Pith. sign in

REVIEW 4 major objections 7 minor 46 references

Informative and Controllable Opinion Summarization

T0 review · 4 major / 7 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 Opinion summarization improves when the extractive pre-selection step is replaced by condensing all reviews into dense vectors, and user needs can be injected at test time by a zero-shot query built from background reviews.

desk verdict A genuinely novel framework that is oversold: the SOTA margin comes mostly from the extractive module that the paper claims to move beyond. read the letter →

arxiv 1909.02322 v2 pith:ZIZFQMAJ submitted 2019-09-05 cs.CL

Reinald Kim Amplayo , Mirella Lapata This is my paper · ORCID
classification cs.CL
keywords opinionsummarizationcondense-abstractframeworkextract-abstractzero-shotcustomizationcontrollablegenerationmulti-documentRottenTomatoesdatasetautoencoder
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 argues that the standard two-stage approach to opinion summarization, which first extracts a few salient reviews and then abstracts over them, throws away information that could be useful for tailoring a summary to a particular user. It proposes a CONDENSE-ABSTRACT framework in which every review is first compressed into a dense vector and the abstractive model attends over all of these encodings. The instantiation, CONDA SUM, is shown to produce summaries that are more informative and more faithful than extract-then-abstract baselines on the Rotten Tomatoes movie review dataset. The paper also introduces a zero-shot technique that lets a user steer the summary toward an aspect or sentiment by supplying background reviews as a query, with no need for customized training examples.

What carries the argument

The central object is the pair of dense encodings produced by the CONDENSE model: a document-level encoding $d$ and word-level encodings $h_i$. The ABSTRACT model fuses all $N$ review encodings via attentive pooling whose query is a mean vector; replacing that query with the mean encoding of background reviews $C_x$ is the zero-shot customization mechanism. A second mechanism is a fusion hinge loss that pushes the fused encoding $d'$ close to the encoding of the gold summary while pushing it away from negative summaries.

What would settle it

Compare the summaries produced with background reviews for 'acting' against those produced with random reviews of the same size; if the acting query does not yield a significantly higher fraction of acting-related content than the random query, the semantic steering claimed by the zero-shot technique is not actually occurring. A second check is to run the customization with a deliberately mismatched background (e.g., plot reviews labeled as acting); the technique should degrade predictably if it truly uses aspect semantics.

Watch

Extended reading notes

Core claim

The central claim is that replacing the extractive pre-selection step with a condensation step over all reviews improves both informativeness and controllability of opinion summaries. CONDA SUM uses an LSTM autoencoder to produce a review encoding and word-level encodings for each review, then fuses all review encodings with an attention pooling, averages word encodings across the corpus, and decodes with attention and copying. At test time the query vector for fusion can be replaced by the average encoding of background reviews about a desired aspect (e.g., acting or plot) or sentiment, generating a summary focused on that need. Experiments on Rotten Tomatoes report gains in ROUGE and METEOR over extract-abstract baselines, and human judgments rate CONDA SUM summaries as more informative and correct; separate elicitation shows the customization technique shifts the proportion of summaries perceived to mention acting, plot, or positive/negative sentiment, but only when the salience-biased extractive module is removed.

Load-bearing premise

The customization technique assumes that the average autoencoder encoding of a handful of background reviews about an aspect (e.g., acting) is a reliable query for what a user means by that aspect for any target movie; if the background reviews are not representative or the encodings do not separate aspects, the generated 'customized' summaries may not actually reflect the requested need.

Editorial extensions

If this is right

  • If the CA framework is right, summaries can be generated from arbitrarily large review sets without a pre-selection bottleneck, since memory scales with per-review encoding rather than the full corpus.
  • Customizing summaries no longer requires training data of customized summaries; any aspect or sentiment expressible in background reviews can be steered at test time.
  • The extractive module, while helpful for general-purpose salience, limits customization; the paper's results imply that controllability and extractive bias trade off.
  • The approach transfers to other multi-document summarization tasks where inputs number in the hundreds, such as product or service reviews.

