Pith. sign in

REVIEW 4 major objections 3 minor 48 references

Improving Generative Cross-lingual Aspect-Based Sentiment Analysis with Constrained Decoding

T0 review · 4 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read The paper reports that restricting a multilingual sequence-to-sequence model’s output to valid aspect-sentiment structures improves cross-lingual ABSA by about 5% on the hardest task and by more than 10% in multi-task setups, while removing

desk verdict A plausible and useful application of constrained decoding to cross-lingual ABSA; the abstract can't support the causal claim on its own, but this deserves a proper look at the full paper. read the letter →

arxiv 2508.10369 v1 pith:7OOXVYRR submitted 2025-08-14 cs.CL

Jakub Šmíd , Pavel Přibáň , Pavel Král This is my paper
classification cs.CL
keywords cross-lingualaspect-basedsentimentanalysisconstraineddecodingsequence-to-sequencegenerationmulti-tasklearninglow-resourcelanguageslargelanguagemodelsstructuredprediction
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

Aspect-based sentiment analysis (ABSA) is the job of pulling an opinion target out of a sentence and labelling the sentiment attached to it; it works well in English but lags in low-resource languages, where existing cross-lingual methods lean on translation tools and narrow task definitions. The paper's central claim is that constraining what a multilingual sequence-to-sequence model may generate—forcing output into valid aspect-and-sentiment structures—improves cross-lingual performance by about 5% on the most complex ABSA task and by more than 10% when one model is trained for several ABSA tasks at once. The method is evaluated across seven languages and six ABSA tasks, and the paper reports it beats existing systems and sets new benchmarks for tasks that had not been explored cross-lingually. A sympathetic reader would care because the approach attacks exactly the bottleneck that keeps ABSA out of reach for many languages: dependence on unreliable machine translation.

What carries the argument

Constrained decoding in a sequence-to-sequence generator: during output generation, the decoder is allowed to produce only token sequences that correspond to valid ABSA annotations according to the task's structured schema. Because the same schema applies across languages, the constraint is language-independent and lets one trained model handle multiple tasks without translation.

What would settle it

Take the trained multilingual sequence-to-sequence model and decode the same test examples twice—once with the constrained decoder active and once with the constraint removed (for example, free-form beam search)—then score both on the same ABSA metrics across all seven languages and six tasks. If the constrained pass does not hold roughly a 5% edge on the hardest task and more than 10% on multi-task settings, the claimed improvement is not caused by constrained decoding.

Watch

Extended reading notes

Core claim

The discovery, stated on the paper's own terms, is that constrained decoding is the active ingredient: a sequence-to-sequence generator trained to output ABSA annotations performs substantially better when its decoder is restricted to legal target-aspect-sentiment combinations, especially for the hardest joint task and in multi-task models. This removes the need for external translation tools because the structured label space is shared across languages. The same mechanism lets a single model solve six ABSA tasks in seven languages, with constrained decoding boosting results by more than 10% in multi-task setups. The paper also reports that LLMs perform poorly in zero-shot and few-shot setti

Load-bearing premise

The load-bearing assumption is that the compared systems differ only in constrained decoding, so the reported 5% and more-than-10% gains are not artefacts of larger models, more training data, or different evaluation settings on the baseline side.

Editorial extensions

If this is right

  • Cross-lingual ABSA pipelines can drop external translation services, removing a source of error and cost.
  • A single multilingual sequence-to-sequence model can solve several ABSA tasks at once, with constrained decoding giving more than a 10% boost in multi-task training.
  • Previously unexplored language–task combinations now have concrete benchmark results, providing a baseline for low-resource languages.
  • Large language models are not a drop-in solution: zero-shot and few-shot performance is weak, and fine-tuning is competitive only with substantially longer training and inference times.

Reading between the lines

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

  • A testable extension: the same constrained-decoding idea should transfer to other structured cross-lingual tasks such as named-entity recognition or event extraction, since output validity is language-independent there too.
  • The paper leaves implicit that the multi-task gain may grow with the number of tasks or output complexity; an ablation varying the number of ABSA tasks in one model could check this.
  • The headline percentages are only as strong as the baseline match; re-running the comparison with identical model sizes, training data, and compute would separate the constraint's contribution from capacity effects.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 3 minor

Summary. The manuscript, as represented by the abstract, proposes constrained decoding for sequence-to-sequence models in cross-lingual aspect-based sentiment analysis (ABSA). It claims that this approach improves cross-lingual performance by 5% on average for the most complex task, and by more than 10% in multi-task setups, without relying on external translation tools. The method is said to be evaluated across seven languages and six ABSA tasks, surpassing state-of-the-art methods. The abstract also reports that LLMs underperform in zero-shot and few-shot settings but achieve competitive results when fine-tuned, with trade-offs in training and inference time. No further technical content, experimental details, or full text are provided in the version available for review.

