Pith. sign in

REVIEW 3 major objections 5 minor 35 references

Bilingual Lexicon Induction with Semi-supervision in Non-Isometric Embedding Spaces

T0 review · 3 major / 5 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 Relaxing the isometry assumption between embedding spaces yields state-of-the-art bilingual lexicon induction, especially for etymologically distant languages.

desk verdict Solid semi-supervised BLI method with a useful GH-distance diagnostic, undermined by an internally inconsistent SOTA claim that must be fixed. read the letter →

arxiv 1908.06625 v1 pith:FLUAW7P6 submitted 2019-08-19 cs.CL cs.LG

Barun Patra , Joel Ruben Antony Moniz , Sarthak Garg , Matthew R. Gormley , Graham Neubig This is my paper · ORCID
classification cs.CLcs.LG
keywords bilinguallexiconinductionsemi-supervisedlearningwordembeddingsorthogonalityassumptionGromov-Hausdorffdistanceadversarialtraininghubnessfilteringcross-lingualmapping
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

The paper sets out to test a hidden premise in bilingual lexicon induction (BLI): that the two word-embedding spaces can be aligned by an orthogonal rotation, i.e., that they are isometric. It argues, with a Gromov-Hausdorff distance estimate, that this isometry assumption weakens as languages become etymologically distant, and that standard orthogonal-mapping methods therefore underperform precisely there. To fix this, the paper proposes BLISS, a semi-supervised framework that jointly optimizes an adversarial distribution-matching loss over all unlabeled embeddings, a supervised alignment loss on a small seed dictionary, and a weak orthogonality constraint that encourages, but does not force, the mapping to be orthogonal. On the MUSE benchmark the paper reports state-of-the-art accuracy on 15 of 18 language pairs, with the clearest gains on non-isometric pairs, and shows that even 50 seed word pairs stabilize training that purely unsupervised methods fail to complete. If correct, this means minimal human-provided word pairs plus a flexible mapping can make bilingual lexicon induction work across typologically diverse languages, not just close relatives.

What carries the argument

The load-bearing machinery is the three-term joint loss $L = L_{W|D} + L_{W|S} + L_{W|O}$. $L_{W|D}$ is an adversarial distribution-matching loss (a GAN) that uses all unlabeled word embeddings from both languages; $L_{W|S}$ is a supervised similarity loss on a small seed dictionary, instantiated either as cosine similarity (BLISS(M)) or as the CSLS retrieval criterion (BLISS(R)); $L_{W|O}$ is a weak orthogonality loss that maximizes the similarity between $x$ and $W^T W x$, encouraging $W^T W \approx I$ without hard projection. The second key object is the Gromov-Hausdorff distance estimate, computed through the bottleneck distance between Vietoris-Rips persistence diagrams of the two embedding spaces, which the paper uses as an a-priori diagnostic of how isometric the spaces are. The claim is that jointly optimizing these three losses lets the model bend the mapping where the spaces are non-isometric and stay close to orthogonal where they are.

What would settle it

On the same language pairs, train BLISS(R) with the weak orthogonality term removed and with a hard orthogonal projection in its place; if the soft-loss version does not clearly beat the hard-projection baseline on high-GH-distance pairs (e.g., en-ko, en-hi, en-zh), the paper's explanation for its gains collapses. A second check: compute GH-distance estimates on a held-out set of language pairs and test whether their rank order predicts the size of BLISS's advantage over RCSLS; if not, the diagnostic is not doing the predictive work the paper claims.

Watch

Extended reading notes

Core claim

The paper's central claim is that the orthogonality assumption used in most BLI methods—that one embedding space can be rotated rigidly onto the other—does not hold uniformly, and that relaxing it improves translation quality. The supporting evidence is the Gromov-Hausdorff distance between embedding spaces, which the authors approximate with bottleneck distance and show to be low for etymologically close pairs (ru-uk, en-fr) and high for distant pairs (en-ko, en-hi), while correlating strongly with BLI accuracy. BLISS then operationalizes the relaxation by adding a soft autoencoding loss, $L_{W|O} = -\frac{1}{|X|} \sum_{x_i} f_a(x_i, W^T W x_i)$, to the supervised and adversarial losses, so $W$ is nudged toward orthogonality only when the data supports it. The claimed result is state-of-the-art performance on 15 of 18 MUSE language pairs and on the VecMap dataset, with the largest margins on high-GH-distance pairs, plus the empirical finding that a small amount of supervision dramatically stabilizes the GAN-based training. This is a claim about both method efficacy and about the geometry of embedding spaces: the failure of isometry is real, measurable, and exploitable.