Reading between the lines

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

  • The zero-shot query-vector technique suggests a continuum of control: rather than binary customization, one could interpolate between the general mean query and a need-specific query to trade salience against user focus; the paper does not test this.
  • Because customization depends on the average encoding of background reviews, the method inherits the biases of that background corpus; weighting background reviews by relevance to the target domain is a natural extension.
  • The framework could be combined with contrastive or denoising objectives to make condensed encodings carry more disentangled aspect information, potentially sharpening customization without needing to disable the extractive module.
  • A testable prediction is that the information gain over extract-then-abstract widens as the number of reviews per target grows and as the extractor becomes more imperfect, since condensation loses less information than pre-selection.
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

4 major / 7 minor

Summary. The paper proposes CONDENSE-ABSTRACT (CA), an alternative to the extract-abstract (EA) two-stage framework for opinion summarization. Instead of pre-selecting a subset of reviews, the CA framework condenses every input review into a dense vector and feeds all such encodings to an abstractive decoder. The authors instantiate this as CONDA SUM, using an LSTM autoencoder for condensing and a decoder with attention and copy mechanisms, optionally augmented with a salience-biased BERTCENT extractive step. They also propose a zero-shot customization technique where background reviews representing a user need (e.g., acting, plot, positive/negative sentiment) are averaged and used as the fusion query vector. Experiments on the Rotten Tomatoes dataset report automatic metric gains over several EA and extractive baselines, human Best-Worst Scaling results on 50 movies, and a customization study in which the variant without the extractive step produces summaries that crowdworkers perceive as more aspect- or sentiment-focused. The paper claims that the CA framework produces more informative and more controllable summaries than existing approaches.

Significance. If the central claim were fully supported, the CA framework would be a meaningful step for multi-document opinion summarization: it offers a way to avoid a hard selection bottleneck, and the zero-shot customization idea is practically attractive because it does not require need-specific training data. The paper's strengths include a clean two-stage formulation, a concrete instantiation, careful ablation experiments, human evaluation via Best-Worst Scaling, and reproducible experimental details. However, the reported evidence is in tension with the framework's motivating premise: the best-performing general-purpose system relies on an extractive BERTCENT module, while the variant that enables customization falls below an EA baseline on two automatic metrics. The paper therefore does not currently demonstrate that conditioning on all condense encodings by itself improves informativeness over pre-selected extracts, and the twin claims of informativeness and controllability are not shown in a single system. This is a load-bearing issue that should be addressed before publication.

