Pith. sign in

REVIEW 2 major objections 4 minor 63 references

Discriminative Topic Mining via Category-Name Guided Text Embedding

T0 review · 2 major / 4 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 CatE mines discriminative topics from category names alone: retrieved terms belong to and only belong to the provided category, and the same embeddings improve weakly supervised classification and lexical entailment direction.

desk verdict A genuinely new task framing with strong empirical topic-mining results, but the generative derivation is loose and the BLESS experiment leaves its supervision setup unstated. read the letter →

arxiv 1908.07162 v2 pith:KOY7NGBD submitted 2019-08-20 cs.CL cs.IR

Yu Meng , Jiaxin Huang , Guangyuan Wang , Zihan Wang , Chao Zhang , Yu Zhang , Jiawei Han This is my paper · ORCID
classification cs.CLcs.IR
keywords topicminingdiscriminativeanalysiscategory-nameguidedembeddingtextworddistributionalspecificityweakly-supervisedclassificationlexicalentailmentCatE
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

CatE proposes a task the authors call discriminative topic mining: given only a set of category names, retrieve from a corpus the terms that semantically belong to and only belong to those categories. The method starts with category names as anchor embeddings, learns an embedding space in which the categories are pulled apart, and iteratively selects representative words by combining embedding similarity with a learned word-specificity score. On NYT and Yelp corpora, human judges rate the retrieved top-10 terms as belonging to the right category in 0.913 to 1.000 of cases, outperforming the compared topic-modeling baselines. The same category-discriminative embeddings improve weakly supervised text classification over unsupervised embeddings, and the learned specificity score identifies hypernym-hyponym direction on the BLESS benchmark at higher accuracy than the compared unsupervised baselines.

What carries the argument

The central mechanism is a joint embedding space in which every word is represented by an input vector, a context vector, and a scalar specificity $\kappa_w$; all vectors are constrained to the unit sphere, and the distribution of contexts around a word is modeled as a von Mises-Fisher distribution with concentration $\kappa_w$. A high $\kappa_w$ means the word is semantically specific because it appears in few contexts. Category names are embedded as anchor vectors, and the training objective includes a KL-divergence term that pushes each selected representative word's category distribution toward a one-hot vector for its own category, alongside the usual local- and global-context objectives. Representative words are then chosen by ranking candidates by the product of cosine similarity to the category anchor and low specificity, while requiring $\kappa_w$ to exceed the category name's $\kappa$. The iterative loop, train then select then augment then retrain, is what turns a few category names into a separated embedding space.

What would settle it

On a held-out set of documents with known categories, compute the rewritten topic loss $-\sum_{c\in C}\sum_{w\in S_c} p(c\mid w) + \mathrm{const}$ and compare it with the original negative log-likelihood $-\log p(d\mid c_d)$ from Eq. (2); if the two differ by more than a constant across documents, the generative-model interpretation of CatE is not the implemented objective, and the empirical success rests on the heuristic self-training loop instead.

Watch

Extended reading notes

Core claim

The paper's central claim is that a short list of category names, with no labeled documents and no seed words beyond the names, is sufficient supervision for mining discriminative topics. CatE models text generation as conditioned on user categories, rewrites the topic term in the corpus likelihood as a sum over per-word category assignments, and implements it as an embedding model in which category embeddings act as anchors while each word carries a learned concentration parameter measuring how specific the word is. Representative words are selected by a rank product of similarity to the category anchor and specificity, subject to being more specific than the category name, then fed back into training to separate the categories further. The paper argues that the learned specificity encodes the distributional inclusion hypothesis: a specific term appears in a narrower set of contexts than its hypernym. The reported payoffs are mean accuracy (MACC) of 0.972 and 0.967 on NYT-Location and NYT-Topic, 0.913 and 1.000 on Yelp-Food and Yelp-Sentiment, improved weakly supervised classification over the compared unsupervised embeddings, and 0.895 accuracy on BLESS direction identification.

Load-bearing premise

The method relies on the assumption that a document's fit to a category can be computed purely from the product of its words' per-word category probabilities, and that the iterative self-selection of representative words keeps improving rather than amplifying its own errors.

