastive loss; Appendix A shows smaller values are best for all three datasets.
  • K (number of attention heads in Eq. 7) = 5 after grid search over {3, 4, 5, 6, 7}
    Number of heads in the multi-head attention fusion; selected by grid search and reported in Appendix A.
  • omega_k (per-head temperature in Eq. 7) = Not specified; only constrained to be greater than 0
    Introduced as the temperature controlling score sharpness in multi-head attention; no value, range, or sensitivity analysis is given.
assumptions (4)
  • domain assumption Pretrained CLIP (ViT-B/32) embeddings are appropriate for intra-modal and cross-modal matching in MEL.
    Section 3.2 uses CLIP as the feature extractor; if CLIP's text-image alignment does not transfer to entity descriptions and images, the matching modules inherit the bias.
  • domain assumption MIMIC's unit-consistent objective L_uco is a valid loss for combining matching scores.
    Equation 9 adopts L_uco from [28] without derivation or adaptation; the claim that summing union, textual, visual and cross-modal losses helps depends on it.
  • domain assumption The benchmark splits and baseline numbers copied from MIMIC are directly comparable.
    All baseline rows in Tables 2 and 3 are taken from the MIMIC paper; if input preprocessing or splits differ, the cross-paper comparisons are invalid.
  • domain assumption In-batch negative samples from the same modality improve MEL representations.
    This is the core hypothesis motivating the intra-modal contrastive learning module in Section 3.2.2; it is treated as true by the design and supported only by internal ablations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multi-level Matching Network for Multimodal Entity Linking." pith.science (2026). https://pith.science/paper/WVKRYQ3K

@misc{pith2026241210440,
  author       = {Pith},
  title        = {Pith review of: Multi-level Matching Network for Multimodal Entity Linking},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WVKRYQ3K}},
  note         = {Machine review of arXiv:2412.10440}
}
read the original abstract

Multimodal entity linking (MEL) aims to link ambiguous mentions within multimodal contexts to corresponding entities in a multimodal knowledge base. Most existing approaches to MEL are based on representation learning or vision-and-language pre-training mechanisms for exploring the complementary effect among multiple modalities. However, these methods suffer from two limitations. On the one hand, they overlook the possibility of considering negative samples from the same modality. On the other hand, they lack mechanisms to capture bidirectional cross-modal interaction. To address these issues, we propose a Multi-level Matching network for Multimodal Entity Linking (M3EL). Specifically, M3EL is composed of three different modules: (i) a Multimodal Feature Extraction module, which extracts modality-specific representations with a multimodal encoder and introduces an intra-modal contrastive learning sub-module to obtain better discriminative embeddings based on uni-modal differences; (ii) an Intra-modal Matching Network module, which contains two levels of matching granularity: Coarse-grained Global-to-Global and Fine-grained Global-to-Local, to achieve local and global level intra-modal interaction; (iii) a Cross-modal Matching Network module, which applies bidirectional strategies, Textual-to-Visual and Visual-to-Textual matching, to implement bidirectional cross-modal interaction. Extensive experiments conducted on WikiMEL, RichpediaMEL, and WikiDiverse datasets demonstrate the outstanding performance of M3EL when compared to the state-of-the-art baselines.

Figures

Figures reproduced from arXiv: 2412.10440 by the authors.

Figure 1
Figure 1. An example of MEL. Dotted boxes of different col￾ors represent different features: color purple for mention textual description (mention text), color orange for mention visual context (mention image), color green for entity textual description (entity text), color blue for entity visual context (entity image). EL supports numerous downstream information-retrieval appli￾cations, such as question answering [18, 26, 44… view at source ↗
Figure 2
Figure 2. The structure of the M3EL model, containing three modules: Multimodal Feature Extraction (MFE) with Intra-modal Contrastive Learning (ICL), Intra-modal Matching Network (IMN) and Cross-modal Matching Network (CMN). Att and M-Att denote the attention and multi-heads attention mechanisms, respectively. a full-transformer vision-language pre-trained model in an end￾to-end manner. MIMIC [28] utilizes BERT [8] and CLIP a… view at source ↗
Figure 3
Figure 3. Illustrative comparison of intra-modal contrastive learning with CLIP, where the red dashed lines represent the positive samples, yellow lines denote the negative samples in CLIP, purple and blue lines represent the inner-source and intra-source negative samples. Circles and squares represent textual and visual features, 𝑒 and 𝑚 represent entity and mention, respectively. ⊲ Textual Modal Embedding. We first concaten… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 51 canonical work pages

  1. [1]

    Omar Adjali, Romaric Besançon, Olivier Ferret, Hervé Le Borgne, and Brigitte Grau. 2020. Multimodal Entity Linking for Tweets. In ECIR. Springer, Lisbon, Portugal, 463–478

    work page 2020

  2. [2]