major comments (4)
  1. [§5, Table 6 vs. Table 2] The central attribution claim is undermined by the ablation results. The best configuration, CONDA SUM, includes the salience-biased BERTCENT extractive step. Removing this module drops ROUGE-L from 18.47 to 16.44 and ROUGE-2 from 7.65 to 6.17 (Table 6), which is below the EA baseline BERTCENT+PTGEN at 17.25 ROUGE-L and 7.43 ROUGE-2 (Table 2). Thus, when the model actually uses all condense encodings without extractive pre-selection, it does not beat an EA baseline on two of the five reported metrics. The abstract and introduction frame the contribution as eliminating reliance on pre-selected content, but the reported state-of-the-art margin comes largely from a module that reintroduces exactly that pre-selection. Please either reframe the paper's claim as a framework that offers a trade-off between a general-purpose extractive-augmented mode and a customizable non-extractive mode, or provide evidence that the non-extractive variant outperforms EA baselines, for example with a stronger condense model or a different fusion design.
  2. [§3.3, Customization evaluation in §5] The zero-shot customization technique assumes that the average of CONDENSE encodings of a set of background reviews C_x is a valid query vector for the target movie's reviews. The paper uses |C_x|=1,000 from Ficler and Goldberg (2017), whose aspect labels are heuristic, but it reports no analysis of how the size or composition of C_x affects customization quality. It also does not test robustness to noisy or mismatched background reviews. Since this assumption is load-bearing for the controllability claim, please add sensitivity experiments varying the size of C_x, the source of the background reviews, and the quality of the aspect/sentiment labels, and report the resulting proportion of aspect/sentiment mentions in the customized summaries.
  3. [§5, Table 5] The customization evaluation shows that the without-extracts variant produces summaries that are more aspect- or sentiment-focused than its own non-customized variant, but it does not assess whether customization degrades other summary qualities. The query-vector replacement in Section 3.3 could, in principle, make customized summaries less informative or factually correct even while increasing aspect focus. Please report automatic metrics or human judgments for the customized outputs against the non-customized CONDA SUM and against the state-of-the-art general-purpose system, so that readers can see whether controllability comes at a measurable cost in summary quality.
  4. [§4 and §5] All experiments are on a single dataset, Rotten Tomatoes, which is a movie-review domain with short, relatively homogeneous reviews. Opinion summarization is commonly evaluated on product reviews (e.g., Amazon or Yelp), where aspects such as price, comfort, or battery life matter and where the zero-shot customization method would be more clearly motivated. Please add at least one additional domain or explicitly justify why conclusions from movie reviews should transfer. This is important because the paper's broader claims are about opinion summarization generally, not about movie reviews specifically.
minor comments (7)
  1. [§5] Typo: "Morevoer" should be "Moreover" in the paragraph following Table 3.
  2. [§3.2 and §5] The terminology is inconsistent: the subsection is titled "Salience-biased Extracts" but later text uses "salient-biased extracts". Please choose one term and use it consistently.
  3. [§4, Comparison Systems] BERTCENT is first described as selecting k=1 review nearest to the centroid, but later the text says "For all extractive steps, we set k=5". Clarify whether the BERTCENT baseline uses k=1 or k=5, since this affects the comparison.
  4. [Eq. (7)] The notation in Equation (7) is unclear: W_p is described as a tensor in R^{D_d × D_d × D_d}, but the expression d_i^T W_p \bar{d} as written suggests a matrix-vector product. Please specify the exact tensor contraction or replace W_p with a matrix.
  5. [§4, Training Configuration] The code URL is given as a placeholder ("xxx.yyy.zzz"). Please provide the actual repository URL in the final version.
  6. [Table 2] The rows for REGRESS+S2S have dashes for ROUGE-1/2/L. This is presumably because those metrics were not reported in Wang and Ling (2016); please state this explicitly in the table caption or footnote.
  7. [Figure 1] The caption says "The latter produces more informative and factual summaries" but the comparison in the example is between an EA system and CONDA SUM; please make clear that the example is illustrative and not from a controlled experiment.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the model is trained against gold consensus summaries and evaluated with external human judgments, while the customization query comes from external background reviews.

full rationale

The paper's derivation chain is supervised rather than self-referential. The CONDENSE autoencoder is trained with a reconstruction objective on the reviews themselves (Eq. 5), and the ABSTRACT decoder is trained with maximum likelihood against gold consensus summaries (Eq. 18). The fusion loss (Eq. 19) uses CONDENSE(y), the encoding of the gold summary, as a training target for the fused review encoding; this is a supervised training signal, not a test-time input. Automatic evaluation compares outputs to held-out gold summaries (Table 2), and the human evaluation uses independent crowdworker judgments (Table 4), so no prediction is evaluated against a quantity that was used to fit the model. The zero-shot customization technique replaces the query vector with an average of encodings of background reviews Cx (Section 3.3); those background reviews come from an external corpus (Ficler and Goldberg, 2017), not from the test summaries, so the customization output is not forced by construction. The self-citations in the paper (Amplayo and Lapata 2020; Liu and Lapata 2019; Angelidis and Lapata 2018) are contextual references in related work and do not carry the argument. The salience-biased extractive module (BERTCENT) is disclosed and ablated in Table 6; although the ablation weakens the attribution of the gain to 'all reviews,' it does not make any output equivalent to an input by construction. No load-bearing step reduces to its own input, so no circularity is found.

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

The central claim depends on the autoencoder's ability to retain salient information in a compact vector, the availability of representative background reviews for customization, and the effectiveness of BERTCENT as a salience measure. No new physical or conceptual entities are introduced.

