Pith. sign in

REVIEW 2 major objections 5 minor 53 references

Leveraging Just a Few Keywords for Fine-Grained Aspect Detection Through Weakly Supervised Co-Training

T0 review · 2 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 A weakly supervised student-teacher method using about 30 seed words per aspect beats prior weakly supervised aspect detectors by 14.1 average F1 points and closes 62 percent of the supervised gap.

desk verdict A clean student-teacher trick for weakly supervised aspect detection that mostly earns its 14-point F1 claim; the co-training bonus is real but self-referential and secondary. read the letter →

arxiv 1909.00415 v1 pith:OSM575WZ submitted 2019-09-01 cs.LG cs.CLstat.ML

Giannis Karamanolakis , Daniel Hsu , Luis Gravano This is my paper · ORCID
classification cs.LGcs.CLstat.ML
keywords weaklysupervisedlearningaspectdetectionseedwordsstudent-teachermodelsco-trainingknowledgedistillationopinionminingmultilingual
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 asks whether a few descriptive keywords (seed words) per aspect can replace the thousands of fine-grained labels normally needed to train aspect detectors for product and restaurant reviews. It argues that earlier weakly supervised methods underuse seed words, and proposes a student-teacher pipeline: a bag-of-seed-words classifier labels unlabeled review segments, then a neural student is trained to imitate those soft labels. The student can exploit non-seed words and pre-trained embeddings, so it generalizes beyond the teacher. An iterative co-training step estimates the quality of each seed word from the student's own predictions, which improves both models. Across six product-review domains and six languages of restaurant reviews, the method reports an average gain of 14.1 F1 points over previous weakly supervised approaches and closes 62 percent of the gap to fully supervised models.

What carries the argument

The load-bearing mechanism is a two-model loop: a bag-of-seed-words teacher that scores a segment by the counts of seed words it contains, and a student network trained to match the teacher's soft predictions while also seeing the full segment text. A second component is the seed-word quality vector $\hat{z}_j$ from Equation (4), which reweights seed words according to agreement between the seed word's presence and the student's predicted aspect. This reweighted teacher (Equation 3) turns noisy seed words into graded supervision. The loop is the co-training step that propagates improvements back and forth.

What would settle it

Take a domain from the paper, run the first co-training round, and on the training segments that contain a given seed word compare the student's hard prediction with the true aspect label; if for several seed words the student is wrong on a majority of those segments, then the next teacher round should decrease rather than increase F1. The same design could be tested by adding deliberately misleading seed words (for example, 'great' assigned to every aspect) and checking whether the reported two-to-three-round improvements reverse.

Watch

Extended reading notes

Core claim

The central claim is that seed words are best treated not as anchors for initializing topic embeddings but as direct, noisy supervision signals. The teacher is a softmax over per-aspect seed-word counts (Equation 1), producing soft aspect labels for unlabeled segments; the student is any embedding-based classifier trained with the cross-entropy distillation objective (Equation 2). Because the student sees all words, it learns to associate non-seed words with aspects and outperforms the teacher. Iterative co-training then reweights each seed word by its estimated predictive quality (Equation 4), using the student's hard predictions as a proxy for ground truth, and this improved teacher in turn trains a better student. The paper reports consistent gains over earlier weakly supervised methods and a 62 percent reduction of the performance gap to fully supervised models on the 12 datasets.

Load-bearing premise

The iterative step assumes that the student's own hard predictions on unlabeled training segments are accurate enough to serve as ground truth when estimating each seed word's quality; if the student is systematically wrong on segments containing a seed word, the reweighted teacher will amplify that error.

Editorial extensions

If this is right

  • Detectors for new product domains can be trained from roughly 30 seed words per aspect without collecting segment-level labels.
  • The same recipe works across languages, since the teacher is language-agnostic and the student can use multilingual pretrained embeddings.
  • Stronger student backbones directly improve the weakly supervised result: BERT-based students give the best reported F1.
  • Noisy seed words need not be cleaned by hand; iterative co-training recovers much of their value within two to three rounds.
  • Weak supervision with seed words becomes a practical alternative when fully supervised models are infeasible, cutting the performance gap to supervised training by 62 percent.