Editorial extensions

If this is right

  • Users can extract distinctive vocabularies from a corpus by typing category names, with no labeled documents and no need to choose the number of topics.
  • Swapping the input embeddings of the WeSTClass weakly supervised classifier from unsupervised embeddings to CatE improves micro-F1 on the tested datasets, e.g., NYT-Location from 0.533 to 0.655.
  • The learned specificity value acts as an unsupervised lexical-entailment direction signal, identifying the more specific word in BLESS hypernym-hyponym pairs at 0.895 accuracy.
  • Because retrieved terms must be more specific than the category name, topic results can be ordered coarse-to-fine by $\kappa$ rather than by an opaque topic-word probability.
  • The method requires no document labels, making it applicable to new corpora where annotation is expensive and the user's only input is a list of categories of interest.

Reading between the lines

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

  • A natural stress test is to run CatE with deliberately nested or overlapping category names, such as food and dessert or Europe and France; performance on the more specific category would show whether the anchor-plus-specificity selection truly enforces exclusive membership or merely separates well-chosen anchors.
  • One could apply CatE recursively: once a category's representative terms are retrieved, feed the more specific terms as new category names to mine subcategories and build a taxonomy from corpus statistics alone.
  • The specificity parameter $\kappa_w$ is a free by-product of training and could be used in other lexical-semantics tasks, such as definition ranking or technical-term extraction, where hypernym-like generality is useful and labeled data are scarce.
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

2 major / 4 minor

Summary. The paper proposes a new task, discriminative topic mining, in which the user supplies only a set of category names and the system retrieves terms that belong to and only belong to each category. The proposed method, CatE, jointly learns word, document, and category embeddings by combining a topic-level term, a global document-context term, and a local context term, while also learning a per-word distributional specificity scalar kappa. Category representative words are selected by ranking candidates by embedding similarity to the category name and by low-to-high kappa, and the procedure iteratively expands the representative sets. The paper evaluates CatE on NYT and Yelp benchmarks for topic mining, as an embedding input to WeSTClass for weakly supervised classification, and on BLESS for unsupervised lexical entailment direction identification. The reported results show large gains over baselines in topic mining (MACC up to 1.000), classification (e.g., Micro-F1 0.838 vs. 0.738 for fastText on Yelp-Sentiment), and LE direction (0.895 vs. 0.861 for SLQS). Source code is provided.

Significance. If the reported results hold, CatE is a practically useful method for weakly supervised text analysis: it turns minimal user guidance (category names only) into discriminative embeddings and topic lists, and it is shown to improve a downstream weakly supervised classifier across several domains. The paper's distributional specificity mechanism is a novel component with an independent validation in the BLESS direction-identification task, and the availability of source code is a concrete strength. The main significance is therefore empirical: CatE appears to be an effective recipe for category-name-guided topic mining. The theoretical framing, however, is considerably weaker than the empirical contribution, and the BLESS experiment lacks important reporting on what supervision was used; these issues need to be resolved before the central claims can be taken at face value.

major comments (2)
  1. [Section 3.2, Eq. (2) through Eq. (6)] placeholder
  2. [Section 5.4 and Algorithm 1] placeholder
minor comments (4)
  1. [Table 3] placeholder
  2. [Eq. (12)] placeholder
  3. [Section 5.3] placeholder
  4. [Section 3.2] placeholder

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: CatE's category names are external supervision, its objectives are not defined in terms of the evaluation labels, and the BLESS and MACC benchmarks are external to the training procedure.

full rationale

CatE's inputs are a text corpus and a set of category names, and its outputs are retrieved category-representative terms and learned embeddings. The retrieval criterion in Eq. (12) is based on embedding cosine similarity to category embeddings and on the learned distributional-specificity parameter kappa, while the category embeddings and kappa are optimized through the objectives in Eqs. (2), (6), (7), and (8). None of these objectives is defined in terms of the human MACC labels, the document classification labels, or the BLESS hypernymy labels used for evaluation. MACC is an external human judgment of whether retrieved terms semantically belong to the user-provided categories, which is the evaluation of the task definition rather than a training signal. BLESS is an external benchmark not used in training. The iterative expansion in Algorithm 1 is a self-training or bootstrapping procedure initialized from the externally provided category names, but it does not inject any test or evaluation labels into the loss, so the predictions are not statistically forced by the fit. The paper's Section 3.2 algebraic rewrite of L_topic is mathematically imprecise: the equality to -sum over categories and words of p(c|w) plus a constant does not follow as written from the preceding product decomposition, and the implemented L_topic in Eq. (6) is a KL-divergence regularization rather than a direct transcription of that derivation. Also, Section 5.4 reports the BLESS result without stating what category names, if any, were supplied to CatE, so the 0.895 accuracy may not clearly demonstrate the category-name-guided claim. These are correctness and reporting concerns, not circularity. The self-citations in the paper, such as [30], [31], [32], and [44], are used for design choices, downstream evaluation infrastructure, or preprocessing tools, and they are not the load-bearing justification for the central derivation or the empirical claims. Therefore, no significant circularity is present, and the score is 0.

