Pith. sign in

REVIEW 3 major objections 6 minor 29 references

Refining Dimensions for Improving Clustering-based Cross-lingual Topic Models

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

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

Pith's one-line read Adding an SVD-based dimension refinement step to the clustering-based topic model pipeline suppresses language-dependent dimensions, producing cross-lingual topics with higher coherence and topic quality.

desk verdict Useful SVD-based preprocessing step for cross-lingual clustering topic models; the empirical gain is real, the mechanism story is thinner than the results. read the letter →

arxiv 2412.12433 v1 pith:JX6QOOCG submitted 2024-12-17 cs.CL cs.IRcs.LG

Chia-Hsuan Chang , Tien-Yuan Huang , Yi-Hang Tsai , Chia-Ming Chang , San-Yih Hwang This is my paper · ORCID
classification cs.CLcs.IRcs.LG
keywords cross-lingualtopicmodelingclustering-basedmodelslanguage-dependentdimensionssingularvaluedecompositionmultilinguallanguageCNPMIcoherencedimensionrefinement
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

Clustering-based topic models work by embedding documents with a language model, clustering the embeddings, and turning each cluster into topic words. The paper argues that in a multilingual corpus these models silently cluster by language rather than by meaning, because a few language-dependent dimensions (LDDs) in the embedding carry language identity and dominate the distances the clusterer sees. The proposed fix replaces the usual dimension-reduction step with an SVD-based refinement: u-SVD uses only the orthonormal left factor of the decomposition, and SVD-LR drops the single dimension most separated by language, identified by a two-sample t-test. On an English–Chinese thesis corpus, English–Chinese news, and English–Japanese product reviews, the refined pipeline yields higher cross-lingual coherence and topic quality than the unrefined pipeline and than two recent cross-lingual topic models. The practical stake is that the fix needs no parallel corpus, bilingual dictionary, or fine-tuning.

What carries the argument

The load-bearing object is the singular value decomposition of the document embedding matrix $E \in \mathbb{R}^{m \times d}$, where $m$ is the number of documents and $d$ is the embedding dimension. Writing $E = U\Sigma V^{\top}$, the paper keeps only the left factor $U$ for u-SVD, which makes every reduced dimension unit length and thereby shrinks the scale advantage of any language-heavy direction; for SVD-LR it keeps $U\Sigma$ and removes the column whose values best separate the two languages, where "best separate" is measured by a two-sample t-test of the per-language value distributions. The mechanism relies on SVD's orthogonal decomposition: if one singular direction carries language identity, the orthogonal remaining directions are freer to encode semantic content. The c-TF-IDF summarization step (the standard class-based TF-IDF of the pipeline) then reconstructs topic words from the refined clusters.

What would settle it

Run the refined pipeline on a multilingual corpus whose singular vectors show language differences spread roughly evenly across many directions (for example, by checking the t-statistics of each singular vector before clustering); if u-SVD and SVD-LR then fall back to the unrefined baselines' CNPMI while a UMAP pipeline does not, the SVD-consolidation assumption is falsified. A direct synthetic test: plant language identity in one planted singular direction and confirm the refinement recovers it, then plant language identity in many orthogonal directions and show the recovery fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that adding an SVD-based dimension refinement step to the clustering-based topic model pipeline neutralizes the language-dependent dimensions (LDDs) that multilingual encoders leave in their document embeddings. With the standard pipeline—embed, reduce with UMAP, cluster with K-means, summarize with c-TF-IDF—these LDDs make distance calculations language-sensitive, so clusters and their topic words come out monolingual and misaligned. The paper shows that SVD of the embedding matrix $E = U\Sigma V^{\top}$ consolidates scattered language information into a concentrated set of singular directions; using $U$ alone (u-SVD) or using $U\Sigma$ with the most language-separated dimension removed (SVD-LR) then lets the clusters form by semantics. In experiments on Airiti, ECNews, and Rakuten Amazon, the refined pipeline generally outperforms the original pipeline, the UMAP/SVD baselines, and the compared state-of-the-art cross-lingual topic models.