Reading between the lines

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

  • The student-teacher recipe is task-agnostic: any classification problem whose classes can be described by a few keywords, such as document tagging or review usefulness prediction, could use the same teacher and distillation loop.
  • The quality estimate in Equation (4) is a self-training bootstrap, so its success likely depends on seed words being reasonably distinctive; on domains with heavy lexical overlap between aspects, confidence-weighted or soft student predictions may be more stable.
  • A natural testable extension is to replace the hard argmax used in seed-quality estimation with teacher-student agreement or prediction confidence, which could further stabilize co-training for low-resource languages.
  • The reported gap-closing of 62 percent is measured against a fixed label budget; combining this method with a small number of labeled segments might push weakly supervised detectors past the best supervised baselines in data-scarce settings.
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 / 5 minor

Summary. The paper studies fine-grained aspect detection with no ground-truth labels, using a small set of seed words per aspect. It introduces a teacher model, a bag-of-seed-words classifier that produces soft aspect labels, and a student network (with BOW, word2vec, attention, or BERT encoders) trained to mimic the teacher via a distillation objective. It then proposes iterative co-training (ISWD), where seed-word quality weights are estimated from the student's hard predictions and fed back into a reweighted teacher. The method is evaluated on six OPOSUM product-review domains and six SemEval restaurant-review languages. The paper reports consistent improvements over prior weakly supervised baselines, an average gain of 14.1 F1 points for Student-BERT, and a 62% reduction of the gap to fully supervised models. Ablations, including a no-seed-words student variant and iterative co-training results, are also reported.

Significance. If the results hold, this is a practically useful advance: it demonstrates that a few dozen seed words per aspect can supervise neural aspect detectors that substantially outperform previous weakly supervised methods and close a large portion of the gap to supervised models, across domains and languages. The core single-round comparison is well controlled: Student-ATT and MATE-unweighted share the same seed words, pre-trained embeddings, segment encoder, and classifier, isolating the effect of using the teacher's soft labels as supervision. The consistent direction of improvement across all 12 datasets is a strong point, and the paper is transparent that it does not beat fully supervised models and that non-English restaurant experiments use a single co-training round due to limited training data. The main weaknesses are the unvalidated self-training mechanism for seed-quality estimation and the lack of variance and significance reporting.

major comments (2)
  1. [Section 3.3, Algorithm 1, Eq. (4)] The co-training claim in Section 3.3 rests on a self-training bootstrap. Equation (4) estimates each seed word's quality z_j from the student's hard predictions t_i on segments containing that seed word, but the student is itself trained to mimic the teacher (Eq. (2)), and the teacher is defined entirely by the seed words. Thus z_j is a function of the model's own beliefs rather than an independent estimate of seed-word quality; a systematic teacher error (for example, a wrong seed-word-to-aspect assignment) will be upweighted by Eq. (3), and the stopping rule based on student-teacher disagreement cannot distinguish convergence to a correct solution from convergence to a self-consistent fixed point. The modest gains in Table 5 and Figure 4 are therefore not sufficient evidence for the paper's third contribution that iterative co-training 'copes with noisy seed words.' I ask for a concrete validation of the mechanism: correlate the estimated z_j with oracle seed-word precisions computed from held-out labels, compare teachers trained with estimated weights against one trained with oracle weights, and if possible include a control in which the student's hard labels are replaced by an independent noisy-label estimator.
  2. [Section 4.1 and Section 4.2, Tables 3 and 4] The central quantitative claim is reported without variance or significance information. The text says results are averaged over 5 runs with the configuration that achieves the best validation performance, but only mean micro-F1 is shown in Tables 3 and 4, with no standard deviations, per-run values, or paired tests across datasets. Because the paper claims the method 'consistently outperforms' previous approaches by a large average margin, the absence of error bars or a significance test makes it hard to assess the stability of the improvements. The direction of improvement is uniform across all 12 datasets in the single-round comparison, which is reassuring, but the paper should report standard deviations and, ideally, a paired test (for example, a Wilcoxon signed-rank test over the 12 datasets) for the main comparisons.
minor comments (5)
  1. [Section 4.2, paragraph on prior weakly supervised models] The sentence 'Student-ATT outperforms MATE-unweighted (by 36.3% in product reviews and by 52.2% in restaurant reviews)' is arithmetically inconsistent with Tables 3 and 4. For restaurant reviews, the relative increase of Student-ATT over MATE-unweighted is (44.9 - 26.1)/26.1 = 72.0%, not 52.2%; 52.2% is the relative increase over MATE (29.5). Please correct the sentence or explicitly state that the restaurant comparison is against MATE.
  2. [Section 4.2, final paragraph] The baseline for the headline 'average absolute increase of 14.1 F1 points' is not stated in the same sentence. Please specify whether the comparison is to MATE, MATE-MT, MATE-unweighted, or the best prior baseline per dataset, and give the per-dataset differences so that the average can be reproduced.
  3. [Section 4.1 and Section 3.3] The paper states that MATE uses the validation set (with aspect labels) to estimate seed weights, while the proposed method is unsupervised. However, Section 4.1 also selects the student's parameter configuration using the labeled validation set. This is a common practice, but it should be stated explicitly as a limitation of the overall label-free claim.
  4. [Section 4.2 and Figure 3] The claim that 'the co-training assumption of conditionally independent views is satisfied' for Student-W2V-RSW is not justified: although the student sees only non-seed words and the teacher sees only seed words, the student is trained on the teacher's soft labels, so the two views are not conditionally independent given the label in the sense of Blum and Mitchell (1998). Please rephrase this claim.
  5. [References and Figure 4 caption] The reference 'Srivastava and Sutton, 2071' should be 2017, and the rendering of 'Buciluˇa et al.' includes a stray formatting character. Also, the symbols T_i and S_i in Figure 4 are not defined in the caption or text.