    Ali Ahmadvand, Harshita Sahijwani, Jason Ingyu Choi, and Eugene Agichtein

  3. [3]

    Alex Andonian, Shixing Chen, and Raffay Hamid. 2022. Robust Cross-Modal Representation Learning with Progressive Self-Distillation. In CVPR. IEEE, New Orleans, LA, USA, 16409–16420

    work page 2022

  4. [4]

    Ilaria Bordino, Yelena Mejova, and Mounia Lalmas. 2013. Penguins in sweaters, or serendipitous entity search on user-generated content. In CIKM. ACM, San Francisco, CA, USA, 109–118

    work page 2013

  5. [5]

    Nicola De Cao, Gautier Izacard, Sebastian Riedel, and Fabio Petroni. 2021. Au- toregressive Entity Retrieval. In ICLR. OpenReview.net, online, 1–20

    work page 2021

  6. [6]

    Yixin Cao, Lei Hou, Juanzi Li, and Zhiyuan Liu. 2018. Neural Collective Entity Linking. In COLING. ACL, Santa Fe, New Mexico, USA, 675–686

    work page 2018

  7. [7]

    Tao Cheng, Xifeng Yan, and Kevin Chen-Chuan Chang. 2007. EntityRank: Search- ing Entities Directly and Holistically. In VLDB. ACM, Vienna, Austria, 387–398

    work page 2007

  8. [8]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In NAACL. ACL, Minneapolis, MN, USA, 4171–4186

    work page 2019

Show all 56 references
  1. [9]

    Zi-Yi Dou, Yichong Xu, Zhe Gan, Jianfeng Wang, Shuohang Wang, Lijuan Wang, Chenguang Zhu, Pengchuan Zhang, Lu Yuan, Nanyun Peng, Zicheng Liu, and Michael Zeng. 2022. An Empirical Study of Training End-to-End Vision-and- Language Transformers. In CVPR. IEEE, New Orleans, LA, US...

    2022

  2. [10]

    Zheng Fang, Yanan Cao, Qian Li, Dongjie Zhang, Zhenyu Zhang, and Yanbing Liu. 2019. Joint Entity Linking with Deep Reinforcement Learning. In WWW. ACM, San Francisco, CA, USA, 438–447

    2019

  3. [11]

    Octavian-Eugen Ganea and Thomas Hofmann. 2017. Deep Joint Entity Disam- biguation with Local Neural Attention. In EMNLP. ACL, Copenhagen, Denmark, 2619–2629

    2017

  4. [12]

    Xiaoqi Han, Ru Li, Xiaoli Li, Jiye Liang, Zifang Zhang, and Jeff Z. Pan. 2024. InstructEd: Soft-Instruction Tuning for Model Editing with Hops. In ACL. ACL, Bangkok, Thailand, 14953–14968

    2024

  5. [13]

    Bowei He, Xu He, Yingxue Zhang, Ruiming Tang, and Chen Ma. 2023. Dynami- cally Expandable Graph Convolution for Streaming Recommendation. In WWW. ACM, Austin, TX, USA, 1457–1467

    2023

  6. [14]

    Girshick

    Kaiming He, Haoqi Fan, Yuxin Wu, Saining Xie, and Ross B. Girshick. 2020. Momentum Contrast for Unsupervised Visual Representation Learning. In CVPR. IEEE, Seattle, WA, USA, 9726–9735

    2020

  7. [15]

    Johannes Hoffart, Mohamed Amir Yosef, Ilaria Bordino, Hagen Fürstenau, Man- fred Pinkal, Marc Spaniol, Bilyana Taneva, Stefan Thater, and Gerhard Weikum

  8. [16]

    Zhiwei Hu, Víctor Gutiérrez-Basulto, Zhiliang Xiang, Ru Li, and Jeff Z. Pan. 2023. Multi-view Contrastive Learning for Entity Typing over Knowledge Graphs. In EMNLP. ACL, Singapore, 12950–12963

    2023

  9. [17]

    Zhiwei Hu, Víctor Gutiérrez-Basulto, Zhiliang Xiang, Ru Li, and Jeff Z. Pan. 2024. Leveraging Intra-modal and Inter-modal Interaction for Multi-Modal Entity Alignment. CoRR abs/2404.17590 (2024). arXiv:2404.17590

    2024 arXiv

  10. [18]

    Zhiwei Hu, Víctor Gutiérrez-Basulto, Zhiliang Xiang, Xiaoli Li, Ru Li, and Jeff Z. Pan. 2022. Type-aware Embeddings for Multi-Hop Reasoning over Knowledge Graphs. In IJCAI. ijcai.org, Vienna, Austria, 3078–3084

    2022

  11. [19]