Assumptions & free parameters 3 free parameters · 5 assumptions · 1 invented entities

The central claim rests mainly on the corpus-dependent training signal and the distributional inclusion hypothesis. The paper's own derivation from the likelihood to the regularizer is not valid as written, so the method's grounding is partially heuristic. No completely new physical or formal entities are posited; the only new latent is the per-word specificity scalar kappa_w, which has external support from the BLESS benchmark.

free parameters (3)
  • kappa_w per-word specificity scalar = learned, initialized to 1; category-name values around 0.527-0.566 in Table 6
    Central parameter for selection (Eq. 12) and entailment direction; fitted to corpus without direct supervision. No identifiability or sensitivity analysis is provided.
  • Embedding vectors for words, documents, and categories = learned, dimension p=100
    The log-linear model parameters in Eqs. (3) through (8) are optimized on the corpus; they are the internal representation rather than a first-principles constant.
  • Hyperparameters p=100, h=5, k=5, max_iter=10, negative samples = p=100, h=5, k=5, max_iter=10
    Chosen in Section 5.1 and applied to all methods; no sensitivity analysis is reported.
assumptions (5)
  • ad hoc to paper p(d|cd) is proportional to the product over words in d of p(cd|w)
    The core link from category names to document likelihood is asserted in Section 3.2, not derived. The subsequent rewrite of L_topic to a sum of p(c|w) does not follow, and the implemented Eq. (6) is a KL heuristic.
  • domain assumption Each document belongs to exactly one user category
    The generative process in Eq. (1) draws each document conditioned on one of n categories; datasets are filtered to single-labeled documents in Section 5.1, which limits applicability to multi-topic documents.
  • domain assumption Distributional inclusion hypothesis: hyponyms occur in a subset of the contexts of their hypernyms
    Definition 2 and Section 4.1 rely on this hypothesis to interpret the learned scalar kappa as word distributional specificity; the BLESS experiment provides empirical support but the hypothesis itself is assumed.
  • standard math Continuous softmax over the unit sphere converges to the von Mises Fisher distribution
    Theorem 1 is a standard limiting result; it validates the probabilistic reading of the softmax but does not prove that the trained kappa corresponds to lexical specificity.
  • domain assumption AutoPhrase phrase extraction is accurate and phrases can be treated as atomic words
    Section 5.1 uses AutoPhrase to extract quality phrases that are treated as single words during embedding training; if phrase extraction fails for category names or representative phrases, retrieval quality degrades.
invented entities (1)
  • kappa_w, distributional specificity scalar for each word independent evidence
    purpose: Quantify how specific a word's meaning is; filter candidate representative words and determine lexical entailment direction
    The BLESS lexical entailment direction experiment in Section 5.4 is an external benchmark not used to fit the model; kappa alone gives 0.895 accuracy against unsupervised baselines. However, the mapping from kappa to specificity is not formally derived, and no identifiability result is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Discriminative Topic Mining via Category-Name Guided Text Embedding." pith.science (2026). https://pith.science/paper/KOY7NGBD

@misc{pith2026190807162,
  author       = {Pith},
  title        = {Pith review of: Discriminative Topic Mining via Category-Name Guided Text Embedding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KOY7NGBD}},
  note         = {Machine review of arXiv:1908.07162}
}
read the original abstract

Mining a set of meaningful and distinctive topics automatically from massive text corpora has broad applications. Existing topic models, however, typically work in a purely unsupervised way, which often generate topics that do not fit users' particular needs and yield suboptimal performance on downstream tasks. We propose a new task, discriminative topic mining, which leverages a set of user-provided category names to mine discriminative topics from text corpora. This new task not only helps a user understand clearly and distinctively the topics he/she is most interested in, but also benefits directly keyword-driven classification tasks. We develop CatE, a novel category-name guided text embedding method for discriminative topic mining, which effectively leverages minimal user guidance to learn a discriminative embedding space and discover category representative terms in an iterative manner. We conduct a comprehensive set of experiments to show that CatE mines high-quality set of topics guided by category names only, and benefits a variety of downstream applications including weakly-supervised classification and lexical entailment direction identification.

Figures

Figures reproduced from arXiv: 1908.07162 by the authors.