Load-bearing premise

The method rests on the geometric assumption that a multilingual embedding's language signal gets packed into one or a few singular directions by SVD, so that rescaling or deleting those directions removes language information without damaging semantics; the paper illustrates this on two datasets but does not prove it, and its own sensitivity analysis shows u-SVD loses effectiveness when more dimensions are retained.

Editorial extensions

If this is right

  • The refined pipeline produces cross-lingual topics on English–Chinese and English–Japanese corpora without parallel corpora, bilingual dictionaries, or fine-tuning, lowering the resource barrier for multilingual topic analysis.
  • On the Airiti and ECNews datasets, both u-SVD and SVD-LR beat the original-embedding, UMAP, UMAP-norm, and pure-SVD baselines in CNPMI and Topic Quality; on Rakuten Amazon they match InfoCTM.
  • SVD-LR stays effective when the reduced dimension grows to 200 or 500, while u-SVD degrades, indicating that SVD-LR is the safer variant when language signal is distributed across small-singular-value directions.
  • The benefit is tied to the encoder: both variants improve strongly over Distilled XLM-R and Cohere embeddings, but give only marginal gains over mBERT, whose cross-lingual representations are weaker to begin with.
  • Qualitatively, the refined topics mix words from both languages under one meaning (for example, a single 'financial market' topic), whereas the unrefined pipeline splits the same concept into separate monolingual topics.

Reading between the lines

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

  • As an extension, u-SVD is best read as a general post-hoc rescaling that equalizes singular directions, so the same move could suppress language-identity leakage in other multilingual embedding tasks such as retrieval or classification.
  • The paper's sensitivity result draws a testable boundary: u-SVD should fail when language information spreads across many small-singular-value directions, and the spectrum of per-direction t-statistics could predict which variant will work before clustering.
  • Because the refinement is agnostic to the clusterer and summarizer, it should combine with neural topic models or other embedding consumers, extending the benefit beyond the K-means + c-TF-IDF pipeline tested here.
  • Since only two-language corpora are evaluated, the natural next test is three or more languages, where SVD-LR's single-dimension removal would need to become a multi-dimension removal.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes a dimension-refinement step for clustering-based cross-lingual topic models. The authors observe that multilingual language model embeddings contain language-dependent dimensions (LDDs) that cause document clustering to separate by language rather than by topic. They introduce two SVD-based refinements: u-SVD, which replaces the reduced representation UΣ with the orthonormal left singular vectors U, and SVD-LR, which removes the single singular dimension with the largest two-sample t-statistic between languages from UΣ. The refined embeddings are then clustered with K-means and summarized with c-TF-IDF. Experiments on Airiti Thesis, ECNews, and Rakuten Amazon, using mBERT, Distilled XLM-R, and Cohere multilingual embeddings, report higher CNPMI and topic quality than the original pipeline and than the Cb-CLTM and InfoCTM baselines on most settings.

Significance. If the empirical claims hold, the paper offers a simple, resource-efficient improvement to cross-lingual topic modeling: unlike dictionary-based or parallel-corpus-based models, the proposed pipeline needs no external alignment resources, and the code is released. The qualitative results in Table 4 are compelling and show that the refined pipeline produces top words from both languages within the same topic. However, the significance is currently limited by two issues: the central mechanism for why the SVD refinement works is asserted rather than directly demonstrated, and the quantitative comparisons are reported only as five-seed averages without variance or significance testing. The method's practical value depends on both points, because the improvements on one dataset (Rakuten Amazon) and with one embedding model (mBERT) are small or negligible.