    Wonjae Kim, Bokyung Son, and Ildoo Kim. 2021. ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision. InICML. PMLR, online, 5583–5594

    2021

  12. [20]

    Phong Le and Ivan Titov. 2018. Improving Entity Linking by Modeling Latent Relations between Mentions. In ACL. ACL, Melbourne, Australia, 1595–1604

    2018

  13. [21]

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven C. H. Hoi. 2023. BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models. In ICML. PMLR, Honolulu, Hawaii, USA, 19730–19742

    2023

  14. [22]

    Junnan Li, Dongxu Li, Caiming Xiong, and Steven C. H. Hoi. 2022. BLIP: Boot- strapping Language-Image Pre-training for Unified Vision-Language Understand- ing and Generation. In ICML. PMLR, Baltimore, Maryland, USA, 12888–12900

    2022

  15. [23]

    Selvaraju, Akhilesh Gotmare, Shafiq R

    Junnan Li, Ramprasaath R. Selvaraju, Akhilesh Gotmare, Shafiq R. Joty, Caiming Xiong, and Steven Chu-Hong Hoi. 2021. Align before Fuse: Vision and Lan- guage Representation Learning with Momentum Distillation. In NeurIPS. Curran Associates, online, 9694–9705

    2021

  16. [24]

    Lizi Liao, Yunshan Ma, Xiangnan He, Richang Hong, and Tat-Seng Chua. 2018. Knowledge-aware Multimodal Dialogue Systems. In MM. ACM, Seoul, Republic of Korea, 801–809

    2018

  17. [25]

    Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. RoBERTa: A Robustly Optimized BERT Pretraining Approach. CoRR abs/1907.11692 (2019), 1–13

    2019 arXiv

  18. [26]

    Shayne Longpre, Kartik Perisetla, Anthony Chen, Nikhil Ramesh, Chris DuBois, and Sameer Singh. 2021. Entity-Based Knowledge Conflicts in Question Answer- ing. In EMNLP. ACL, online, 7052–7063

    2021

  19. [27]

    Ilya Loshchilov and Frank Hutter. 2019. Decoupled Weight Decay Regularization. In ICLR. OpenReview.net, New Orleans, LA, USA, 1–11

    2019

  20. [28]

    Pengfei Luo, Tong Xu, Shiwei Wu, Chen Zhu, Linli Xu, and Enhong Chen. 2023. Multi-Grained Multimodal Interaction Network for Entity Linking. InKDD. ACM, Long Beach, CA, USA, 1583–1594

    2023

  21. [29]

    Seungwhan Moon, Leonardo Neves, and Vitor Carvalho. 2018. Multimodal Named Entity Disambiguation for Noisy Social Media Posts. In ACL. ACL, Mel- bourne, Australia, 2000–2008

    2018

  22. [30]

    OpenAI. 2023. Chatgpt. https://chat.openai.com,2023

    2023

  23. [31]

    Weiran Pan, Wei Wei, and Xian-Ling Mao. 2021. Context-aware Entity Typing in Knowledge Graphs. In EMNLP. ACL, online, 2240–2250

    2021

  24. [32]

    Peters, Mark Neumann, Robert L

    Matthew E. Peters, Mark Neumann, Robert L. Logan IV, Roy Schwartz, Vidur Joshi, Sameer Singh, and Noah A. Smith. 2019. Knowledge Enhanced Contextual Word Representations. In EMNLP. ACL, Hong Kong, China, 43–54

    2019

  25. [33]

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. 2021. Learning Transferable Visual Models From Natural Language Supervision. In ICML. PMLR,...

    2021

  26. [34]

    Chenwei Ran, Wei Shen, and Jianyong Wang. 2018. An Attention Factor Graph Model for Tweet Entity Linking. In WWW. ACM, Lyon, France, 1135–1144

    2018

  27. [35]

    Wei Shen, Jianyong Wang, and Jiawei Han. 2015. Entity Linking with a Knowl- edge Base: Issues, Techniques, and Solutions. TKDE 27, 2 (2015), 443–460

    2015

  28. [36]

    Senbao Shi, Zhenran Xu, Baotian Hu, and Min Zhang. 2024. Generative Multi- modal Entity Linking. In COLING. ELRA and ICCL, Torino, Italy, 7654–7665

    2024

  29. [37]

    Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yas- mine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhos- ale, Dan Bikel, Lukas Blecher, Cristian Canton-Ferrer, Moya Chen, Guillem Cucurull, David Esiobu, Jude Fernandes, Jeremy F...