Circularity Check

1 steps flagged · score 4.0 of 10

Co-training's seed-quality estimates are defined by the student's own predictions, but the headline F1 comparison against held-out labels is not circular.

  1. self definitional [Section 3.3, Eq. (4) and Algorithm 1; student trained via Eq. (2) from teacher Eq. (1)/(3)]
    "As no ground-truth labels are available, we follow Khetan et al. (2018) and estimate zj via Maximum Likelihood Estimation using the student's predictions as the current estimate of the ground truth labels. In particular, we assume that the prediction of the student for a training segment si is ti = argmax k pk i . Then, for each seed word we compute the quality estimate for the k-th aspect using the student's predictions for N segments: ˆzk j = (Σ_{i=1}^N 1{c_j^i > 0}1{t_i = k}) / (Σ_{k'} Σ_{i=1}^N 1{c_j^i > 0}1{t_i = k'}). (4)"

    Eq. (4) defines seed-word quality z_j as the distribution of the student's own hard predictions t_i = argmax_k p_k^i on segments containing that seed word. The student p is trained in Eq. (2) to minimize cross-entropy with teacher q, and q (Eq. (1)/(3)) is computed directly from the same seed words, and in later rounds from the same z_j. Thus z_j is not an independent estimate of seed-word quality; it is a summary of the model's inherited beliefs. Updating the teacher with this z and retraining the student closes a fixed-point loop, so the improved teacher is by construction the teacher that agrees with the student.

full rationale

Most of the paper's headline claim is not circular: the student-teacher system is evaluated on held-out human aspect labels (Section 4), and the single-round student already outperforms MATE, ABAE, and LDA-Anchors (Tables 3-4). The seed words are external inputs, and the test labels are not used for training. The only load-bearing self-reference is in iterative co-training: Eq. (4) estimates seed-word quality from the student's own hard predictions, while the student was trained from the teacher built on the same seed words. This is a transparent self-training bootstrap rather than a hidden fit to test labels, and the final evaluation is external. However, because the paper presents the iterative loop as a method for correcting noisy seed words, and because the stopping criterion is a fixed-point agreement condition, the co-training contribution is partially circular: the seed-quality estimates are defined by the model's own predictions. This does not invalidate the empirical comparison, but it lowers the circularity score to 4 rather than 0.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests on four assumptions: the single-label segment formulation, the informativeness of seed words, the value of distillation, and the reliability of student pseudo-labels for seed reweighting. The fitted parameters are the seed-quality weights and standard student hyperparameters. No new physical or conceptual entities are introduced.