Significance. If the empirical claims are substantiated with controlled comparisons, the proposed method could be a valuable contribution to cross-lingual ABSA, particularly by removing dependence on translation tools and enabling multi-task learning. However, the supplied text provides only aggregate performance figures with no baseline identities, dataset names, task definitions, error analysis, or reproducibility artifacts. The significance of the work therefore cannot be assessed from the available material. The paper does not include machine-checked proofs, code/data links, or parameter-free derivations that would partially offset the lack of experimental detail.

major comments (4)
  1. [Abstract] The central causal claim—that constrained decoding drives the reported 5% and 10%+ improvements—is unsupported because no baseline systems are named and no model sizes, training data, hyperparameters, decoding settings, or evaluation protocols are given. The observed gains could equally be attributed to differences in model capacity, dataset construction, or evaluation scripts. A controlled comparison in which constrained decoding is the only toggled variable is essential but absent.
  2. [Abstract] The experimental scope is vague: the seven languages and six ABSA tasks are not enumerated, and 'the most complex task' is not defined. Without naming the datasets, label schemes, and task formulations, the claimed benchmarks cannot be reproduced or compared with prior work.
  3. [Abstract] No evaluation statistics are reported: there are no standard deviations, confidence intervals, significance tests, or error analyses. The claim of 'surpassing state-of-the-art methods' is therefore not verifiable. Additionally, no code or data links are provided, which further limits reproducibility.
  4. [Abstract] The LLM comparison is also underspecified. Zero-shot, few-shot, and fine-tuning results are mentioned without details on prompts, shot counts, parameter scales, training compute, or inference-time measurements. The statement that fine-tuned LLMs are 'competitive' is not meaningful without a specified metric and comparison set.
minor comments (3)
  1. [Abstract] The abstract uses vague formulations such as '5% on average for the most complex task' and 'boosting results by more than 10%.' Please specify the exact metric (e.g., macro-F1) and the precise comparison configuration (constrained vs. unconstrained decoding with the same base model).
  2. [Abstract] The seven languages and six tasks should be explicitly listed in the abstract or introduction, since they are central to the claimed cross-lingual breadth.
  3. [Abstract] The phrase 'state-of-the-art methods' should be replaced with concrete method names and citations, so readers can assess the comparison.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found in the abstract; the claims are empirical benchmark comparisons with no self-referential derivation chain.

full rationale

The supplied material contains only the abstract; no full text, equations, or experimental details are available to inspect. The abstract's claims are empirical: constrained decoding improves cross-lingual ABSA performance by 5% on the most complex task and by more than 10% in multi-task settings, evaluated across seven languages and six tasks. These are comparative benchmark statements, not quantities defined through fitted parameters or equations. There is no visible derivation that reduces a predicted quantity to an input by construction, no fitted parameter being renamed as a prediction, and no self-citation invoked as load-bearing evidence. While the abstract does not name baselines or fully control for confounds, that is a concern about causal attribution and experimental rigor, not circularity. The hard rule requires quoting a specific reduction; none can be exhibited from the abstract alone. Therefore the honest finding is no significant circularity, score 0.

Assumptions & free parameters 1 free parameters · 2 assumptions · 0 invented entities

The abstract introduces no new physical or mathematical entities. The audit is limited by absence of full text.

free parameters (1)
  • Unspecified model and decoding hyperparameters = not reported
    The abstract reports empirical gains without listing architecture choices, training configurations, or the grammar used for constrained decoding. These choices determine the reported percentages and cannot be audited from the abstract alone.
assumptions (2)
  • domain assumption ABSA task definitions and annotation schemas are aligned across the seven languages and six tasks.
    The claimed cross-lingual performance and benchmark comparisons require that the same aspect and sentiment labels mean the same thing in each language.
  • domain assumption Baseline systems cited as state-of-the-art are implemented fairly and comparably.
    The 5% and 10% improvement claims are relative; without seeing baseline configurations, the comparison assumes fair implementation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Improving Generative Cross-lingual Aspect-Based Sentiment Analysis with Constrained Decoding." pith.science (2026). https://pith.science/paper/7OOXVYRR