Load-bearing premise

The load-bearing premise is that the three losses in equation (6) can be jointly optimized so that the soft orthogonality penalty is outweighed by the supervised and adversarial losses exactly when the spaces are non-isometric, without causing training instability or harmful local optima; the paper provides no theoretical guarantee of this balance, only empirical results.

Editorial extensions

If this is right

  • A small seed dictionary (tens to hundreds of word pairs) is enough to convert a failing unsupervised method into a working bilingual lexicon induction system, including for distant language pairs such as English–Chinese and English–Korean.
  • Purely unsupervised methods are at a systematic disadvantage for fine-grained word choice: they can align semantic clusters but miss distinctions within clusters (e.g., aunt vs. grandmother); adding supervision inside those clusters corrects this.
  • The Gromov-Hausdorff distance between embedding spaces can serve as an a-priori predictor of how hard a language pair will be for orthogonal-mapping methods, since it correlates strongly with method accuracy and with the empirical non-orthogonality of the learned mapping.
  • Near-orthogonal rather than strictly orthogonal mappings are sufficient for state-of-the-art BLI; a hard orthogonality projection is not necessary and is sensitive to hyperparameters.
  • Providing even minimal supervision stabilizes the adversarial training procedure, avoiding the failure-to-converge cases seen in unsupervised GAN-based methods, particularly when embeddings are low-quality or the language pair is distant.
  • The paper's own error analysis (polysemy and antonymy) suggests that the remaining failures are not geometric but semantic; a testable extension is to combine BLISS with sense-aware or contextualized embeddings to see whether the non-isometry error shrinks further.
  • Because the method works with 50–500 seed pairs, it could serve as a practical warm-start for low-resource machine translation systems, where such lexicons are often available from WordNet-style resources or Wikipedia links.

Reading between the lines

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

  • If the GH-distance diagnostic is as predictive as reported, it could be used to decide a priori how much seed supervision a language pair needs, and to extrapolate the method's likely success to resource-poor languages where no seed dictionary exists.
  • The weak orthogonality constraint resembles a regularizer rather than a geometric prior; a natural extension would be to learn the trade-off between $L_{W|O}$ and the other losses per language pair rather than fixing the coefficient, which the paper does not explore.
  • The paper's own error analysis (polysemy and antonymy) suggests that the remaining failures are not geometric but semantic; a testable extension is to combine BLISS with sense-aware or contextualized embeddings to see whether the non-isometry error shrinks further.
  • Because the method works with 50–500 seed pairs, it could serve as a practical warm-start for low-resource machine translation systems, where such lexicons are often available from WordNet-style resources or Wikipedia links.
  • If the GH-distance diagnostic is as predictive as reported, it could be used to decide a priori how much seed supervision a language pair needs, and to extrapolate the method's likely success to resource-poor languages where no seed dictionary exists.
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 / 5 minor

Summary. The paper studies bilingual lexicon induction (BLI) and makes two main contributions. First, it proposes a quantitative diagnostic based on the Gromov–Hausdorff (GH) distance to measure how close two embedding spaces are to being isometric, and it reports correlations between this distance and BLI accuracy. Second, it proposes BLISS, a semi-supervised framework that combines an adversarial distribution-matching loss, a supervised loss on a small seed lexicon, and a weak orthogonality (cyclic consistency) loss, together with iterative Procrustes refinement and a hubness filtering technique. Experiments on MUSE and VecMap datasets compare two instantiations of BLISS against supervised, unsupervised, and semi-supervised baselines. The abstract claims state-of-the-art results on 15 of 18 MUSE language pairs and improved stability from supervision, especially for etymologically distant language pairs.

