Pith. sign in

REVIEW 3 major objections 6 minor 56 references

ViCo: Word Embeddings from Visual Co-occurrences

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Visual co-occurrences in annotated images encode word meanings that text-only embeddings miss.

desk verdict Solid, useful paper with a genuine negative result about random embeddings; the downstream gains are too small and unreplicated to carry the 'gains on all tasks' claim. read the letter →

arxiv 1908.08527 v1 pith:XRKB75KI submitted 2019-08-22 cs.CV cs.CL

Tanmay Gupta , Alexander Schwing , Derek Hoiem This is my paper · ORCID
classification cs.CVcs.CL
keywords visualco-occurrenceswordembeddingsmulti-tasklog-bilinearmodelGloVezero-shotgeneralizationvision-languagetasksGenomeImageNet
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 word embeddings can be learned from visual co-occurrence—two words co-occur visually when both apply to the same image or region—and that these embeddings capture visual distinctions that text corpora obscure. From object and attribute annotations in Visual Genome and ImageNet, the authors count four kinds of co-occurrence: object-attribute, attribute-attribute, context, and object-hypernym. A multi-task log-bilinear model learns one compact vector per word shared across all four co-occurrence types. The paper claims that concatenating ViCo with GloVe outperforms GloVe alone on unsupervised clustering, zero-shot-like transfer to unseen CIFAR-100 classes, and all five downstream applications tested. It also claims that on supervised vision-language tasks, random word vectors perform nearly as well as learned ones—a result the authors present as contrary to conventional wisdom.

What carries the argument

The load-bearing mechanism is a multi-task log-bilinear objective. A shared word vector $w_i\in\mathbb{R}^d$ is projected through a co-occurrence-type-specific transformation $\varphi_t$ into a dedicated space, where a GloVe-style squared loss approximates the log of the observed count $X^t_{ij}$ plus a hinge-style penalty for zero-count pairs. Four count matrices corresponding to object-attribute, attribute-attribute, context, and object-hypernym co-occurrences are computed by expanding synset annotations across all member words. This formulation lets one compact embedding encode several senses of relatedness at once; the paper shows a linear $\varphi_t$ matches the concatenation of separately trained embeddings at lower dimension.

What would settle it

Train ViCo using only directly attested region labels with no synonym expansion and no WordNet hypernyms, and compare its clustering and zero-shot performance against GloVe; if the gains over GloVe disappear, the reported advantage comes from WordNet and annotation structure, not from visual co-occurrence itself. More narrowly, replacing each synset's words with its single most frequent surface form should preserve visual signal while removing synonym-expansion effects.

Watch

Extended reading notes

Core claim

ViCo's central claim is that visual co-occurrence statistics carry genuine word-meaning signal beyond text: words that describe the same visible stuff (salad and chopped, glass and clear, fork and spoon) end up closer in the learned space even when they rarely co-occur in prose. The paper demonstrates this by training a multi-task extension of GloVe's log-bilinear objective on four co-occurrence matrices, then showing that GloVe+ViCo beats GloVe alone on clustering, zero-shot generalization, and five downstream applications. A separate claim, supported by the same experiments, is that supervised vision-language models are largely insensitive to the choice of word embeddings: random vectors match learned ones on all four vision-language tasks, while the word-only discriminative-attributes task still strongly favors learned embeddings.

Load-bearing premise

The counting pipeline assumes that every synonym listed for an annotated concept truly applies to every image region carrying that concept label and that the region annotations are exhaustive; if labels are missing or synonyms are used loosely, the co-occurrence counts encode annotation and WordNet structure rather than visual co-occurrence.

Editorial extensions

If this is right

  • Using ViCo in place of or alongside GloVe should improve any task where word similarity must reflect visible properties, such as attribute-based retrieval or zero-shot object recognition.
  • Because the multi-task model shares one compact embedding across co-occurrence types, adding new co-occurrence sources need not inflate embedding dimensionality.
  • For supervised vision-language models, pretrained word embeddings may be replaceable by random vectors without much loss, freeing capacity or simplifying pipelines.
  • Word embeddings can be obtained from image annotation alone, which matters for domains or languages with little text but available visual labels.