free parameters (2)
  • k (number of BERTCENT extracts) = 5
    Chosen by hand and tuned on the development set (Section 4). The salience-biased extractive step depends on this value, and the ablation shows it affects ROUGE-L.
  • Number of negative summaries in fusion loss = 5
    Equation (19) samples five negative summaries per training instance. This is a hand-chosen hyperparameter that shapes the fusion loss.
assumptions (4)
  • domain assumption A review can be compressed into a fixed-length vector (the last hidden state of a BiLSTM) without losing information needed for summary generation.
    This is the basis of the CONDENSE model (Section 3.1). If the autoencoder loses aspect or sentiment information, the abstractive decoder cannot recover it.
  • domain assumption Background reviews representing a user need (e.g., acting, plot, positive or negative sentiment) are available at test time.
    Section 3.3 explicitly states: 'we assume access to background reviews to represent the user need.' The customization technique relies on these reviews to form the query vector.
  • domain assumption BERTCENT, a centroid-based extractor, selects reviews that are salient and useful for general-purpose summaries.
    Section 3.2 introduces the salience-biased extractive step using BERTCENT. The ablations show it improves ROUGE-L, but the assumption is that the centroid of BERT review encodings approximates salience.
  • domain assumption The gold consensus summaries in the Rotten Tomatoes dataset are reliable training targets for opinion summarization.
    The generation loss (Equation 18) trains against these gold summaries, which are editor-written and may not capture all valid summaries.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Informative and Controllable Opinion Summarization." pith.science (2026). https://pith.science/paper/ZIZFQMAJ

@misc{pith2026190902322,
  author       = {Pith},
  title        = {Pith review of: Informative and Controllable Opinion Summarization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZIZFQMAJ}},
  note         = {Machine review of arXiv:1909.02322}
}
read the original abstract

Opinion summarization is the task of automatically generating summaries for a set of reviews about a specific target (e.g., a movie or a product). Since the number of reviews for each target can be prohibitively large, neural network-based methods follow a two-stage approach where an extractive step first pre-selects a subset of salient opinions and an abstractive step creates the summary while conditioning on the extracted subset. However, the extractive model leads to loss of information which may be useful depending on user needs. In this paper we propose a summarization framework that eliminates the need to rely only on pre-selected content and waste possibly useful information, especially when customizing summaries. The framework enables the use of all input reviews by first condensing them into multiple dense vectors which serve as input to an abstractive model. We showcase an effective instantiation of our framework which produces more informative summaries and also allows to take user preferences into account using our zero-shot customization technique. Experimental results demonstrate that our model improves the state of the art on the Rotten Tomatoes dataset and generates customized summaries effectively.

Figures

Figures reproduced from arXiv: 1909.02322 by the authors.

Figure 1
Figure 1. Three out of 150 reviews for the movie “Coach Carter”, and summaries written by the editor, and generated by a model following the EXTRACT￾ABSTRACT approach and the proposed CONDENSE￾ABSTRACT framework. The latter produces more infor￾mative and factual summaries whilst allowing to con￾trol aspects of the generated summary (such as the act￾ing or plot of the movie). The vast majority of previous work (Hu and Liu, 200… view at source ↗
Figure 2
Figure 2. Illustration of EA and CA frameworks for opinion summarization. In the CA framework, users can obtain need-specific summaries at test time (e.g., give me a summary focusing on acting). Figure 2b). The CONDENSE model first represents the input reviews as encodings, aiming to condense their meaning and distill information relating to sentiment and various aspects of the target being reviewed. The ABSTRACT model then f… view at source ↗
Figure 3
Figure 3. Examples of general-purpose and need￾specific opinion summaries for the movie “Gremlins”, generated by two versions of CONDASUM. We also show the consensus summary (GOLD). Words/phrases in color highlight aspects pertaining to positive and neg￾ative. More examples can be found in the Appendix. discard information relevant to the user’s need. We thus expect this model to be less effective for cus￾tomization than COND… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Examples of general-purpose and need-specific opinion summaries generated by four systems. We also [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: Examples of general-purpose and need-specific opinion summaries generated by four systems. We also [PITH_FULL_IMAGE:figures/full_fig_p015_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

46 extracted references · 18 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]

    Reinald Kim Amplayo and Mirella Lapata. 2020. https://www.aclweb.org/anthology/2020.acl-main.175 Unsupervised opinion summarization with noising and denoising . In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 1934--1945, Online. Association for Computational Linguistics

    work page 2020

  4. [4]

    Stefanos Angelidis and Mirella Lapata. 2018. https://doi.org/10.18653/v1/D18-1403 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, pages 3675--3686, Brussels, Belgium. Association for Computational Linguistics

    work page doi:10.18653/v1/d18-1403 2018

  5. [5]

    Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio. 2014. https://openreview.net/forum?id=rkE3y85ee Neural machine translation by jointly learning to align and translate . In Proceedings of the 3rd International Conference on Learning Representations, San Diego, California

    work page 2014

  6. [6]

    Sasha Blair-Goldensohn, Kerry Hannan, Ryan McDonald, Tyler Neylon, George Reis, and Jeff Reynar. 2008. https://research.google.com/pubs/archive/34368.pdf Building a sentiment summarizer for local service reviews . In Proceedings of the WWW Workshop on NLP Challenges in the Information Explosion Era (NLPIX), Beijing, China

    work page 2008

  7. [7]

    Arthur Bra z inskas, Mirella Lapata, and Ivan Titov. 2020. https://www.aclweb.org/anthology/2020.acl-main.461 Unsupervised opinion summarization as copycat-review generation . In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5151--5169, Online. Association for Computational Linguistics

    work page 2020

  8. [8]

    Giuseppe Carenini, Jackie Chi Kit Cheung, and Adam Pauls. 2013. https://doi.org/10.1111/j.1467-8640.2012.00417.x Mutli-document summarization of evaluative text . Computational Intelligence, 29(4):545--576

    arXiv 2013

Show all 46 references
  1. [9]

    Eric Chu and Peter Liu. 2019. http://proceedings.mlr.press/v97/chu19b.html M ean S um: A neural model for unsupervised multi-document abstractive summarization . In Proceedings of the 36th International Conference on Machine Learning, volume 97 of Proceedings of Machine Learni...

    2019

  2. [10]

    Conrad, Jochen L

    Jack G. Conrad, Jochen L. Leidner, Frank Schilder, and Ravi Kondadadi. 2009. https://doi.org/10.1145/1568234.1568253 Query-based opinion summarization for legal blog entries . In Proceedings of the 12th International Conference on Artificial Intelligence and Law, ICAIL '09, pa...

    2009

  3. [11]

    Michael Denkowski and Alon Lavie. 2014. https://doi.org/10.3115/v1/W14-3348 Meteor universal: Language specific translation evaluation for any target language . In Proceedings of the Ninth Workshop on Statistical Machine Translation, pages 376--380, Baltimore, Maryland, USA. A...

    2014 doi

  4. [12]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. https://doi.org/10.18653/v1/N19-1423 BERT : Pre-training of deep bidirectional transformers for language understanding . In Proceedings of the 2019 Conference of the North A merican Chapter of the Associat...

    2019 doi

  5. [13]

    Giuseppe Di Fabbrizio, Amanda Stent, and Robert Gaizauskas. 2014. https://doi.org/10.3115/v1/W14-4408 A hybrid approach to multi-document summarization of opinions in reviews . In Proceedings of the 8th International Natural Language Generation Conference ( INLG ) , pages 54--...

    2014 doi

  6. [14]

    G\" u nes Erkan and Dragomir R. Radev. 2004. http://dl.acm.org/citation.cfm?id=1622487.1622501 Lexrank: Graph-based lexical centrality as salience in text summarization . J. Artif. Int. Res., 22(1):457--479

    2004

  7. [15]

    Angela Fan, David Grangier, and Michael Auli. 2018. https://doi.org/10.18653/v1/W18-2706 Controllable abstractive summarization . In Proceedings of the 2nd Workshop on Neural Machine Translation and Generation, pages 45--54, Melbourne, Australia. Association for Computational ...

    2018 doi

  8. [16]

    Jessica Ficler and Yoav Goldberg. 2017. https://doi.org/10.18653/v1/W17-4912 Controlling linguistic style aspects in neural language generation . In Proceedings of the Workshop on Stylistic Variation, pages 94--104, Copenhagen, Denmark. Association for Computational Linguistics

    2017 doi

  9. [17]

    Kavita Ganesan, ChengXiang Zhai, and Jiawei Han. 2010. https://www.aclweb.org/anthology/C10-1039 O pinosis: A graph based approach to abstractive summarization of highly redundant opinions . In Proceedings of the 23rd International Conference on Computational Linguistics (Coli...