@misc{pith2026250810369,
  author       = {Pith},
  title        = {Pith review of: Improving Generative Cross-lingual Aspect-Based Sentiment Analysis with Constrained Decoding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7OOXVYRR}},
  note         = {Machine review of arXiv:2508.10369}
}
read the original abstract

While aspect-based sentiment analysis (ABSA) has made substantial progress, challenges remain for low-resource languages, which are often overlooked in favour of English. Current cross-lingual ABSA approaches focus on limited, less complex tasks and often rely on external translation tools. This paper introduces a novel approach using constrained decoding with sequence-to-sequence models, eliminating the need for unreliable translation tools and improving cross-lingual performance by 5\% on average for the most complex task. The proposed method also supports multi-tasking, which enables solving multiple ABSA tasks with a single model, with constrained decoding boosting results by more than 10\%. We evaluate our approach across seven languages and six ABSA tasks, surpassing state-of-the-art methods and setting new benchmarks for previously unexplored tasks. Additionally, we assess large language models (LLMs) in zero-shot, few-shot, and fine-tuning scenarios. While LLMs perform poorly in zero-shot and few-shot settings, fine-tuning achieves competitive results compared to smaller multilingual models, albeit at the cost of longer training and inference times. We provide practical recommendations for real-world applications, enhancing the understanding of cross-lingual ABSA methodologies. This study offers valuable insights into the strengths and limitations of cross-lingual ABSA approaches, advancing the state-of-the-art in this challenging research domain.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

48 extracted references · 27 canonical work pages

  1. [1]

    In: Walker, M., Ji, H., Stent, A

    Akhtar, M.S., Sawant, P., Sen, S., Ekbal, A., Bhattacharyya, P.: Solving data sparsity for aspect based sentiment analysis using cross-linguality and multi-linguality. In: Walker, M., Ji, H., Stent, A. (eds.) Proceedings of the 2018 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Vol...

    work page doi:10.18653/v1/n18-1053 2018

  2. [2]

    In: Park, J.C., Arase, Y., Hu, B., Lu, W., Wijaya, D., Purwarianti, A., Krisnadhi, A.A

    Bang, Y., Cahyawijaya, S., Lee, N., Dai, W., Su, D., Wilie, B., Lovenia, H., Ji, Z., Yu, T., Chung, W., Do, Q.V., Xu, Y., Fung, P.: A multitask, multilingual, multimodal evaluation of C hat GPT on reasoning, hallucination, and interactivity. In: Park, J.C., Arase, Y., Hu, B., Lu, W., Wijaya, D., Purwarianti, A., Krisnadhi, A.A. (eds.) Proceedings of the 1...

    2023

  3. [3]

    In: 9th International Conference on Learning Representations, ICLR 2021, Virtual Event, Austria, May 3-7, 2021

    Cao, N.D., Izacard, G., Riedel, S., Petroni, F.: Autoregressive entity retrieval. In: 9th International Conference on Learning Representations, ICLR 2021, Virtual Event, Austria, May 3-7, 2021. OpenReview.net (2021), https://openreview.net/forum?id=5k8F6UU39V

    work page 2021

  4. [4]

    In: Jurafsky, D., Chai, J., Schluter, N., Tetreault, J

    Conneau, A., Khandelwal, K., Goyal, N., Chaudhary, V., Wenzek, G., Guzm \'a n, F., Grave, E., Ott, M., Zettlemoyer, L., Stoyanov, V.: Unsupervised cross-lingual representation learning at scale. In: Jurafsky, D., Chai, J., Schluter, N., Tetreault, J. (eds.) Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. pp. 8440--...

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

  5. [5]

    In: Proceedings of the 37th International Conference on Neural Information Processing Systems

    Dettmers, T., Pagnoni, A., Holtzman, A., Zettlemoyer, L.: Qlora: efficient finetuning of quantized llms. In: Proceedings of the 37th International Conference on Neural Information Processing Systems. NIPS '23, Curran Associates Inc., Red Hook, NY, USA (2023)

    work page 2023

  6. [6]

    In: Burstein, J., Doran, C., Solorio, T

    Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: BERT : Pre-training of deep bidirectional transformers for language understanding. In: Burstein, J., Doran, C., Solorio, T. (eds.) Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers)...

    doi:10.18653/v1/n19-1423 2019

  7. [7]