Reading between the lines

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

  • If visual co-occurrence signal is as strong as reported, similar counting over other dense annotation sources (video captions, pixel-level segmentations, image tags) could produce embeddings for a much larger vocabulary without synset expansion.
  • The near-parity of random and learned embeddings on vision-language tasks suggests those benchmarks may be dominated by visual features and model architecture; a clean test would sweep vocabulary coverage or embedding quality while holding everything else fixed.
  • The type-specific projections could be reused as a cheap sense-disentangling tool: querying which projection gives high similarity reveals whether two words are related by appearance, function, hypernymy, or context.
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 ViCo, a method for learning word embeddings from visual co-occurrences. It formulates a multi-task extension of GloVe's log-bilinear objective (Eq. 2) in which a single shared word vector is projected through type-specific linear transforms into separate spaces for four co-occurrence types: object-attribute, attribute-attribute, context, and object-hypernym, with counts derived from VisualGenome and ImageNet annotations. The authors evaluate GloVe+ViCo against GloVe, random vectors, and vis-w2v on unsupervised clustering over manually annotated word categories, a zero-shot-like CIFAR-100 transfer protocol, five downstream tasks (discriminative attributes, caption-image retrieval, VQA, referring expression comprehension, image captioning), and qualitative analogies. The main claims are that GloVe+ViCo outperforms GloVe across these evaluations and that random embeddings perform comparably to learned embeddings on supervised vision-language tasks.

Significance. If the results hold, the paper offers a simple and scalable way to inject visual statistics into word embeddings and provides one of the first systematic demonstrations that frozen word embeddings have limited impact in several supervised vision-language models. The multi-task log-bilinear formulation is clean and the w/o WordNet ablation is a useful control. The clustering and zero-shot-like results are more substantial than the downstream results, which rest on small, unreplicated differences. The finding about random embeddings is interesting and potentially important, but the evidence for it is entangled with the same statistical reliability issues.

major comments (3)
  1. [Sec. 4.3, Table 5] The central claim that GloVe+ViCo outperforms GloVe on all downstream tasks rests on Table 5 differences that are mostly 0.1-0.3 absolute points, with no standard errors or significance tests. For example, VQA Overall is 67.7 vs 67.5, VQA Other is 58.4 vs 58.3, RefCOCO+ TestA is 75.5 vs 75.3, BLEU-4 is 0.291 vs 0.290, and SPICE is 0.168 vs 0.167. The image captioning row is particularly concerning: the random-embedding row alone reports BLEU-4 0.296, CIDEr 0.910, and SPICE 0.170, which are all higher than the corresponding GloVe+ViCo values (0.291, 0.894, 0.168), suggesting that captioning metrics can move around by amounts larger than the claimed gains. The authors should report multiple random seeds or bootstrap confidence intervals for every row of Table 5, or explicitly tone down the all-task claim. This is load-bearing because the abstract and conclusion claim gains on all tasks.
  2. [Sec. 4.2] The zero-shot-like protocol is described as 'slightly different' from typical zero-shot because images of unseen CIFAR-100 classes are used to compute the word co-occurrences used for embedding training. The paper then concludes that GloVe+ViCo 'generalizes to unseen classes better than GloVe.' This is not a genuine zero-shot transfer test, since the unseen class names and their visual co-occurrence statistics are available to the embedding. I do not see equation-level circularity, but the wording overstates what is measured. The authors should either re-frame this as a novel-class classification experiment with supervised embeddings, or add an experiment in which the embeddings are trained without any annotations from the unseen classes. At minimum, the limitations paragraph should state explicitly that the zero-shot result does not test whether ViCo can transfer to classes absent from its training data.
  3. [Sec. 3.3, Table 3] The co-occurrence counting expands each synset to all of its words, so a region annotated with object synset {man, person, adult, mammal} and attribute {smiling} contributes counts for (man, smiling), (person, smiling), (adult, smiling), and (mammal, smiling). This means WordNet structure is injected into the co-occurrence statistics, not only through the explicit Object-Hypernym type but also through synonym expansion in the other types. The w/o WordNet ablation in Table 3 is a good control and shows that WordNet is not the sole contributor, but it is reported only for the clustering analysis. Since the zero-shot-like and downstream claims are among the paper's main contributions, the authors should provide w/o WordNet variants for those evaluations as well, or justify why the clustering ablation is sufficient to bound the effect.
minor comments (6)
  1. [Sec. 4.2, Learning] The sentence 'The model is trained with a batch size of0.01 for 50 epochs' appears to contain a typo ('of0.01'); please clarify the batch size and the learning rate.
  2. [Sec. 4.2, Model Selection] The best model is selected using seen-class accuracy on the test set; a held-out validation split would be cleaner, since the test set is also used to compute the unseen-class accuracy.
  3. [Sec. 4.2, Fig. 5] The text says results are averaged over four runs and that 7x4x4=112 models were trained; the description of how the 112 models are distributed across splits, embeddings, and runs is easy to misread and should be made explicit.
  4. [Abstract and Table 5 caption] The abstract states that random embeddings 'perform comparably' to learned embeddings on all supervised vision-language tasks, while the Table 5 caption says GloVe+ViCo outperforms random 'for all tasks except Image Captioning.' These statements are consistent only if 'comparably' is interpreted loosely; the wording should be tightened to avoid the appearance that random embeddings match learned embeddings everywhere.
  5. [Sec. 4.4] The analogy evaluation is based on only 30 hand-picked pairings and is reported as a count of correct answers with no candidate-set construction details; it should be presented as qualitative evidence rather than a rigorous evaluation.
  6. [Sec. 3.2, Eq. (2)] The weighting function f(X) is set to 1 for all co-occurrence counts, so rare and frequent co-occurrences contribute equally in the first term; the paper should state whether this choice was evaluated, since GloVe's original weighting is designed to downweight rare pairs.