    2010

  10. [18]

    Ng, and Bita Nejat

    Shima Gerani, Yashar Mehdad, Giuseppe Carenini, Raymond T. Ng, and Bita Nejat. 2014. https://doi.org/10.3115/v1/D14-1168 Abstractive summarization of product reviews using discourse structure . In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Proc...

    2014 doi

  11. [19]

    Hinton, Nitish Srivastava, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov

    Geoffrey E. Hinton, Nitish Srivastava, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. 2012. http://arxiv.org/abs/1207.0580 Improving neural networks by preventing co-adaptation of feature detectors . CoRR, abs/1207.0580

    2012 arXiv

  12. [20]

    Sepp Hochreiter and J\" u rgen Schmidhuber. 1997. https://doi.org/10.1162/neco.1997.9.8.1735 Long short-term memory . Neural Comput., 9(8):1735--1780

    1997 doi

  13. [21]

    Minqing Hu and Bing Liu. 2004. https://doi.org/10.1145/1014052.1014073 Mining and summarizing customer reviews . In Proceedings of the Tenth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD '04, pages 168--177, New York, NY, USA. ACM

    2004

  14. [22]

    Yuta Kikuchi, Graham Neubig, Ryohei Sasano, Hiroya Takamura, and Manabu Okumura. 2016. https://doi.org/10.18653/v1/D16-1140 Controlling output length in neural encoder-decoders . In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 1...

    2016 doi

  15. [23]

    Hyun Duk Kim, Kavita Ganesan, Parikshit Sondhi, and ChengXiang Zhai. 2011. https://core.ac.uk/download/pdf/4827130.pdf Comprehensive review of opinion summarization . Technical report, University of Illinois at Urbana-Champaign

    2011

  16. [24]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. 2015. https://arxiv.org/pdf/1412.6980 Adam: A method for stochastic optimization . In Proceedings of the 4th International Conference on Learning Representations, San Diego, CA

    2015 arXiv

  17. [25]

    Svetlana Kiritchenko and Saif Mohammad. 2017. https://doi.org/10.18653/v1/P17-2074 Best-worst scaling more reliable than rating scales: A case study on sentiment intensity annotation . In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (...

    2017 doi

  18. [26]

    Kevin Lerman, Sasha Blair-Goldensohn, and Ryan McDonald. 2009. https://www.aclweb.org/anthology/E09-1059 Sentiment summarization: Evaluating and learning user preferences . In Proceedings of the 12th Conference of the E uropean Chapter of the ACL ( EACL 2009) , pages 514--522,...

    2009

  19. [27]

    Chin-Yew Lin. 2004. https://www.aclweb.org/anthology/W04-1013 ROUGE : A package for automatic evaluation of summaries . In Text Summarization Branches Out, pages 74--81, Barcelona, Spain. Association for Computational Linguistics

    2004

  20. [28]

    Peter J Liu, Yu-An Chung, and Jie Ren. 2019. Summae: Zero-shot abstractive text summarization using length-agnostic auto-encoders. arXiv preprint arXiv:1910.00998

    2019 arXiv

  21. [29]

    Peter J Liu, Mohammad Saleh, Etienne Pot, Ben Goodrich, Ryan Sepassi, Lukasz Kaiser, and Noam Shazeer. 2018. https://openreview.net/forum?id=Hyg0vbWC- Generating Wikipedia by summarizing long sequences . In Proceedings of the 7th International Conference on Learning Representa...

    2018

  22. [30]

    Yang Liu and Mirella Lapata. 2019. https://doi.org/10.18653/v1/P19-1500 Hierarchical transformers for multi-document summarization . In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 5070--5081, Florence, Italy. Association for C...

    2019 doi

  23. [31]

    Louviere, Terry N

    Jordan J. Louviere, Terry N. Flynn, and A. A. J. Marley. 2015. https://doi.org/10.1017/CBO9781107337855 Best-Worst Scaling: Theory, Methods and Applications . Cambridge University Press

    2015 doi

  24. [32]

    Subhabrata Mukherjee and Sachindra Joshi. 2013. https://www.aclweb.org/anthology/I13-1065 Sentiment aggregation using C oncept N et ontology . In Proceedings of the Sixth International Joint Conference on Natural Language Processing, pages 570--578, Nagoya, Japan. Asian Federa...