    Dubey, A., Jauhri, A., Pandey, A., Kadian, A., Al-Dahle, A., et al.: The llama 3 herd of models (2024), https://arxiv.org/abs/2407.21783

    arXiv 2024

  8. [8]

    In: Vanderwende, L., Daum \'e III, H., Kirchhoff, K

    Dyer, C., Chahuneau, V., Smith, N.A.: A simple, fast, and effective reparameterization of IBM model 2. In: Vanderwende, L., Daum \'e III, H., Kirchhoff, K. (eds.) Proceedings of the 2013 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies. pp. 644--648. Association for Computational Lingu...

    work page 2013

Show all 48 references
  1. [9]

    In: Proceedings of the 29th International Conference on Computational Linguistics

    Gao, T., Fang, J., Liu, H., Liu, Z., Liu, C., Liu, P., Bao, Y., Yan, W.: LEGO - ABSA : A prompt-based task assemblable unified generative framework for multi-task aspect-based sentiment analysis. In: Proceedings of the 29th International Conference on Computational Linguistics...

    2022

  2. [10]

    In: Rogers, A., Boyd-Graber, J., Okazaki, N

    Gou, Z., Guo, Q., Yang, Y.: M v P : Multi-view prompting improves aspect sentiment tuple prediction. In: Rogers, A., Boyd-Graber, J., Okazaki, N. (eds.) Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). pp. 4380--4...

    2023 doi

  3. [11]

    ICLR 1(2), 3 (2022)

    Hu, E.J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., Wang, L., Chen, W., et al.: Lora: Low-rank adaptation of large language models. ICLR 1(2), 3 (2022)

    2022

  4. [12]

    In: III, H.D., Singh, A

    Hu, J., Ruder, S., Siddhant, A., Neubig, G., Firat, O., Johnson, M.: XTREME : A massively multilingual multi-task benchmark for evaluating cross-lingual generalisation. In: III, H.D., Singh, A. (eds.) Proceedings of the 37th International Conference on Machine Learning. Procee...

    2020

  5. [13]

    In: Goldberg, Y., Kozareva, Z., Zhang, Y

    Hu, M., Wu, Y., Gao, H., Bai, Y., Zhao, S.: Improving aspect sentiment quad prediction via template-order data augmentation. In: Goldberg, Y., Kozareva, Z., Zhang, Y. (eds.) Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing. pp. 7889--7900....

    2022 doi

  6. [14]

    In: Burstein, J., Doran, C., Solorio, T

    Jebbara, S., Cimiano, P.: Z ero-shot cross-lingual opinion target extraction. In: Burstein, J., Doran, C., Solorio, T. (eds.) Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume...

    2019 doi

  7. [15]

    In: Proceedings of the Nineteenth Conference on Computational Natural Language Learning

    Klinger, R., Cimiano, P.: Instance selection improves cross-lingual model training for fine-grained sentiment analysis. In: Proceedings of the Nineteenth Conference on Computational Natural Language Learning. pp. 153--163. Association for Computational Linguistics, Beijing, Ch...

    2015 doi

  8. [16]

    In: Zong, C., Strube, M

    Lambert, P.: Aspect-level cross-lingual sentiment classification with constrained SMT . In: Zong, C., Strube, M. (eds.) Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Proces...

    2015 doi

  9. [17]

    Li, X., Bing, L., Zhang, W., Li, Z., Lam, W.: Unsupervised cross-lingual adaptation for sequence tagging and beyond (2021), https://arxiv.org/abs/2010.12405

    2021 arXiv

  10. [18]

    Lin, N., Fu, Y., Lin, X., Yang, A., Jiang, S.: Cl-xabsa: Contrastive learning for cross-lingual aspect-based sentiment analysis (2023)

    2023

  11. [19]

    Loshchilov, I., Hutter, F.: Decoupled weight decay regularization (2019)

    2019

  12. [20]

    In: Muresan, S., Nakov, P., Villavicencio, A

    Mao, Y., Shen, Y., Yang, J., Zhu, X., Cai, L.: S eq2 P ath: Generating sentiment tuples as paths of a tree. In: Muresan, S., Nakov, P., Villavicencio, A. (eds.) Findings of the Association for Computational Linguistics: ACL 2022. pp. 2215--2225. Association for Computational L...