Circularity Check

0 steps flagged · score 0.0 of 10

No construction-level circularity: ViCo's objective fits co-occurrence counts, and all evaluations use independent labels or external tasks.

full rationale

The paper's central derivation is Eq. (2), a multi-task log-bilinear regression whose targets are the visual co-occurrence counts X_t computed in Sec. 3.3; no downstream evaluation label or clustering label enters this objective. The unsupervised clustering evaluation uses manually assigned coarse and fine categories on 495 VisualGenome words (Sec. 4.1), the zero-shot-like evaluation uses CIFAR-100 class probabilities from Eq. (4) with a classifier trained only on seen classes (Sec. 4.2), and all downstream applications are external benchmarks (Sec. 4.3). None of these quantities is the fitted log-count or a deterministic transform of the fitted parameters alone, so no prediction is equivalent by construction to the model's input. The only author self-citation ([12]) is background context about vision-language models and is not load-bearing. The paper explicitly flags the non-standard nature of its zero-shot-like setup: 'annotations associated with images of unseen categories in VisualGenome or ImageNet may be used to compute word co-occurrences while learning word embeddings' (Sec. 4.2). This is an evaluation-leakage limitation that weakens the strength of the transfer claim, but it is an acknowledged experimental nuance rather than a derivational circularity. Likewise, the small Table 5 gains over GloVe without significance tests are an evidential weakness, not a case of fitting a parameter and renaming it a prediction. Therefore no circular step is established.

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

The central claims rest on the validity of the co-occurrence matrices (free of annotation noise), the multi-task objective (Eq. 2) with its hyperparameters, and the appropriateness of the evaluation protocols. No new physical or conceptual entities are introduced.

free parameters (5)
  • Embedding dimension d = 100 (linear), 200 (select)
    Chosen for compactness/performance trade-off; Table 4 shows 50-200 dimensions perform similarly.
  • Transform type phi_t = linear(100)
    Selected based on clustering and downstream performance; select is the concatenation baseline.
  • Weighting function f(X) = 1 (constant)
    Set to constant instead of GloVe's original weighting; stated in Sec. 3.2.
  • Max-term coefficient = 1 (equal weight)
    Second term in Eq. (2) is added with coefficient 1; no tuning reported.
  • Zero-shot temperature epsilon = learned
    Learnable temperature in Eq. (4).
assumptions (4)
  • standard math Log-bilinear factorization of log co-occurrence counts learns useful semantic embeddings (GloVe).
    Eq. (1) is taken from Pennington et al. and is the basis of the multi-task extension.
  • domain assumption Visual Genome and ImageNet annotations are reliable enough that the derived co-occurrence counts reflect visual co-occurrence.
    Sec. 3.3 uses these annotations without validation; noisy annotations would produce noisy embeddings.
  • domain assumption All words in an annotated synset are mutually co-occurring and equally applicable to the region or image.
    Sec. 3.3 increments counts for all word pairs across all synset pairs, assuming synonyms and hypernyms are interchangeable.
  • ad hoc to paper The zero-shot-like protocol is a valid transfer measure even though unseen CIFAR-100 classes appear in the embedding training data.
    Sec. 4.2 explicitly acknowledges this overlap, making the transfer evaluation a weaker form of zero-shot.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ViCo: Word Embeddings from Visual Co-occurrences." pith.science (2026). https://pith.science/paper/XRKB75KI