Figure 1
Figure 1. Word Distributional Specificity. 4.4 Selecting Category Representative Words Finally, the learned distributional specificity can be used to impose the constraint that class representative words should belong to the category. Specifically, a category representative word must have higher distributional specificity than the category name. However, we also want to avoid selecting too specific terms as category repre￾sen… view at source ↗
Figure 2
Figure 2. Dataset statistics. 5 EXPERIMENTS 5.1 Experiment Setup Datasets. We use two datasets, the New York Times annotated corpus (NYT) [42], the recently released Yelp Dataset Challenge (Yelp) 4 . NYT and Yelp each has two sets of categories: NYT: topic and location; Yelp: food type and sentiment. For NYT, we first select the major categories (with more than 100 documents) from topics and locations, and then collect docume… view at source ↗
Figure 3
Figure 3. Discriminative embedding space training for topic mining. [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

63 extracted references · 57 canonical work pages

  1. [1]

    David Andrzejewski and Xiaojin Zhu. 2009. Latent Dirichlet Allocation with Topic-in-Set Knowledge. In HLT-NAACL

    work page 2009

  2. [2]

    Marco Baroni and Alessandro Lenci. 2011. How we BLESSed distributional semantic evaluation. In EMNLP

    work page 2011

  3. [3]

    Kayhan Batmanghelich, Ardavan Saeedi, Karthik Narasimhan, and Sam Gersh- man. 2016. Nonparametric spherical topic modeling with word embeddings. In ACL. 537

    work page 2016

  4. [4]

    David Blei and John Lafferty. 2006. Correlated topic models. In NIPS. 147

    work page 2006

  5. [5]

    David M Blei and Jon D Mcauliffe. 2008. Supervised topic models. In NIPS. 121–128

    work page 2008

  6. [6]

    Blei, Andrew Y

    David M. Blei, Andrew Y. Ng, and Michael I. Jordan. 2003. Latent Dirichlet Allocation. In NIPS

    work page 2003

  7. [7]

    Piotr Bojanowski, Edouard Grave, Armand Joulin, and Tomas Mikolov. 2017. En- riching Word Vectors with Subword Information. Transactions of the Association for Computational Linguistics 5 (2017), 135–146

    2017

  8. [8]

    Ming-Wei Chang, Lev-Arie Ratinov, Dan Roth, and Vivek Srikumar. 2008. Impor- tance of Semantic Representation: Dataless Classification. In AAAI

    work page 2008

Show all 63 references
  1. [9]

    Chaitanya Chemudugunta, Padhraic Smyth, and Mark Steyvers. 2008. Combining concept hierarchies and statistical topic models. In CIKM. 1469–1470

    2008

  2. [10]

    Rajarshi Das, Manzil Zaheer, and Chris Dyer. 2015. Gaussian lda for topic models with word embeddings. In ACL. 795–804

    2015

  3. [11]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In NAACL-HLT

    2019

  4. [12]

    Shallue, Mohammad Norouzi, Andrew M

    Bhuwan Dhingra, Christopher J. Shallue, Mohammad Norouzi, Andrew M. Dai, and George E. Dahl. 2018. Embedding Text in Hyperbolic Spaces. In TextGraphs@NAACL-HLT

    2018

  5. [13]

    Dieng, Francisco J

    Adji B. Dieng, Francisco J. R. Ruiz, and David M. Blei. 2019. Topic Modeling in Embedding Spaces. ArXiv abs/1907.04907 (2019)

    2019 arXiv

  6. [14]

    Zhicheng Dou, Ruihua Song, and Ji-Rong Wen. 2007. A large-scale evaluation and analysis of personalized search strategies. In WWW

    2007

  7. [15]

    Foster and Roland Kuhn

    George F. Foster and Roland Kuhn. 2007. Mixture-Model Adaptation for SMT. In WMT@ACL

    2007

  8. [16]

    Gallagher, Kyle Reing, David C

    Ryan J. Gallagher, Kyle Reing, David C. Kale, and Greg Ver Steeg. 2017. Anchored Correlation Explanation: Topic Modeling with Minimal Domain Knowledge. TACL (2017)

    2017

  9. [17]

    Octavian-Eugen Ganea, Gary Bécigneul, and Thomas Hofmann. 2018. Hyperbolic Entailment Cones for Learning Hierarchical Embeddings. In ICML

    2018

  10. [18]

    Thomas L Griffiths, Michael I Jordan, Joshua B Tenenbaum, and David M Blei

  11. [19]