free parameters (2)
  • Seed word quality weights z_j^k = Per-seed vectors estimated from student predictions, values not reported
    Introduced in Eq. (3) and estimated by Eq. (4) from the student's hard predictions on unlabeled training segments; the iterative co-training gains depend on these fitted weights.
  • Student training hyperparameters (learning rate, batch size, dropout, L2 weight) = 0.005, 50, 0.5, not specified
    Chosen by hand and tuned on the labeled validation set (Section 4.1 and Appendix A.2); they affect which student configuration is reported and the magnitude of the gains.
assumptions (4)
  • domain assumption Each review segment has exactly one aspect label from K predefined aspects, with the K-th aspect being 'General'.
    The problem definition in Section 2.4 frames multiclass single-label segmentation and treats 'General' as a catch-all for segments with no seed words.
  • domain assumption Seed words are weakly positive indicators: presence of a seed word from G_k in a segment is evidence for aspect k.
    The teacher model in Eq. (1) is built entirely on this premise; if the supplied seed words are not predictive, the teacher is uninformative and the student inherits noise.
  • domain assumption A student trained to imitate the teacher's soft labels can generalize beyond the teacher, including to non-seed words.
    The distillation objective in Eq. (2) assumes the teacher's weak labels carry enough signal for a stronger model to learn; this is validated empirically, not proven.
  • ad hoc to paper The student's hard predictions on unlabeled data can serve as a ground-truth proxy for estimating seed-word quality in Eq. (4).
    This is borrowed from MBEM (Khetan et al., 2018); the paper does not prove the convergence or reliability conditions hold for its specific student-teacher loop, so it is a load-bearing assumption of the co-training contribution.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Leveraging Just a Few Keywords for Fine-Grained Aspect Detection Through Weakly Supervised Co-Training." pith.science (2026). https://pith.science/paper/OSM575WZ

@misc{pith2026190900415,
  author       = {Pith},
  title        = {Pith review of: Leveraging Just a Few Keywords for Fine-Grained Aspect Detection Through Weakly Supervised Co-Training},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OSM575WZ}},
  note         = {Machine review of arXiv:1909.00415}
}
read the original abstract

User-generated reviews can be decomposed into fine-grained segments (e.g., sentences, clauses), each evaluating a different aspect of the principal entity (e.g., price, quality, appearance). Automatically detecting these aspects can be useful for both users and downstream opinion mining applications. Current supervised approaches for learning aspect classifiers require many fine-grained aspect labels, which are labor-intensive to obtain. And, unfortunately, unsupervised topic models often fail to capture the aspects of interest. In this work, we consider weakly supervised approaches for training aspect classifiers that only require the user to provide a small set of seed words (i.e., weakly positive indicators) for the aspects of interest. First, we show that current weakly supervised approaches do not effectively leverage the predictive power of seed words for aspect detection. Next, we propose a student-teacher approach that effectively leverages seed words in a bag-of-words classifier (teacher); in turn, we use the teacher to train a second model (student) that is potentially more powerful (e.g., a neural network that uses pre-trained word embeddings). Finally, we show that iterative co-training can be used to cope with noisy seed words, leading to both improved teacher and student models. Our proposed approach consistently outperforms previous weakly supervised approaches (by 14.1 absolute F1 points on average) in six different domains of product reviews and six multilingual datasets of restaurant reviews.

Figures

Figures reproduced from arXiv: 1909.00415 by the authors.

Figure 1
Figure 1. Example of product review with aspect an [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Our student-teacher approach for segment [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Our weakly supervised co-training approach [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Co-training performance for each round re [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

53 extracted references · 45 canonical work pages

  1. [1]

    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.all mid.sentence after.sentence after.block STRINGS urlintro eprinturl eprintpr...

  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]

    Stefanos Angelidis and Mirella Lapata. 2018. Summarizing opinions: Aspect extraction meets sentiment prediction and they are both weakly supervised. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing

    work page 2018

  4. [4]

    Sanjeev Arora, Rong Ge, Yoni Halpern, David Mimno, Ankur Moitra, David Sontag, Yichen Wu, and Michael Zhu. 2013. A practical algorithm for topic modeling with provable guarantees. In Proceedings of the International Conference on International Conference on Machine Learning

    work page 2013

  5. [5]

    Sanjeev Arora, Yingyu Liang, and Tengyu Ma. 2017. A simple but tough-to-beat baseline for sentence embeddings. In Proceedings of the International Conference on Learning Representations

    work page 2017

  6. [6]

    Jimmy Ba and Rich Caruana. 2014. Do deep nets really need to be deep? In Advances in Neural Information Processing Systems

    work page 2014

  7. [7]

    Maria-Florina Balcan, Avrim Blum, and Ke Yang. 2005. Co-training and expansion: Towards bridging theory and practice. In Advances in Neural Information Processing Systems

    work page 2005

  8. [8]

    David M Blei, Andrew Y Ng, and Michael I Jordan. 2003. Latent dirichlet allocation. Journal of Machine Learning Research, 3(Jan):993--1022