@misc{pith2026190808527,
  author       = {Pith},
  title        = {Pith review of: ViCo: Word Embeddings from Visual Co-occurrences},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XRKB75KI}},
  note         = {Machine review of arXiv:1908.08527}
}
read the original abstract

We propose to learn word embeddings from visual co-occurrences. Two words co-occur visually if both words apply to the same image or image region. Specifically, we extract four types of visual co-occurrences between object and attribute words from large-scale, textually-annotated visual databases like VisualGenome and ImageNet. We then train a multi-task log-bilinear model that compactly encodes word "meanings" represented by each co-occurrence type into a single visual word-vector. Through unsupervised clustering, supervised partitioning, and a zero-shot-like generalization analysis we show that our word embeddings complement text-only embeddings like GloVe by better representing similarities and differences between visual concepts that are difficult to obtain from text corpora alone. We further evaluate our embeddings on five downstream applications, four of which are vision-language tasks. Augmenting GloVe with our embeddings yields gains on all tasks. We also find that random embeddings perform comparably to learned embeddings on all supervised vision-language tasks, contrary to conventional wisdom.

Figures

Figures reproduced from arXiv: 1908.08527 by the authors.

Figure 1
Figure 1. Visual co-occurrences are a rich source of informa￾tion for learning word meanings. The figure shows regions an￾notated with words and attributes in an image, and the four types of visual co-occurrences used for learning ViCo embeddings. Attribute co-occurrence between an object in an image re￾gion and the region’s attributes; (2) Attribute-Attribute co￾occurrence of a region; (3) Context co-occurrence which capture… view at source ↗
Figure 2
Figure 2. Log-bilinear models and our multi-task extension. We show loss computation of different approaches for learning word embeddings wi and wj for words i and j. The embeddings are denoted by colored vertical bars. (i) shows GloVe’s log-bilinear model. (ii) is our multi-task extension to learn from multiple co-occurrence matrices. Word embeddings wi and wj are projected into a dedicated space for each co-occurrence type … view at source ↗
Figure 3
Figure 3. Rich sense of relatedness through multiple co￾occurrences. Different notions of word relatedness exist but cur￾rent word embeddings do not provide a way to disentangle those. Since ViCo is learned from multiple types of co-occurrences with dedicated embedding spaces for each (obtained through transfor￾mations φt), it can provide a richer sense of relatedness. The figure shows cosine similarities computed in GloVe, V… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Unsupervised Clustering Analysis. (a,b) Qualitative evaluation with t-SNE: Plots show that ViCo augmented GloVe results in tighter, more homogenous clusters than GloVe. Marker shape encodes the annotated coarse category and color denotes if the word is used more freque…
Figure 5
Figure 5. Figure 5: compares chance performance (1/|U|), random vectors, GloVe, and GloVe+ViCo on four seen/unseen splits [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 49 canonical work pages

  1. [1]

    Spice: Semantic propositional image cap- tion evaluation

    Peter Anderson, Basura Fernando, Mark Johnson, and Stephen Gould. Spice: Semantic propositional image cap- tion evaluation. In ECCV, 2016. 8

    work page 2016

  2. [2]

    Localizing mo- ments in video with natural language

    Lisa Anne Hendricks, Oliver Wang, Eli Shechtman, Josef Sivic, Trevor Darrell, and Bryan Russell. Localizing mo- ments in video with natural language. In ICCV, 2017. 1

    work page 2017

  3. [3]

    Vqa: Visual question answering

    Stanislaw Antol, Aishwarya Agrawal, Jiasen Lu, Margaret Mitchell, Dhruv Batra, C Lawrence Zitnick, and Devi Parikh. Vqa: Visual question answering. In ICCV, 2015. 8

    work page 2015

  4. [4]

    Bullinaria and J

    John A. Bullinaria and J. P. Levy. Extracting semantic rep- resentations from word co-occurrence statistics: a computa- tional study. Behavior research methods, 2007. 2

    work page 2007

  5. [5]