    2013

  25. [33]

    Ramesh Nallapati, Feifei Zhai, and Bowen Zhou. 2017. http://dl.acm.org/citation.cfm?id=3298483.3298681 Summarunner: A recurrent neural network based sequence model for extractive summarization of documents . In Proceedings of the Thirty-First AAAI Conference on Artificial Inte...

    2017

  26. [34]

    Ramesh Nallapati, Bowen Zhou, Cicero dos Santos, C a g lar Gul c ehre, and Bing Xiang. 2016. https://doi.org/10.18653/v1/K16-1028 Abstractive text summarization using sequence-to-sequence RNN s and beyond . In Proceedings of The 20th SIGNLL Conference on Computational Natural ...

    2016 doi

  27. [35]

    Jeffrey Pennington, Richard Socher, and Christopher Manning. 2014. https://doi.org/10.3115/v1/D14-1162 G love: Global vectors for word representation . In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing ( EMNLP ) , pages 1532--1543, Doha,...

    2014 doi

  28. [36]

    Ana-Maria Popescu and Oren Etzioni. 2005. https://www.aclweb.org/anthology/H05-1043 Extracting product features and opinions from reviews . In Proceedings of Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing, pages 339--346...

    2005

  29. [37]

    Radev, Hongyan Jing, and Malgorzata Budzikowska

    Dragomir R. Radev, Hongyan Jing, and Malgorzata Budzikowska. 2000. https://www.aclweb.org/anthology/W00-0403 Centroid-based summarization of multiple documents: sentence extraction, utility-based evaluation, and user studies . In NAACL-ANLP 2000 Workshop: Automatic Summarization

    2000

  30. [38]

    Rush, Sumit Chopra, and Jason Weston

    Alexander M. Rush, Sumit Chopra, and Jason Weston. 2015. https://doi.org/10.18653/v1/D15-1044 A neural attention model for abstractive sentence summarization . In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 379--389, Lisbon, Po...

    2015 doi

  31. [39]

    Liu, and Christopher D

    Abigail See, Peter J. Liu, and Christopher D. Manning. 2017. https://doi.org/10.18653/v1/P17-1099 Get to the point: Summarization with pointer-generator networks . In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers...

    2017 doi

  32. [40]

    Rico Sennrich, Barry Haddow, and Alexandra Birch. 2016. https://doi.org/10.18653/v1/N16-1005 Controlling politeness in neural machine translation via side constraints . In Proceedings of the 2016 Conference of the North A merican Chapter of the Association for Computational Li...

    2016 doi

  33. [41]

    Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. 2014. http://dl.acm.org/citation.cfm?id=2627435.2670313 Dropout: A simple way to prevent neural networks from overfitting . J. Mach. Learn. Res., 15(1):1929--1958

    2014

  34. [42]

    Ilya Sutskever, Oriol Vinyals, and Quoc V. Le. 2014. http://dl.acm.org/citation.cfm?id=2969033.2969173 Sequence to sequence learning with neural networks . In Proceedings of the 27th International Conference on Neural Information Processing Systems - Volume 2, NIPS'14, pages 3...

    2014

  35. [43]

    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. http://dl.acm.org/citation.cfm?id=3295222.3295349 Attention is all you need . In Proceedings of the 31st International Conference on Neural Infor...

    2017

  36. [44]

    Oriol Vinyals, Meire Fortunato, and Navdeep Jaitly. 2015. http://dl.acm.org/citation.cfm?id=2969442.2969540 Pointer networks . In Proceedings of the 28th International Conference on Neural Information Processing Systems - Volume 2, NIPS'15, pages 2692--2700, Cambridge, MA, USA...

    2015

  37. [45]

    Lu Wang and Wang Ling. 2016. https://doi.org/10.18653/v1/N16-1007 Neural network-based abstract generation for opinions and arguments . In Proceedings of the 2016 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technol...

    2016 doi

  38. [46]

    Xinyuan Zhang, Ruiyi Zhang, Manzil Zaheer, and Amr Ahmed. 2020. Unsupervised abstractive dialogue summarization for tete-a-tetes. arXiv preprint arXiv:2009.06851

    2020 arXiv

Pith tools

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