major comments (3)
  1. [Section 4.1, Tables 2 and 3] The central empirical claim, that the updated pipeline generally outperforms state-of-the-art cross-lingual topic models, is not supported with uncertainty estimates. All results are averages over five seeds, but no standard deviations, confidence intervals, or significance tests are reported. On Rakuten Amazon the margins over InfoCTM are tiny: SVD-LR has CNPMI 0.032 vs. InfoCTM's 0.033, and u-SVD has CNPMI 0.037 vs. 0.033; the TQ margins are also small (0.019 and 0.025 vs. 0.028). In Table 3, with mBERT, neither proposed method clearly improves over the baselines. Without a measure of run-to-run variation, the reader cannot tell whether the reported advantages on Airiti and ECNews are reliable. I ask the authors to report per-seed results, error bars, and appropriate paired significance tests across seeds (or across datasets) for the headline comparisons.
  2. [Section 2.2 and Section 4.3 (Figure 3)] The paper's stated mechanism is that SVD consolidates scattered LDDs into one or a few singular directions, so that u-SVD's rescaling or SVD-LR's removal of one dimension neutralizes language information. This is not directly verified and is partially contradicted by the paper's own sensitivity analysis. Figure 2 shows only the top three t-statistics for two datasets, not the full distribution of language signal across singular dimensions. Figure 3 shows that u-SVD degrades as the reduced dimension r increases from 100 to 500, and the text attributes this to language information in small-singular-value dimensions that u-SVD amplifies. That is evidence that LDDs are not fully concentrated into one or a few leading directions, which makes the SVD-LR choice of removing exactly one dimension seem arbitrary and weakens the generalizability argument. I ask for a direct quantitative test of language-signal concentration, for example the cumulative t-statistic mass or language-classification accuracy as a function of retained singular dimensions, or a revision of the mechanistic claims to match what the experiments actually establish.
  3. [Algorithm 1 and Section 2.2] The SVD-LR design uses a two-sample t-test to select the dimension with the largest mean difference between the two languages. With datasets of tens of thousands of documents, essentially every dimension may be statistically significant, and choosing the maximum t-statistic is not principled as a choice of the dimension most harmful to clustering. The paper does not report how concentrated the t-statistics are, nor whether removing two or more dimensions changes the results. Since the only method-specific free choice in SVD-LR is which dimension(s) to remove, this selection rule should be validated, for instance by ablating over the number of removed dimensions and by reporting the distribution of t-statistics across the reduced dimensions.
minor comments (6)
  1. [Section 2.2] The sentence 'owing to the orthogonal decomposition property of SVD, when one dimension retains language information, the remaining dimensions are more likely to capture other types of information' is informal; please state precisely which orthogonality property is being invoked and how it implies the concentration claim.
  2. [Algorithm 1] The notation U, Σ, VT = SVD(E, r) is not fully defined: it should state whether Σ is the r-by-r truncated diagonal matrix and whether the left singular vectors U are used in their truncated form. In the text, 'UΣ' appears without explicitly defining the truncated matrices.
  3. [Section 4.2 and Limitations] The model is referred to as 'Distilled XLM-R' in Section 4.2 but as 'XLM-R' in the Limitations section; please use one consistent name to avoid ambiguity about whether the distilled or full model was evaluated.
  4. [Table 2 caption] The caption does not state that N=15 and K=50; these are given in the text of Section 4.1, but including them in the table caption would make the table self-contained.
  5. [Figure 3] The figure would be easier to read if the three datasets were distinguished with a legend or explicit curve labels, since the current description relies on the reader inferring the mapping from the text.
  6. [Section 4.4] The qualitative topics in Tables 1 and 4 are described as 'manually sampled'; please state the selection criterion, or note whether all topics were inspected, to reduce concerns about selective reporting.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the refinement methods are transparent preprocessing transformations evaluated by an external coherence metric, not fitted to that metric.