Significance. If the empirical claims are substantiated, the paper would make a useful contribution: it offers a pre-hoc isometry diagnostic that correlates with BLI performance, and a semi-supervised objective that appears to outperform pure supervised and pure unsupervised baselines on several language pairs, particularly when supervision is scarce. The release of code and the inclusion of a stability analysis are also strengths that support reproducibility. However, the headline state-of-the-art claim is currently supported by inconsistent counts across the abstract, introduction, and results tables, and the relationship between the weak orthogonality loss and the paper's critique of the orthogonality assumption is not directly demonstrated. These issues block acceptance in the current form, but they are addressable in revision.

major comments (3)
  1. [Abstract; Section 1; Section 4.3; Tables 3 and 4] The central empirical claim is not supported by the reported experiments. The abstract states 'state of the art results on 15 of 18 language pairs on the MUSE dataset'; Section 1 states 'Our best model outperforms the state-of-the-art on 10 of 16 language pairs on the MUSE datasets'; Section 4.3 reports that BLISS(R) gives the best accuracy across all baselines for 6 of 10 language pairs (Table 3), and Table 4 reports 2 VecMap pairs. No table or appendix reports 18 or 16 language pairs; Table 3 contains 10 MUSE direction pairs. Since the SOTA claim in the abstract is the paper's headline result, the authors must either supply the full pair-level results for 18 (or 16) pairs or revise the claim to match the 10 pairs actually presented. The distinction matters because Table 3 also shows GeoMM outperforming BLISS(R) on de-en and en-zh, so the choice of pair set and counting convention changes the verdict.
  2. [Section 4.4; Table 2] The supporting comparison with RCSLS is also inconsistently counted. Section 4.4 says 'BLISS(R) substantially outperforms RCSLS on 6 of 9 language pairs' and points to Table 2. Reading the BLISS(R) and RCSLS rows in Table 2 for the 9 language pairs where RCSLS has non-star values gives BLISS(R) higher on 7 pairs, lower on 1 (en-ru), and tied on 1 (es-fr). Please reconcile this count and make explicit whether ties are counted as outperformance. As it stands, the evidence for the 'particularly well when non-isometric' claim depends on a count that does not match the table.
  3. [Section 3.2, Eqs. (5)-(6); Table 7] The weak orthogonality constraint in Eq. (5) encourages W^T W approximately equal to the identity and is therefore a softened version of the hard orthogonality assumption that motivates the paper. The claim that the supervised and adversarial losses outweigh LW|O exactly when the spaces are non-isometric is an empirical assumption, not demonstrated in the paper. Table 7 shows that the autoencoding loss alone is sensitive to its scaling in the unsupervised setting, but the paper does not report the analogous analysis for BLISS, nor does it report ||I - W^T W|| for the learned BLISS mappings. Without this evidence, the reader cannot tell whether BLISS is actually learning non-orthogonal mappings for high-GH pairs or simply reverting to an orthogonal map with a small regularizer. Please add a sensitivity analysis and, if possible, a direct measurement of the learned W's deviation from orthogonality across language pairs.
minor comments (5)
  1. [Abstract and Section 2] The name is spelled 'Gromov-Hausdroff' in the abstract and 'Gromov Hausdorff' elsewhere; use 'Gromov-Hausdorff' consistently.
  2. [Eq. (3)] The second sum is written over xi in Y but uses the target variable; this should be written as a sum over y_i in Y (or equivalent) to avoid a type error.
  3. [Table 2] The column layout mixes per-language-pair values with correlation columns, and the meaning of '*' entries is not defined in the caption. Please add a caption that explains the row and column structure.
  4. [Table 3] Several accuracy advantages are below one point (e.g., en-es 84.3 vs. 84.1); the paper should state whether these differences are stable across runs and ideally report variance or significance.
  5. [Conclusion] The phrase 'orthogonal gains' is ambiguous and should be rephrased, for example as 'further gains' or 'additional gains'.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the GH-distance analysis and BLISS objectives are computed from external embeddings and benchmark dictionaries, not from the outputs they predict.