    Thomas Hofmann. 1999. Probabilistic Latent Semantic Indexing. In SIGIR

    1999

  12. [20]

    Jiaxin Huang, Yiqing Xie, Yu Meng, Jiaming Shen, Yunyi Zhang, and Jiawei Han

  13. [21]

    Jagadeesh Jagarlamudi, Hal Daumé, and Raghavendra Udupa. 2012. Incorporating Lexical Priors into Topic Models. In EACL

    2012

  14. [22]

    Yoon Kim. 2014. Convolutional Neural Networks for Sentence Classification. In EMNLP

    2014

  15. [23]

    Simon Lacoste-Julien, Fei Sha, and Michael I Jordan. 2009. DiscLDA: Discrimina- tive learning for dimensionality reduction and classification. In NIPS. 897–904

    2009

  16. [24]

    Jey Han Lau, David Newman, and Timothy Baldwin. 2014. Machine Reading Tea Leaves: Automatically Evaluating Topic Coherence and Topic Model Quality. In EACL

    2014

  17. [25]

    Wei Li and Andrew McCallum. 2006. Pachinko allocation: DAG-structured mixture models of topic correlations. In ICML. 577–584

    2006

  18. [26]

    Yang Liu, Zhiyuan Liu, Tat-Seng Chua, and Maosong Sun. 2015. Topical Word Embeddings. In AAAI

    2015

  19. [27]

    Laurens van der Maaten and Geoffrey Hinton. 2008. Visualizing data using t-SNE. Journal of machine learning research 9, Nov (2008), 2579–2605

    2008

  20. [28]

    Xian-Ling Mao, Zhao-Yan Ming, Tat-Seng Chua, Si Li, Hongfei Yan, and Xiaoming Li. 2012. SSHLDA: a semi-supervised hierarchical topic model. In EMNLP. 800– 809

    2012

  21. [29]

    Qiaozhu Mei, Xuehua Shen, and ChengXiang Zhai. 2007. Automatic labeling of multinomial topic models. In KDD

    2007

  22. [30]

    Yu Meng, Jiaxin Huang, Guangyuan Wang, Chao Zhang, Honglei Zhuang, Lance Kaplan, and Jiawei Han. 2019. Spherical Text Embedding. In NeurIPS

    2019

  23. [31]

    Yu Meng, Jiaming Shen, Chao Zhang, and Jiawei Han. 2018. Weakly-Supervised Neural Text Classification. In CIKM

    2018

  24. [32]

    Yu Meng, Jiaming Shen, Chao Zhang, and Jiawei Han. 2019. Weakly-Supervised Hierarchical Text Classification. In AAAI

    2019

  25. [33]

    Corrado, and Jeffrey Dean

    Tomas Mikolov, Ilya Sutskever, Kai Chen, Gregory S. Corrado, and Jeffrey Dean

  26. [34]

    Dat Quoc Nguyen, Richard Billingsley, Lan Du, and Mark Johnson. 2015. Im- proving topic models with latent feature word representations. TACL 3 (2015), 299–313

    2015

  27. [35]

    Kim Anh Nguyen, Maximilian Köper, Sabine Schulte im Walde, and Ngoc Thang Vu. 2017. Hierarchical Embeddings for Hypernymy Detection and Directionality. In EMNLP

    2017

  28. [36]

    Maximilian Nickel and Douwe Kiela. 2017. Poincaré Embeddings for Learning Hierarchical Representations. In NIPS

    2017

  29. [37]

    Maximilian Nickel and Douwe Kiela. 2018. Learning Continuous Hierarchies in the Lorentz Model of Hyperbolic Geometry. In ICML

    2018

  30. [38]

    Jeffrey Pennington, Richard Socher, and Christopher D. Manning. 2014. Glove: Global Vectors for Word Representation. InEMNLP

    2014

  31. [39]

    Daniel Ramage, David Hall, Ramesh Nallapati, and Christopher D Manning. 2009. Labeled LDA: A supervised topic model for credit attribution in multi-labeled corpora. In EMNLP. 248–256

    2009

  32. [40]

    Michal Rosen-Zvi, Thomas Griffiths, Mark Steyvers, and Padhraic Smyth. 2004. The author-topic model for authors and documents. In UAI. 487–494

    2004

  33. [41]

    Timothy N Rubin, America Chambers, Padhraic Smyth, and Mark Steyvers. 2012. Statistical topic models for multi-label document classification. Machine learning 88, 1-2 (2012), 157–208

    2012

  34. [42]