    2003

Show all 53 references
  1. [9]

    Avrim Blum and Tom Mitchell. 1998. Combining labeled and unlabeled data with co-training. In Proceedings of the Annual Conference on Computational Learning Theory

    1998

  2. [10]

    Cristian Buciluǎ, Rich Caruana, and Alexandru Niculescu-Mizil. 2006. Model compression. In Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

    2006

  3. [11]

    Minmin Chen, Kilian Q Weinberger, and John Blitzer. 2011. Co-training for domain adaptation. In Advances in Neural Information Processing Systems

    2011

  4. [12]

    Kevin Clark, Minh-Thang Luong, Christopher D Manning, and Quoc Le. 2018. Semi-supervised sequence modeling with cross-view training. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing

    2018

  5. [13]

    Michael Collins and Yoram Singer. 1999. Unsupervised models for named entity classification. In 1999 Joint SIGDAT Conference on Empirical Methods in Natural Language Processing and Very Large Corpora

    1999

  6. [14]

    Andrew M Dai and Quoc V Le. 2015. Semi-supervised sequence learning. In Advances in Neural Information Processing Systems

    2015

  7. [15]

    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: Human...

    2019

  8. [16]

    Vanessa Wei Feng and Graeme Hirst. 2012. Text-level discourse parsing with rich linguistic features. In Proceedings of the Annual Meeting of the Association for Computational Linguistics

    2012

  9. [17]

    Tommaso Furlanello, Zachary Lipton, Michael Tschannen, Laurent Itti, and Anima Anandkumar. 2018. Born-again neural networks. In International Conference on Machine Learning

    2018

  10. [18]

    Ruidan He, Wee Sun Lee, Hwee Tou Ng, and Daniel Dahlmeier. 2017. An unsupervised neural attention model for aspect extraction. In Proceedings of the Annual Meeting of the Association for Computational Linguistics

    2017

  11. [19]

    Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. 2015. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531

    2015 arXiv

  12. [20]

    Jeremy Howard and Sebastian Ruder. 2018. Universal language model fine-tuning for text classification. In Proceedings of the Annual Meeting of the Association for Computational Linguistics

    2018

  13. [21]

    Minqing Hu and Bing Liu. 2004. Mining and summarizing customer reviews. In Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

    2004

  14. [22]

    Mohit Iyyer, Anupam Guha, Snigdha Chaturvedi, Jordan Boyd-Graber, and Hal Daum \'e III. 2016. Feuding families and former friends: Unsupervised learning for dynamic fictional relationships. In Proceedings of the 2016 Conference of the North American Chapter of the Association ...

    2016

  15. [23]

    Giannis Karamanolakis, Daniel Hsu, and Luis Gravano. 2019. Training neural networks for aspect extraction using descriptive keywords only. In Proceedings of the Second Learning from Limited Labeled Data Workshop

    2019

  16. [24]

    Ashish Khetan, Zachary C Lipton, and Anima Anandkumar. 2018. Learning from noisy singly-labeled data. In Proceedings of the International Conference on Learning Representations

    2018

  17. [25]

    Yoon Kim. 2014. Convolutional neural networks for sentence classification. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing

    2014

  18. [26]

    Yoon Kim and Alexander M Rush. 2016. Sequence-level knowledge distillation. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing

    2016

  19. [27]

    Diederik P Kingma and Jimmy Ba. 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980

    2014 arXiv

  20. [28]

    Bing Liu. 2012. Sentiment analysis and opinion mining. Synthesis lectures on human language technologies, 5(1):1--167

    2012

  21. [29]

    Pengfei Liu, Shafiq Joty, and Helen Meng. 2015. Fine-grained opinion mining with recurrent neural networks and word embeddings. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing

    2015

  22. [30]

    David Lopez-Paz, L \'e on Bottou, Bernhard Sch \"o lkopf, and Vladimir Vapnik. 2016. Unifying distillation and privileged information. In Proceedings of the International Conference on Learning Representations

    2016

  23. [31]

    Bin Lu, Myle Ott, Claire Cardie, and Benjamin K Tsou. 2011. Multi-aspect sentiment analysis with topic models. In 2011 IEEE International Conference on Data Mining Workshops. IEEE

    2011

  24. [32]

    Jeffrey Lund, Connor Cook, Kevin Seppi, and Jordan Boyd-Graber. 2017. Tandem anchoring: A multiword anchor approach for interactive topic modeling. In Proceedings of the Annual Meeting of the Association for Computational Linguistics