    2022 doi

  13. [21]

    Minaee, S., Mikolov, T., Nikzad, N., Chenaghlu, M., Socher, R., Amatriain, X., Gao, J.: Large language models: A survey (2024)

    2024

  14. [22]

    Mitra, A., Corro, L.D., Mahajan, S., Codas, A., Simoes, C., Agarwal, S., Chen, X., Razdaibiedina, A., Jones, E., Aggarwal, K., Palangi, H., Zheng, G., Rosset, C., Khanpour, H., Awadallah, A.: Orca 2: Teaching small language models how to reason (2023)

    2023

  15. [23]

    OpenAI: Openai: Introducing chatgpt (Nov 2022), https://openai.com/blog/chatgpt

    2022

  16. [24]

    In: Bethard, S., Carpuat, M., Cer, D., Jurgens, D., Nakov, P., Zesch, T

    Pontiki, M., Galanis, D., Papageorgiou, H., Androutsopoulos, I., Manandhar, S., AL-Smadi, M., Al-Ayyoub, M., Zhao, Y., Qin, B., De Clercq, O., Hoste, V., Apidianaki, M., Tannier, X., Loukachevitch, N., Kotelnikov, E., Bel, N., Jim \'e nez-Zafra, S.M., Eryi g it, G.: S em E val...

    2016

  17. [25]

    In: Nakov, P., Zesch, T., Cer, D., Jurgens, D

    Pontiki, M., Galanis, D., Papageorgiou, H., Manandhar, S., Androutsopoulos, I.: S em E val-2015 task 12: Aspect based sentiment analysis. In: Nakov, P., Zesch, T., Cer, D., Jurgens, D. (eds.) Proceedings of the 9th International Workshop on Semantic Evaluation ( S em E val 201...

    2015 doi

  18. [26]

    In: Nakov, P., Zesch, T

    Pontiki, M., Galanis, D., Pavlopoulos, J., Papageorgiou, H., Androutsopoulos, I., Manandhar, S.: S em E val-2014 task 4: Aspect based sentiment analysis. In: Nakov, P., Zesch, T. (eds.) Proceedings of the 8th International Workshop on Semantic Evaluation ( S em E val 2014). pp...

    2014 doi

  19. [27]

    Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., Liu, P.J.: Exploring the limits of transfer learning with a unified text-to-text transformer. J. Mach. Learn. Res. 21, 140:1--140:67 (2020), http://jmlr.org/papers/v21/20-074.html

    2020

  20. [28]

    In: Riloff, E., Chiang, D., Hockenmaier, J., Tsujii, J

    Schmitt, M., Steinheber, S., Schreiber, K., Roth, B.: Joint aspect and polarity classification for aspect-based sentiment analysis with end-to-end neural networks. In: Riloff, E., Chiang, D., Hockenmaier, J., Tsujii, J. (eds.) Proceedings of the 2018 Conference on Empirical Me...

    2018 doi

  21. [29]

    Shazeer, N., Stern, M.: Adafactor: Adaptive learning rates with sublinear memory cost (2018)

    2018

  22. [30]

    In: De Clercq, O., Barriere, V., Barnes, J., Klinger, R., Sedoc, J., Tafreshi, S

    S m \' d, J., Priban, P., Kral, P.: LL a MA -based models for aspect-based sentiment analysis. In: De Clercq, O., Barriere, V., Barnes, J., Klinger, R., Sedoc, J., Tafreshi, S. (eds.) Proceedings of the 14th Workshop on Computational Approaches to Subjectivity, Sentiment, & So...

    2024

  23. [31]

    In: Proceedings of the 17th International Conference on Agents and Artificial Intelligence - Volume 2: ICAART

    S m \' d, J., Priban, P., Kral, P.: Advancing cross-lingual aspect-based sentiment analysis with llms and constrained decoding for sequence-to-sequence models. In: Proceedings of the 17th International Conference on Agents and Artificial Intelligence - Volume 2: ICAART. pp. 75...

    2025 doi

  24. [32]

    In: Calzolari, N., Kan, M.Y., Hoste, V., Lenci, A., Sakti, S., Xue, N

    S m \' d, J., P r ib \'a n , P., Prazak, O., Kral, P.: C zech dataset for complex aspect-based sentiment analysis tasks. In: Calzolari, N., Kan, M.Y., Hoste, V., Lenci, A., Sakti, S., Xue, N. (eds.) Proceedings of the 2024 Joint International Conference on Computational Lingui...