    Microsoft coco captions: Data collection and evaluation server

    Xinlei Chen, Hao Fang, Tsung-Yi Lin, Ramakrishna Vedan- tam, Saurabh Gupta, Piotr Doll ´ar, and C Lawrence Zitnick. Microsoft coco captions: Data collection and evaluation server. arXiv preprint arXiv:1504.00325, 2015. 6

    arXiv 2015

  6. [6]

    Visual dialog

    Abhishek Das, Satwik Kottur, Khushi Gupta, Avi Singh, Deshraj Yadav, Jos´e MF Moura, Devi Parikh, and Dhruv Ba- tra. Visual dialog. In CVPR, 2017. 1

    work page 2017

  7. [7]

    Deerwester, Susan T

    Scott C. Deerwester, Susan T. Dumais, Thomas K. Landauer, George W. Furnas, and Richard A. Harshman. Indexing by latent semantic analysis. JASIS, 1990. 2

    work page 1990

  8. [8]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In CVPR, 2009. 2

    2009

Show all 56 references
  1. [9]

    Bert: Pre-training of deep bidirectional trans- formers for language understanding.CoRR, abs/1810.04805,

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidirectional trans- formers for language understanding.CoRR, abs/1810.04805,

    arXiv

  2. [10]

    Vse++: Improved visual-semantic embeddings

    Fartash Faghri, David J Fleet, Jamie Ryan Kiros, and Sanja Fidler. Vse++: Improved visual-semantic embeddings. BMVC, 2018. 8

    2018

  3. [11]

    Making the v in vqa matter: Elevating the role of image understanding in visual question answer- ing

    Yash Goyal, Tejas Khot, Douglas Summers-Stay, Dhruv Ba- tra, and Devi Parikh. Making the v in vqa matter: Elevating the role of image understanding in visual question answer- ing. In CVPR, 2017. 8

    2017

  4. [12]

    Aligned image-word representations improve induc- tive transfer across vision-language tasks

    Tanmay Gupta, Kevin Shih, Saurabh Singh, and Derek Hoiem. Aligned image-word representations improve induc- tive transfer across vision-language tasks. In ICCV, 2017. 1

    2017

  5. [13]

    Incorporating visual features into word embed- dings: A bimodal autoencoder-based approach

    Mika Hasegawa, Tetsunori Kobayashi, and Yoshihiko Hayashi. Incorporating visual features into word embed- dings: A bimodal autoencoder-based approach. In IWCS,

  6. [14]

    Deep semantic role labeling: What works and whats next

    Luheng He, Kenton Lee, Mike Lewis, and Luke Zettlemoyer. Deep semantic role labeling: What works and whats next. In ACL, 2017. 1

    2017

  7. [15]

    Yu Jiang, Vivek Natarajan, Xinlei Chen, Marcus Rohrbach, Dhruv Batra, and Devi Parikh. Pythia. https:// github.com/facebookresearch/pythia, 2018. 8

    2018

  8. [16]

    Deep visual-semantic align- ments for generating image descriptions

    Andrej Karpathy and Li Fei-Fei. Deep visual-semantic align- ments for generating image descriptions. In CVPR, 2015. 1

    2015

  9. [17]

    Adam: A method for stochastic optimization

    Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. ICLR, 2015. 2, 7

    2015

  10. [18]

    Satwik Kottur, Ramakrishna Vedantam, Jos ´e M. F. Moura, and Devi Parikh. Visual word2vec (vis-w2v): Learning visually grounded word embeddings using abstract scenes. CVPR, 2016. 3, 6, 7

    2016

  11. [19]

    Semeval-2018 task 10: Capturing discriminative attributes

    Alicia Krebs, Alessandro Lenci, and Denis Paperno. Semeval-2018 task 10: Capturing discriminative attributes. In International Workshop on Semantic Evaluation, 2018. 8

    2018

  12. [20]

    Learning multiple layers of features from tiny images

    Alex Krizhevsky and Geoffrey Hinton. Learning multiple layers of features from tiny images. Technical report, Cite- seer, 2009. 5

    2009

  13. [21]

    Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis

    Joseph B Kruskal. Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis. Psychometrika,

  14. [22]