    2017

  25. [33]

    Thang Luong, Hieu Pham, and Christopher D Manning. 2015. Effective approaches to attention-based neural machine translation. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing

    2015

  26. [34]

    Julian McAuley, Christopher Targett, Qinfeng Shi, and Anton Van Den Hengel. 2015. Image-based recommendations on styles and substitutes. In Proceedings of the International ACM SIGIR Conference on Research and Development in Information Retrieval

    2015

  27. [35]

    David McClosky, Eugene Charniak, and Mark Johnson. 2006. Effective self-training for parsing. In Proceedings of the 2006 Conference of the North American Chapter of the Association of Computational Linguistics. Association for Computational Linguistics

    2006

  28. [36]

    Nagarajan Natarajan, Inderjit S Dhillon, Pradeep K Ravikumar, and Ambuj Tewari. 2013. Learning with noisy labels. In Advances in Neural Information Processing Systems

    2013

  29. [37]

    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

    2014

  30. [38]

    Maria Pontiki, Dimitris Galanis, Haris Papageorgiou, Ion Androutsopoulos, Suresh Manandhar, AL-Smadi Mohammad, Mahmoud Al-Ayyoub, Yanyan Zhao, Bing Qin, Orph \'e e De Clercq, et al. 2016. Semeval-2016 task 5: Aspect based sentiment analysis. In Proceedings of the 10th Internat...

    2016

  31. [39]

    Soujanya Poria, Erik Cambria, and Alexander Gelbukh. 2016. Aspect extraction for opinion mining with a deep convolutional neural network. Knowledge-Based Systems, 108:42--49

    2016

  32. [40]

    Alec Radford, Karthik Narasimhan, Tim Salimans, and Ilya Sutskever. 2018. Improving language understanding by generative pre-training. https://blog.openai.com/language-unsupervised

    2018

  33. [41]

    Sebastian Ruder, Parsa Ghaffari, and John G Breslin. 2016. A hierarchical model of reviews for aspect-based sentiment analysis. Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing

    2016

  34. [42]

    Sebastian Ruder and Barbara Plank. 2018. Strong baselines for neural semi-supervised learning under domain shift. In Proceedings of the Annual Meeting of the Association for Computational Linguistics

    2018

  35. [43]

    Akash Srivastava and Charles Sutton. 2071. Autoencoding variational inference for topic models. In Proceedings of the International Conference on Learning Representations

  36. [44]

    Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. 2014. Dropout: a simple way to prevent neural networks from overfitting. The Journal of Machine Learning Research, 15(1):1929--1958

    2014

  37. [45]

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, ukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Advances in Neural Information Processing Systems

    2017

  38. [46]

    Weiran Wang. 2019. Everything old is new again: A multi-view learning approach to learning using privileged information and distillation. arXiv preprint arXiv:1903.03694

    2019 arXiv

  39. [47]

    Jason Weston, Samy Bengio, and Nicolas Usunier. 2011. Wsabie: Scaling up to large vocabulary image annotation. In Twenty-Second International Joint Conference on Artificial Intelligence

    2011

  40. [48]

    John Wieting, Mohit Bansal, Kevin Gimpel, and Karen Livescu. 2015. Towards universal paraphrastic sentence embeddings. In Proceedings of the International Conference on Learning Representations

    2015

  41. [49]

    John Wieting and Kevin Gimpel. 2017. Revisiting recurrent networks for paraphrastic sentence embeddings. In Proceedings of the Annual Meeting of the Association for Computational Linguistics

    2017

  42. [50]

    Zichao Yang, Diyi Yang, Chris Dyer, Xiaodong He, Alex Smola, and Eduard Hovy. 2016. Hierarchical attention networks for document classification. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language...

    2016

  43. [51]

    David Yarowsky. 1995. Unsupervised word sense disambiguation rivaling supervised methods. In Annual Meeting of the Association for Computational Linguistics

    1995

  44. [52]

    Lei Zhang, Shuai Wang, and Bing Liu. 2018. Deep learning for sentiment analysis: A survey. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, page e1253

    2018

  45. [53]

    Zhi-Hua Zhou and Ming Li. 2005. Tri-training: Exploiting unlabeled data using three classifiers. IEEE Transactions on Knowledge & Data Engineering, (11):1529--1541

    2005

Pith tools

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