full rationale

The derivation chain is self-contained. The isometry estimate in Section 2 is computed from mean-centered, unit-normed embeddings via the Bottleneck-distance approximation of the Gromov-Hausdorff distance (Eq. 2), independently of BLISS's learned mapping, and its correlations in Table 2 are against accuracies of externally defined baselines and of BLISS after evaluation. BLISS's supervised loss (Eq. 4) uses ground-truth seed pairs, its distribution-matching loss (Eq. 3) uses unlabeled embeddings, and its cyclic loss (Eq. 5) is a regularizer, not a quantity later reported as the model's output. Benchmark accuracies in Tables 3-5 are computed by CSLS nearest-neighbor retrieval against held-out dictionaries, so none of the reported predictions re-enters the fitted objectives. The paper contains no load-bearing self-citation: the only self-citation (Qi et al. 2018) concerns pre-trained embeddings for NMT and is contextual. The abstract's '15 of 18' state-of-the-art claim is inconsistent with the introduction's '10 of 16' and with Table 3's 6 of 10, but that is a factual reporting inconsistency, not a circular derivation. No self-definitional step, fitted-input-as-prediction, or uniqueness-imported-via-self-citation is present.

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

The central claim depends on a handful of hyperparameters (loss scaling, hubness threshold, training size) and four domain assumptions. No new entities are introduced. The free parameters are not fitted to test labels, but they are tuned on validation and could affect the reported gains.

free parameters (4)
  • Autoencoding loss scaling weight (lambda_O) = not explicitly stated; Table 7 explores 1e-2, 1e-3, 1e-4
    Weight of LW|O in Eq. (6); chosen by validation performance. It controls the degree of orthogonality enforcement and directly affects the method's central behavior.
  • Hubness filtering threshold = 20
    Words in the target domain with more than 20 source-domain nearest neighbors are filtered out during iterative Procrustes refinement. Set empirically.
  • Number of words for GAN training = 75000
    Top 75000 words per language are used for adversarial distribution matching; this is a hyperparameter that affects training scale.
  • Learning rate and decay schedule = lr=0.1, decay=0.98 per round, halved on no improvement
    Standard optimization hyperparameters; reported in Appendix A.4.
assumptions (4)
  • domain assumption Mean-centering and unit-normalizing embeddings makes orthogonal transforms equivalent to isometric transforms between the spaces.
    Stated in the footnote in Section 2. This justifies using orthogonal mappings when discussing isometry.
  • standard math The bottleneck distance between first-order Vietoris-Rips persistence diagrams lower bounds the Gromov-Hausdorff distance.
    Invoked in Section 2 and Appendix A.1, citing Chazal et al. (2009). Used to approximate GH distance.
  • domain assumption CSLS is an effective nearest-neighbor retrieval metric for bilingual lexicon induction.
    Adopted from Lample et al. (2018), Section 3.3. The method relies on CSLS for both retrieval and iterative refinement.
  • domain assumption The unsupervised CSLS metric correlates well with ground-truth translation accuracy and is a valid stopping criterion.
    Stated in Section 4.2. The paper cites Lample et al. (2018) for this correlation but does not provide quantitative evidence in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Bilingual Lexicon Induction with Semi-supervision in Non-Isometric Embedding Spaces." pith.science (2026). https://pith.science/paper/FLUAW7P6

@misc{pith2026190806625,
  author       = {Pith},
  title        = {Pith review of: Bilingual Lexicon Induction with Semi-supervision in Non-Isometric Embedding Spaces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FLUAW7P6}},
  note         = {Machine review of arXiv:1908.06625}
}
read the original abstract

Recent work on bilingual lexicon induction (BLI) has frequently depended either on aligned bilingual lexicons or on distribution matching, often with an assumption about the isometry of the two spaces. We propose a technique to quantitatively estimate this assumption of the isometry between two embedding spaces and empirically show that this assumption weakens as the languages in question become increasingly etymologically distant. We then propose Bilingual Lexicon Induction with Semi-Supervision (BLISS) --- a semi-supervised approach that relaxes the isometric assumption while leveraging both limited aligned bilingual lexicons and a larger set of unaligned word embeddings, as well as a novel hubness filtering technique. Our proposed method obtains state of the art results on 15 of 18 language pairs on the MUSE dataset, and does particularly well when the embedding spaces don't appear to be isometric. In addition, we also show that adding supervision stabilizes the learning procedure, and is effective even with minimal supervision.

Figures

Figures reproduced from arXiv: 1908.06625 by the authors.