full rationale

I walked the claimed derivation chain: document embeddings are decomposed by SVD, u-SVD retains the orthonormal left singular vectors, and SVD-LR removes the single reduced dimension with the largest two-sample t-statistic between language groups. No equation in the paper is fitted to CNPMI, Diversity, or Topic Quality, and the proposed transformations contain no parameter tuned to the evaluation metric. The only data-dependent choice, which reduced dimension to remove in SVD-LR, is selected by a language-mean t-test and is explicitly stated as part of the method's mechanism rather than a hidden fit to topic coherence. The paper's self-citations are used as a baseline method (Cb-CLTM), a dataset source (Airiti Thesis), and related-work context, but the central claim does not depend on any load-bearing theorem imported from the authors' prior work. The assertion that SVD consolidates scattered language-dependent dimensions is an empirical geometric claim illustrated in Figure 2; it may be under-supported and is challenged by the u-SVD sensitivity results in Section 4.3, but that is a correctness and generalizability risk, not circularity. The empirical comparisons against external baselines and the coherence-based evaluation are self-contained, so the appropriate finding is no significant circularity.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on several empirical assumptions about the geometry of multilingual embeddings, plus a few hyperparameters. No new entities are postulated.

free parameters (3)
  • Reduced dimension size r = 100 (with 200 and 500 in sensitivity)
    Chosen by the authors without a stated criterion; the number of SVD components kept in the refinement step.
  • Number of topics K = 50
    Set to match the InfoCTM setting; controls the number of clusters in KMeans.
  • Number of top words N for evaluation = 15
    Hyperparameter for CNPMI and Diversity; not part of the model itself.
assumptions (4)
  • domain assumption SVD of the document embedding matrix consolidates language-dependent variation into a small number of singular directions.
    Section 2.2 and Figure 2 illustrate this for the datasets used, but the method requires it to hold more generally.
  • domain assumption The dimension with the largest two-sample t-statistic between languages captures most of the language signal.
    Algorithm 1 line 6; assumes a mean shift along one singular direction is the main carrier of language identity.
  • domain assumption Scaling all SVD directions to unit length (u-SVD) reduces LDD influence because LDDs have large singular values.
    Section 2.2; Section 4.3 shows u-SVD degrades when more dimensions are kept, so this assumption is not universally safe.
  • domain assumption Topic coherence measured by CNPMI on the same corpus is a valid proxy for topic interpretability.
    Evaluation section uses comparable corpus statistics; standard in the field but not an external gold standard.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Refining Dimensions for Improving Clustering-based Cross-lingual Topic Models." pith.science (2026). https://pith.science/paper/JX6QOOCG

@misc{pith2026241212433,
  author       = {Pith},
  title        = {Pith review of: Refining Dimensions for Improving Clustering-based Cross-lingual Topic Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JX6QOOCG}},
  note         = {Machine review of arXiv:2412.12433}
}
read the original abstract

Recent works in clustering-based topic models perform well in monolingual topic identification by introducing a pipeline to cluster the contextualized representations. However, the pipeline is suboptimal in identifying topics across languages due to the presence of language-dependent dimensions (LDDs) generated by multilingual language models. To address this issue, we introduce a novel, SVD-based dimension refinement component into the pipeline of the clustering-based topic model. This component effectively neutralizes the negative impact of LDDs, enabling the model to accurately identify topics across languages. Our experiments on three datasets demonstrate that the updated pipeline with the dimension refinement component generally outperforms other state-of-the-art cross-lingual topic models.

Figures

Figures reproduced from arXiv: 2412.12433 by the authors.

Figure 1
Figure 1. Two resultant scenarios of clustering-based [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Top 3 language-dependent dimensions, sorted by t-statistic values, for original embeddings and embeddings [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Sensitivity analysis of u-SVD and SVD-LR [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Top 3 language-dependent dimensions, sorted by t-statistic values, for original embeddings and embeddings [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 15 canonical work pages

  1. [1]

    online" 'onlinestring :=

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

  2. [2]

    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 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Federico Bianchi, Silvia Terragni, Dirk Hovy, Debora Nozza, and Elisabetta Fersini. 2021. Cross-lingual Contextualized Topic Models with Zero-shot Learning . In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics : Main Volume , pages 1676--1683, Online . Association for Computational Linguistics

    work page 2021

  4. [4]