    2023 arXiv

  30. [38]

    Denny Vrandecic and Markus Krötzsch. 2014. Wikidata: a free collaborative knowledgebase. Commun. ACM 57, 10 (2014), 78–85

    2014

  31. [39]

    Meng Wang, Haofen Wang, Guilin Qi, and Qiushuo Zheng. 2020. Richpedia: A Large-Scale, Comprehensive Multi-Modal Knowledge Graph. Big Data Res. 22 (2020), 100159

    2020

  32. [40]

    Peng Wang, Jiangheng Wu, and Xiaohang Chen. 2022. Multimodal Entity Linking with Gated Hierarchical Fusion and Contrastive Training. InSIGIR. ACM, Madrid, Spain, 938–948

    2022

  33. [41]

    Xuwu Wang, Junfeng Tian, Min Gui, Zhixu Li, Rui Wang, Ming Yan, Lihan Chen, and Yanghua Xiao. 2022. WikiDiverse: A Multimodal Entity Linking Dataset with Diversified Contextual Topics and Entity Types. In ACL. ACL, Dublin, Ireland, 4785–4797

    2022

  34. [42]

    Ledell Wu, Fabio Petroni, Martin Josifoski, Sebastian Riedel, and Luke Zettle- moyer. 2020. Scalable Zero-shot Entity Linking with Dense Entity Retrieval. In EMNLP. ACL, online, 6397–6407

    2020

  35. [43]

    Shangyu Xing, Fei Zhao, Zhen Wu, Chunhui Li, Jianbing Zhang, and Xinyu Dai. 2023. DRIN: Dynamic Relation Interactive Network for Multimodal Entity Linking. In MM. ACM, Ottawa, ON, Canada, 3599–3608

    2023

  36. [44]

    Wenhan Xiong, Mo Yu, Shiyu Chang, Xiaoxiao Guo, and William Yang Wang

  37. [45]

    Zhichao Yan, Jiapu Wang, Jiaoyan Chen, Xiaoli Li, Ru Li, and Jeff Z Pan. 2024. Atomic Fact Decomposition Helps Attributed Question Answering. CoRR abs/2410.16708 (2024), 1–17

    2024 arXiv

  38. [46]

    Chengmei Yang, Bowei He, Yimeng Wu, Chao Xing, Lianghua He, and Chen Ma

  39. [47]

    Improving Question Answering over Incomplete KBs with Knowledge- Aware Reader. In ACL. ACL, Florence, Italy, 4258–4264

  40. [48]

    Xiyuan Yang, Xiaotao Gu, Sheng Lin, Siliang Tang, Yueting Zhuang, Fei Wu, Zhigang Chen, Guoping Hu, and Xiang Ren. 2019. Learning Dynamic Context Augmentation for Global Entity Linking. In EMNLP. ACL, Hong Kong, China, 271–281. Multi-level Matching Network for Multimodal Entit...

    2019

  41. [49]

    Wen-tau Yih, Ming-Wei Chang, Xiaodong He, and Jianfeng Gao. 2015. Semantic Parsing via Staged Query Graph Generation: Question Answering with Knowl- edge Base. In ACL. ACL, Beijing, China, 1321–1331

    2015

  42. [50]

    Qiushuo Zheng, Hao Wen, Meng Wang, and Guilin Qi. 2022. Visual Entity Linking via Multi-modal Learning. Data Intell. 4, 1 (2022), 1–19

    2022

  43. [51]

    Jinyu Yang, Jiali Duan, Son Tran, Yi Xu, Sampath Chanda, Liqun Chen, Belinda Zeng, Trishul Chilimbi, and Junzhou Huang. 2022. Vision-Language Pre-Training with Triple Contrastive Learning. In CVPR. IEEE, New Orleans, LA, USA, 15650– 15659

    2022

  44. [52]

    Gehler, and Thomas Brox

    Mohammadreza Zolfaghari, Yi Zhu, Peter V. Gehler, and Thomas Brox. 2021. CrossCLR: Cross-modal Contrastive Learning For Multi-modal Video Represen- tations. In ICCV. IEEE, Montreal, QC, Canada, 1430–1439. APPENDIX A Parameter Sensitivity We carry out parameter sensitivity expe...

    2021

  45. [55]

    Yanqiao Zhu, Yichen Xu, Feng Yu, Qiang Liu, Shu Wu, and Liang Wang. 2021. Graph Contrastive Learning with Adaptive Augmentation. In WWW. ACM, online, 2069–2080

    2021

  46. [2011]

    In EMNLP

    Robust Disambiguation of Named Entities in Text. In EMNLP. ACL, Edin- burgh, UK, 782–792

  47. [2019]

    ConCET: Entity-Aware Topic Classification for Open-Domain Conversa- tional Agents. In CIKM. ACM, Beijing, China, 1371–1380

  48. [2023]

    MMEL: A Joint Learning Framework for Multi-Mention Entity Linking. In UAI. PMLR, Pittsburgh, PA, USA, 2411–2421

Pith tools

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