    Word embeddings through hellinger pca

    R ´emi Lebret and Ronan Collobert. Word embeddings through hellinger pca. In EACL, 2014. 2

    2014

  15. [23]

    End-to-end neural coreference resolution

    Kenton Lee, Luheng He, Mike Lewis, and Luke Zettlemoyer. End-to-end neural coreference resolution. EMNLP, 2017. 1

    2017

  16. [24]

    Neural word embedding as implicit matrix factorization

    Omer Levy and Yoav Goldberg. Neural word embedding as implicit matrix factorization. In NIPS, 2014. 2

    2014

  17. [25]

    Microsoft coco: Common objects in context

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll´ar, and C Lawrence Zitnick. Microsoft coco: Common objects in context. In ECCV, 2014. 6

    2014

  18. [26]

    Lund and Curt Burgess

    K. Lund and Curt Burgess. Producing high-dimensional se- mantic spaces from lexical co-ocurrence. 1996. 2

    1996

  19. [27]

    Comprehension- guided referring expressions

    Ruotian Luo and Gregory Shakhnarovich. Comprehension- guided referring expressions. In CVPR, 2017. 1

    2017

  20. [28]

    Flipdial: A generative model for two-way visual dialogue

    Daniela Massiceti, N Siddharth, Puneet K Dokania, and Philip HS Torr. Flipdial: A generative model for two-way visual dialogue. In CVPR, 2018. 1

    2018

  21. [29]

    Corrado, and Jeffrey Dean

    Tomas Mikolov, Kai Chen, Gregory S. Corrado, and Jeffrey Dean. Efficient estimation of word representations in vector space. CoRR, abs/1301.3781, 2013. 2

    2013 arXiv

  22. [30]

    Cor- rado, and Jeffrey Dean

    Tomas Mikolov, Ilya Sutskever, Kai Chen, Gregory S. Cor- rado, and Jeffrey Dean. Distributed representations of words and phrases and their compositionality. In NIPS, 2013. 2

    2013

  23. [31]

    Linguis- tic regularities in continuous space word representations

    Tomas Mikolov, Wen tau Yih, and Geoffrey Zweig. Linguis- tic regularities in continuous space word representations. In HLT-NAACL, 2013. 2, 8

    2013

  24. [32]

    Wordnet: a lexical database for english

    George A Miller. Wordnet: a lexical database for english. ACM, 1995. 2

    1995

  25. [33]

    The big book of concepts

    Gregory Murphy. The big book of concepts . MIT press,

  26. [34]

    The measurement of meaning

    Charles Egerton Osgood, George J Suci, and Percy H Tan- nenbaum. The measurement of meaning. University of Illi- nois press, 1957. 2

    1957

  27. [35]

    Bleu: a method for automatic evaluation of machine translation

    Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. Bleu: a method for automatic evaluation of machine translation. In ACL, 2002. 8

    2002

  28. [36]

    Parikh, Oscar T ¨ackstr¨om, Dipanjan Das, and Jakob Uszkoreit

    Ankur P. Parikh, Oscar T ¨ackstr¨om, Dipanjan Das, and Jakob Uszkoreit. A decomposable attention model for natural lan- guage inference. In EMNLP, 2016. 1

    2016

  29. [37]

    Glove: Global vectors for word representation

    Jeffrey Pennington, Richard Socher, and Christopher Man- ning. Glove: Global vectors for word representation. In EMNLP, 2014. 2, 3, 5

    2014

  30. [38]

    Semi-supervised sequence tagging with bidirectional language models

    Matthew E Peters, Waleed Ammar, Chandra Bhagavatula, and Russell Power. Semi-supervised sequence tagging with bidirectional language models. ACL, 2017. 1

    2017

  31. [39]

    Peters, Mark Neumann, Mohit Iyyer, Matt Gard- ner, Christopher Clark, Kenton Lee, and Luke S

    Matthew E. Peters, Mark Neumann, Mohit Iyyer, Matt Gard- ner, Christopher Clark, Kenton Lee, and Luke S. Zettle- moyer. Deep contextualized word representations. In NAACL-HLT, 2018. 3

    2018

  32. [40]

    Plummer, Paige Kordas, M

    Bryan A. Plummer, Paige Kordas, M. Hadi Kiapour, Shuai Zheng, Robinson Piramuthu, and Svetlana Lazebnik. Con- ditional image-text embedding networks. In ECCV, 2018. 1