    David M Blei, Andrew Y Ng, and Michael I Jordan. 2003. Latent dirichlet allocation. J. Mach. Learn. Res., 3(Jan):993--1022

    work page 2003

  5. [5]

    Chia-Hsuan Chang and San-Yih Hwang. 2021. https://doi.org/10.1007/s10115-021-01555-7 A word embedding-based approach to cross-lingual topic modeling . Knowledge and Information Systems, 63(6):1529--1555

    work page doi:10.1007/s10115-021-01555-7 2021

  6. [6]

    Chia-Hsuan Chang, San-Yih Hwang, and Tou-Hsiang Xui. 2018. https://doi.org/10.1109/BigDataCongress.2018.00010 Incorporating Word Embedding into Cross-Lingual Topic Modeling . In 2018 IEEE International Congress on Big Data ( BigData Congress ) , pages 17--24, San Francisco, CA, USA . IEEE

    arXiv 2018

  7. [7]

    Chia-Ming Chang, Chia-Hsuan Chang, and San-Yih Hwang. 2020. https://doi.org/10.1002/pra2.229 Employing word mover's distance for cross-lingual plagiarized text detection . In Proceedings of the Association for Information Science and Technology , volume 57, page e229

    work page doi:10.1002/pra2.229 2020

  8. [8]

    Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzm \'a n, Edouard Grave, Myle Ott, Luke Zettlemoyer, and Veselin Stoyanov. 2020. https://doi.org/10.18653/v1/2020.acl-main.747 Unsupervised Cross-lingual Representation Learning at Scale . In Proceedings of the 58th Annual Meeting of the Association for Comp...

    doi:10.18653/v1/2020.acl-main.747 2020

Show all 29 references
  1. [9]

    Crain, Ke Zhou, Shuang-Hong Yang, and Hongyuan Zha

    Steven P. Crain, Ke Zhou, Shuang-Hong Yang, and Hongyuan Zha. 2012. https://doi.org/10.1007/978-1-4614-3223-4_5 Dimensionality Reduction and Topic Modeling : From Latent Semantic Indexing to Latent Dirichlet Allocation and Beyond . In Charu C. Aggarwal and ChengXiang Zhai, edi...

    2012 doi

  2. [10]

    Dumais, George W

    Scott Deerwester, Susan T. Dumais, George W. Furnas, Thomas K. Landauer, and Richard Harshman. 1990. https://doi.org/10.1002/(SICI)1097-4571(199009)41:6<391::AID-ASI1>3.0.CO;2-9 Indexing by latent semantic analysis . Journal of the American Society for Information Science, 41(...

    1990 doi

  3. [11]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT : Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Huma...

    2019

  4. [12]

    Adji B Dieng, Francisco J R Ruiz, and David M Blei. 2020. https://doi.org/10.1162/tacl_a_00325 Topic Modeling in Embedding Spaces . Transactions of the Association for Computational Linguistics, 8:439--453

    2020 doi

  5. [13]

    Philipp Dufter and Hinrich Sch \"u tze. 2020. https://doi.org/10.18653/v1/2020.emnlp-main.358 Identifying Elements Essential for BERT 's Multilinguality . In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing ( EMNLP ) , pages 4423--4437, On...

    2020 doi

  6. [14]

    Maarten Grootendorst. 2022. BERTopic : Neural topic modeling with a class-based TF-IDF procedure

    2022

  7. [15]

    Shudong Hao and Michael J Paul. 2020. https://doi.org/10.1162/coli_a_00369 An Empirical Study on Crosslingual Transfer in Probabilistic Topic Models . Comput. Linguist., 46(1):95--134

    2020 doi

  8. [16]

    Yuening Hu, Ke Zhai, Vladimir Eidelman, and Jordan Boyd-Graber . 2014. https://doi.org/10.3115/v1/P14-1110 Polylingual Tree-Based Topic Models for Translation Domain Adaptation . In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics ( Volum...