Figure 1
Figure 1. A toy dataset demonstrating the shortcomings of unsupervised distribution matching. Fig. a) and b) show [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Training Stability of different language pairs (en-de), (en-ru), (en-zh) [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Fraction of errors coming from polysemy in the source/target side and antonymy, for the language pairs [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 25 canonical work pages

  1. [1]

    Mikel Artetxe, Gorka Labaka, and Eneko Agirre. 2016. Learning principled bilingual mappings of word embeddings while preserving monolingual invariance. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 2289--2294

    work page 2016

  2. [2]

    Mikel Artetxe, Gorka Labaka, and Eneko Agirre. 2017. http://aclweb.org/anthology/P17-1042 Learning bilingual word embeddings with (almost) no bilingual data . In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 451--462, Vancouver, Canada. Association for Computational Linguistics

    work page 2017

  3. [3]

    Mikel Artetxe, Gorka Labaka, and Eneko Agirre. 2018 a . Generalizing and improving bilingual word embedding mappings with a multi-step framework of linear transformations. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence (AAAI-18)

    work page 2018

  4. [4]

    Mikel Artetxe, Gorka Labaka, and Eneko Agirre. 2018 b . A robust self-learning method for fully unsupervised cross-lingual mappings of word embeddings. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), volume 1, pages 789--798

    work page 2018

  5. [5]

    Piotr Bojanowski, Edouard Grave, Armand Joulin, and Tomas Mikolov. 2017. Enriching word vectors with subword information. Transactions of the Association for Computational Linguistics, 5:135--146

    2017

  6. [6]

    Hailong Cao, Tiejun Zhao, Shu Zhang, and Yao Meng. 2016. A distribution-based model to learn bilingual word embeddings. In Proceedings of COLING 2016, the 26th International Conference on Computational Linguistics: Technical Papers, pages 1818--1827

    work page 2016

  7. [7]

    Fr \'e d \'e ric Chazal, David Cohen-Steiner, Leonidas J Guibas, Facundo M \'e moli, and Steve Y Oudot. 2009. Gromov-hausdorff stable signatures for shapes using persistence. In Computer Graphics Forum, volume 28, pages 1393--1403. Wiley Online Library

    work page 2009

  8. [8]

    Georgiana Dinu and Marco Baroni. 2014. http://arxiv.org/abs/1412.6568 Improving zero-shot learning by mitigating the hubness problem . volume abs/1412.6568

    arXiv 2014

Show all 35 references
  1. [9]

    Herbert Edelsbrunner and Dmitriy Morozov. 2013. Persistent homology: theory and practice

    2013

  2. [10]

    Manaal Faruqui and Chris Dyer. 2014. Improving vector space word representations using multilingual correlation. In Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, pages 462--471

    2014

  3. [11]

    Edouard Grave, Armand Joulin, and Quentin Berthet. 2018. Unsupervised alignment of embeddings with wasserstein procrustes. arXiv preprint arXiv:1805.11222

    2018 arXiv

  4. [12]

    Aria Haghighi, Percy Liang, Taylor Berg-Kirkpatrick, and Dan Klein. 2008. http://www.aclweb.org/anthology/P/P08/P08-1088 Learning bilingual lexicons from monolingual corpora . In Proceedings of ACL-08: HLT, pages 771--779, Columbus, Ohio. Association for Computational Linguistics

    2008

  5. [13]

    Ann Irvine and Chris Callison-Burch. 2013. http://www.aclweb.org/anthology/W13-2233 Combining bilingual and comparable corpora for low resource machine translation . In Proceedings of the Eighth Workshop on Statistical Machine Translation, pages 262--270, Sofia, Bulgaria. Asso...