    2018

  33. [41]

    Phrase local- ization and visual relationship detection with comprehensive image-language cues

    Bryan A Plummer, Arun Mallya, Christopher M Cervantes, Julia Hockenmaier, and Svetlana Lazebnik. Phrase local- ization and visual relationship detection with comprehensive image-language cues. In ICCV, 2017. 1

    2017

  34. [42]

    Improving language understanding by gener- ative pre-training

    Alec Radford. Improving language understanding by gener- ative pre-training. 2018. 3

    2018

  35. [43]

    Smith, and Yejin Choi

    Hannah Rashkin, Maarten Sap, Emily Allaway, Noah A. Smith, and Yejin Choi. Event2mind: Commonsense infer- ence on events, intents, and reactions. In ACL, 2018. 1

    2018

  36. [44]

    Se- mantic distance and the verification of semantic relations

    Lance J Rips, Edward J Shoben, and Edward E Smith. Se- mantic distance and the verification of semantic relations. Journal of verbal learning and verbal behavior, 1973. 2

    1973

  37. [45]

    Ross and Gregory L

    Brian H. Ross and Gregory L. Murphy. Food for thought: Cross-classification and category organization in a complex real-world domain. Cognitive Psychology, 1999. 2

    1999

  38. [46]

    Bidirectional attention flow for machine comprehension

    Minjoon Seo, Aniruddha Kembhavi, Ali Farhadi, and Han- naneh Hajishirzi. Bidirectional attention flow for machine comprehension. ICLR, 2017. 1

    2017

  39. [47]

    Where to look: Focus regions for visual question answering

    Kevin J Shih, Saurabh Singh, and Derek Hoiem. Where to look: Focus regions for visual question answering. InCVPR,

  40. [48]

    Zettlemoyer, and Ido Dagan

    Gabriel Stanovsky, Julian Michael, Luke S. Zettlemoyer, and Ido Dagan. Supervised open information extraction. In NAACL-HLT, 2018. 1

    2018

  41. [49]

    Learn- ing type-aware embeddings for fashion compatibility

    Mariya I Vasileva, Bryan A Plummer, Krishna Dusad, Shreya Rajpal, Ranjitha Kumar, and David Forsyth. Learn- ing type-aware embeddings for fashion compatibility. In ECCV, 2018. 1

    2018

  42. [50]

    Cider: Consensus-based image description evalua- tion

    Ramakrishna Vedantam, C Lawrence Zitnick, and Devi Parikh. Cider: Consensus-based image description evalua- tion. In CVPR, 2015. 8

    2015

  43. [51]

    Show and tell: A neural image caption gen- erator

    Oriol Vinyals, Alexander Toshev, Samy Bengio, and Du- mitru Erhan. Show and tell: A neural image caption gen- erator. In CVPR, 2015. 8

    2015

  44. [52]

    Learning two-branch neural networks for image-text match- ing tasks

    Liwei Wang, Yin Li, Jing Huang, and Svetlana Lazebnik. Learning two-branch neural networks for image-text match- ing tasks. TPAMI, 2019. 1

    2019

  45. [53]

    Zero-shot recognition via semantic embeddings and knowledge graphs

    Xiaolong Wang, Yufei Ye, and Abhinav Gupta. Zero-shot recognition via semantic embeddings and knowledge graphs. In CVPR, 2018. 1

    2018

  46. [54]

    Mattnet: Modular at- tention network for referring expression comprehension

    Licheng Yu, Zhe Lin, Xiaohui Shen, Jimei Yang, Xin Lu, Mohit Bansal, and Tamara L Berg. Mattnet: Modular at- tention network for referring expression comprehension. In CVPR, 2018. 8

    2018

  47. [55]

    Pythia v0.1: the winning entry to the vqa challenge 2018

    Yu Jiang*, Vivek Natarajan*, Xinlei Chen*, Marcus Rohrbach, Dhruv Batra, and Devi Parikh. Pythia v0.1: the winning entry to the vqa challenge 2018. arXiv preprint arXiv:1807.09956, 2018. 1, 8

    2018 arXiv

  48. [56]

    Bringing semantics into focus using visual abstraction

    C Lawrence Zitnick and Devi Parikh. Bringing semantics into focus using visual abstraction. In CVPR, 2013. 6

    2013

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.