    2014 doi

  9. [17]

    Ehsan Kamalloo, Xinyu Zhang, Odunayo Ogundepo, Nandan Thakur, David Alfonso-Hermelo , Mehdi Rezagholizadeh, and Jimmy Lin. 2023. https://doi.org/10.48550/arXiv.2305.06300 Evaluating Embedding APIs for Information Retrieval

    2023 doi arXiv

  10. [18]

    Jey Han Lau, David Newman, and Timothy Baldwin. 2014. Machine reading tea leaves: Automatically evaluating topic coherence and topic model quality. In Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics , pages 530--539

    2014

  11. [19]

    Jind r ich Libovick \'y , Rudolf Rosa, and Alexander Fraser. 2020. https://doi.org/10.18653/v1/2020.findings-emnlp.150 On the language neutrality of pre-trained multilingual representations . In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 1663-...

    2020 doi

  12. [20]

    Leland McInnes, John Healy, and James Melville. 2020. https://doi.org/10.48550/arXiv.1802.03426 UMAP : Uniform Manifold Approximation and Projection for Dimension Reduction

    2020 doi arXiv

  13. [21]

    David Mimno, Hanna M Wallach, Jason Naradowsky, David A Smith, and Andrew McCallum. 2009. https://doi.org/10.3115/1699571.1699627 Polylingual topic models . In Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing Volume 2 - EMNLP '09 , volume ...

    2009

  14. [22]

    Tiziano Piccardi and Robert West. 2021. https://doi.org/10.1145/3442381.3449805 Crosslingual Topic Modeling with WikiPDA . In Proceedings of the Web Conference 2021 , pages 3032--3041, Ljubljana Slovenia . ACM

    2021

  15. [24]

    Telmo Pires, Eva Schlinger, and Dan Garrette. 2019 b . https://doi.org/10.18653/v1/P19-1493 How multilingual is multilingual BERT ? In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4996--5001, Florence, Italy. Association for Co...

    2019 doi

  16. [25]

    Nils Reimers and Iryna Gurevych. 2020. https://doi.org/10.18653/v1/2020.emnlp-main.365 Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillation . In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing ( EMNLP ) , pages...

    2020 doi

  17. [26]

    Suzanna Sia, Ayush Dalmia, and Sabrina J Mielke. 2020. https://doi.org/10.18653/v1/2020.emnlp-main.135 Tired of Topic Models ? Clusters of Pretrained Word Embeddings Make for Fast and Good Topics too! In Proceedings of the 2020 Conference on Empirical Methods in Natural Langua...

    2020 doi

  18. [27]

    Xiaobao Wu, Xinshuai Dong, Thong Nguyen, Chaoqun Liu, Liang-Ming Pan, and Anh Tuan Luu. 2023. https://doi.org/10.1609/aaai.v37i11.26612 InfoCTM : A mutual information maximization perspective of cross-lingual topic modeling . In Proceedings of the Thirty-Seventh AAAI Conferenc...

    2023 doi

  19. [28]

    Michelle Yuan, Benjamin Van Durme, and Jordan L Ying. 2018. Multilingual Anchoring : Interactive Topic Modeling and Alignment Across Languages . In Advances in Neural Information Processing Systems , volume 31. Curran Associates, Inc

    2018

  20. [29]

    Zihan Zhang, Meng Fang, Ling Chen, and Mohammad Reza Namazi Rad. 2022. https://doi.org/10.18653/v1/2022.naacl-main.285 Is Neural Topic Modelling Better than Clustering ? An Empirical Study on Clustering with Contextual Embeddings for Topics . In Proceedings of the 2022 Confere...

    2022 doi

  21. [30]

    Wei Zhao, Steffen Eger, Johannes Bjerva, and Isabelle Augenstein. 2021. https://doi.org/10.18653/v1/2021.starsem-1.22 Inducing Language-Agnostic Multilingual Representations . In Proceedings of * SEM 2021: The Tenth Joint Conference on Lexical and Computational Semantics , pag...

    2021 doi

Pith tools

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