    2013

  6. [14]

    Pratik Jawanpuria, Arjun Balgovind, Anoop Kunchukuttan, and Bamdev Mishra. 2019. Learning multilingual word embeddings in latent metric space: a geometric approach. Transaction of the Association for Computational Linguistics (TACL), 7:107--120

    2019

  7. [15]

    Armand Joulin, Piotr Bojanowski, Tomas Mikolov, Herv \'e J \'e gou, and Edouard Grave. 2018. Loss in translation: Learning bilingual word mapping with a retrieval criterion. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 2979--2984

    2018

  8. [16]

    Yova Kementchedjhieva, Sebastian Ruder, Ryan Cotterell, and Anders S gaard. 2018. Generalizing procrustes analysis for better bilingual dictionary induction. In Proceedings of the 22nd Conference on Computational Natural Language Learning, pages 211--220

    2018

  9. [17]

    Alexandre Klementiev, Ivan Titov, and Binod Bhattarai. 2012. Inducing crosslingual distributed representations of words. pages 1459--1474

    2012

  10. [18]

    Guillaume Lample, Alexis Conneau, Marc'Aurelio Ranzato, Ludovic Denoyer, and Hervé Jégou. 2018. https://openreview.net/forum?id=H196sainb Word translation without parallel data . In International Conference on Learning Representations

    2018

  11. [19]

    Thang Luong, Hieu Pham, and Christopher D Manning. 2015. Bilingual word representations with monolingual quality in mind. In Proceedings of the 1st Workshop on Vector Space Modeling for Natural Language Processing, pages 151--159

    2015

  12. [20]

    Tomas Mikolov, Quoc V Le, and Ilya Sutskever. 2013 a . Exploiting similarities among languages for machine translation

    2013

  13. [21]

    Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg S Corrado, and Jeff Dean. 2013 b . http://papers.nips.cc/paper/5021-distributed-representations-of-words-and-phrases-and-their-compositionality.pdf Distributed representations of words and phrases and their compositionality . In C....

    2013

  14. [22]

    Jeffrey Pennington, Richard Socher, and Christopher Manning. 2014. Glove: Global vectors for word representation. In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), pages 1532--1543

    2014

  15. [23]

    Ye Qi, Devendra Sachan, Matthieu Felix, Sarguna Padmanabhan, and Graham Neubig. 2018. http://www.aclweb.org/anthology/N18-2084 When and why are pre-trained word embeddings useful for neural machine translation? In Proceedings of the 2018 Conference of the North American Chapte...

    2018

  16. [24]

    Milo s Radovanovi \'c , Alexandros Nanopoulos, and Mirjana Ivanovi \'c . 2010. Hubs in space: Popular nearest neighbors in high-dimensional data. volume 11, pages 2487--2531

    2010

  17. [25]

    Sebastian Ruder. 2017. http://arxiv.org/abs/1706.04902 A survey of cross-lingual embedding models . CoRR, abs/1706.04902

    2017 arXiv

  18. [26]

    Samuel L Smith, David HP Turban, Steven Hamblin, and Nils Y Hammerla. 2017. Offline bilingual word vectors, orthogonal transformations and the inverted softmax

    2017

  19. [27]

    Anders S gaard, Sebastian Ruder, and Ivan Vuli \'c . 2018. On the limitations of unsupervised bilingual dictionary induction. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), volume 1, pages 778--788

    2018

  20. [28]

    Min Xiao and Yuhong Guo. 2014. Distributed word representation learning for cross-lingual dependency parsing. In Proceedings of the Eighteenth Conference on Computational Natural Language Learning, pages 119--129

    2014

  21. [29]

    Chao Xing, Dong Wang, Chao Liu, and Yiye Lin. 2015. Normalized word embedding and orthogonal transform for bilingual word translation. In Proceedings of the 2015 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technolog...

    2015

  22. [30]

    Meng Zhang, Yang Liu, Huanbo Luan, and Maosong Sun. 2017 a . Adversarial training for unsupervised bilingual lexicon induction. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), volume 1, pages 1959--1970

    2017

  23. [31]

    Meng Zhang, Yang Liu, Huanbo Luan, and Maosong Sun. 2017 b . Earth mover's distance minimization for unsupervised bilingual lexicon induction. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 1934--1945

    2017

  24. [32]

    Yuan Zhang, David Gaddy, Regina Barzilay, and Tommi Jaakkola. 2016. Ten pairs to tag-multilingual pos tagging via coarse mapping between embeddings. Association for Computational Linguistics

    2016

  25. [33]

    Will Y Zou, Richard Socher, Daniel Cer, and Christopher D Manning. 2013. Bilingual word embeddings for phrase-based machine translation. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1393--1398

    2013

  26. [34]

    URL: " 'urlintro :=

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

  27. [35]

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