    Evan Sandhaus. 2008. The New York Times Annotated Corpus

    2008

  35. [43]

    Enrico Santus, Alessandro Lenci, Qin Lu, and Sabine Schulte im Walde. 2014. Chasing Hypernyms in Vector Spaces with Entropy. In EACL

    2014

  36. [44]

    Voss, and Jiawei Han

    Jingbo Shang, Jialu Liu, Meng Jiang, Xiang Ren, Clare R. Voss, and Jiawei Han

  37. [45]

    Yangqiu Song and Dan Roth. 2014. On Dataless Hierarchical Text Classification. In AAAI

    2014

  38. [46]

    Jian Tang, Meng Qu, and Qiaozhu Mei. 2015. PTE: Predictive Text Embedding through Large-scale Heterogeneous Text Networks. In KDD

    2015

  39. [47]

    Alexandru Tifrea, Gary Bécigneul, and Octavian-Eugen Ganea. 2019. Poincaré Glove: Hyperbolic Word Embeddings. In ICLR

    2019

  40. [48]

    Gomez, Lukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. Attention Is All You Need. In NIPS

    2017

  41. [49]

    Ivan Vulic, Daniela Gerz, Douwe Kiela, Felix Hill, and Anna Korhonen. 2017. HyperLex: A Large-Scale Evaluation of Graded Lexical Entailment.Computational Linguistics (2017)

    2017

  42. [50]

    Guoyin Wang, Chunyuan Li, Wenlin Wang, Yizhe Zhang, Dinghan Shen, Xinyuan Zhang, Ricardo Henao, and Lawrence Carin. 2018. Joint Embedding of Words and Labels for Text Classification. In ACL

    2018

  43. [51]

    Weir, and Diana McCarthy

    Julie Weeds, David J. Weir, and Diana McCarthy. 2004. Characterising Measures of Lexical Distributional Similarity. In COLING

    2004

  44. [52]

    Bruce Croft

    Xing Wei and W. Bruce Croft. 2006. LDA-based document models for ad-hoc retrieval. In SIGIR

    2006

  45. [53]

    Hongteng Xu, Wenlin Wang, Wei Liu, and Lawrence Carin. 2018. Distilled wasserstein learning for word embedding and topic modeling. In NIPS. 1716– 1725

    2018

  46. [54]

    Guangxu Xun, Vishrawas Gopalakrishnan, Fenglong Ma, Yaliang Li, Jing Gao, and Aidong Zhang. 2016. Topic discovery for short texts using word embeddings. In ICDM. 1299–1304

    2016

  47. [55]

    Guangxu Xun, Yaliang Li, Jing Gao, and Aidong Zhang. 2017. Collaboratively Improving Topic Discovery and Word Embeddings by Coordinating Global and Local Contexts. In KDD

    2017

  48. [56]

    Smola, and Ed- uard H

    Zichao Yang, Diyi Yang, Chris Dyer, Xiaodong He, Alexander J. Smola, and Ed- uard H. Hovy. 2016. Hierarchical Attention Networks for Document Classification. In HLT-NAACL

    2016

  49. [57]

    Sadler, Michelle T

    Chao Zhang, Fangbo Tao, Xiusi Chen, Jiaming Shen, Meng Jiang, Brian M. Sadler, Michelle T. Vanni, and Jiawei Han. 2018. TaxoGen: Constructing Topical Concept Taxonomy by Adaptive Term Embedding and Clustering. In KDD

    2018

  50. [58]

    Yu Zhang, Frank F Xu, Sha Li, Yu Meng, Xuan Wang, Qi Li, and Jiawei Han. 2019. HiGitClass: Keyword-Driven Hierarchical Classification of GitHub Repositories. In ICDM

    2019

  51. [59]

    Maayan Zhitomirsky-Geffet and Ido Dagan. 2005. The Distributional Inclusion Hypotheses and Lexical Entailment. In ACL

    2005

  52. [2004]

    Hierarchical topic models and the nested Chinese restaurant process. In NIPS. 17–24

  53. [2013]

    Distributed Representations of Words and Phrases and their Composition- ality. In NIPS

  54. [2018]

    IEEE Transactions on Knowledge and Data Engineering 30 (2018), 1825–1837

    Automated Phrase Mining from Massive Text Corpora. IEEE Transactions on Knowledge and Data Engineering 30 (2018), 1825–1837

    2018

  55. [2020]

    Guiding Corpus-based Set Expansion by Auxiliary Sets Generation and Co-Expansion. In WWW

Pith tools

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