    2024

  25. [33]

    In: Zong, C., Xia, F., Li, W., Navigli, R

    Tang, Y., Tran, C., Li, X., Chen, P.J., Goyal, N., Chaudhary, V., Gu, J., Fan, A.: Multilingual translation from denoising pre-training. In: Zong, C., Xia, F., Li, W., Navigli, R. (eds.) Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021. pp. 3450--3466...

    2021 doi

  26. [34]

    Touvron, H., Martin, L., Stone, K., Albert, P., et al.: Llama 2: Open foundation and fine-tuned chat models (2023)

    2023

  27. [35]

    Advances in neural information processing systems 30 (2017)

    Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, ., Polosukhin, I.: Attention is all you need. Advances in neural information processing systems 30 (2017)

    2017

  28. [36]

    doi:10.1016/j.inffus.2025.103073, https://www.sciencedirect.com/science/article/pii/S1566253525001460

    S m \' d, J., Kr \'a l, P.: Cross-lingual aspect-based sentiment analysis: A survey on tasks, approaches, and challenges (2025). doi:10.1016/j.inffus.2025.103073, https://www.sciencedirect.com/science/article/pii/S1566253525001460

    2025

  29. [37]

    Proceedings of the AAAI Conference on Artificial Intelligence 34(05), 9122--9129 (Apr 2020)

    Wan, H., Yang, Y., Du, J., Liu, Y., Qi, K., Pan, J.Z.: Target-aspect-sentiment joint detection for aspect-based sentiment analysis. Proceedings of the AAAI Conference on Artificial Intelligence 34(05), 9122--9129 (Apr 2020). doi:10.1609/aaai.v34i05.6447, https://ojs.aaai.org/i...

    2020 doi

  30. [38]

    In: 2018 International joint conference on neural networks (IJCNN)

    Wang, F., Lan, M., Wang, W.: Towards a one-stop solution to both aspect extraction and sentiment analysis tasks with neural multi-task learning. In: 2018 International joint conference on neural networks (IJCNN). pp. 1--8. IEEE (2018)

    2018

  31. [39]

    In: IJCAI

    Wang, W., Pan, S.J.: Transition-based adversarial network for cross-lingual aspect extraction. In: IJCAI. pp. 4475--4481 (2018)

    2018

  32. [40]

    In: Liu, Q., Schlangen, D

    Wolf, T., Debut, L., Sanh, V., Chaumond, J., Delangue, C., Moi, A., Cistac, P., Rault, T., Louf, R., Funtowicz, M., Davison, J., Shleifer, S., von Platen, P., Ma, C., Jernite, Y., Plu, J., Xu, C., Le Scao, T., Gugger, S., Drame, M., Lhoest, Q., Rush, A.: Transformers: State-of...

    2020

  33. [41]

    In: Toutanova, K., Rumshisky, A., Zettlemoyer, L., Hakkani-Tur, D., Beltagy, I., Bethard, S., Cotterell, R., Chakraborty, T., Zhou, Y

    Xue, L., Constant, N., Roberts, A., Kale, M., Al-Rfou, R., Siddhant, A., Barua, A., Raffel, C.: m T 5: A massively multilingual pre-trained text-to-text transformer. In: Toutanova, K., Rumshisky, A., Zettlemoyer, L., Hakkani-Tur, D., Beltagy, I., Bethard, S., Cotterell, R., Ch...

    2021

  34. [42]

    In: Moens, M.F., Huang, X., Specia, L., Yih, S.W.t

    Zhang, W., Deng, Y., Li, X., Yuan, Y., Bing, L., Lam, W.: Aspect sentiment quad prediction as paraphrase generation. In: Moens, M.F., Huang, X., Specia, L., Yih, S.W.t. (eds.) Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. pp. 9209--921...

    2021 doi

  35. [43]

    In: Duh, K., Gomez, H., Bethard, S

    Zhang, W., Deng, Y., Liu, B., Pan, S., Bing, L.: Sentiment analysis in the era of large language models: A reality check. In: Duh, K., Gomez, H., Bethard, S. (eds.) Findings of the Association for Computational Linguistics: NAACL 2024. pp. 3881--3906. Association for Computati...

    2024 doi

  36. [44]

    In: Moens, M.F., Huang, X., Specia, L., Yih, S.W.t

    Zhang, W., He, R., Peng, H., Bing, L., Lam, W.: Cross-lingual aspect-based sentiment analysis with aspect term code-switching. In: Moens, M.F., Huang, X., Specia, L., Yih, S.W.t. (eds.) Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. pp....

    2021 doi

  37. [45]

    Zhang, W., Li, X., Deng, Y., Bing, L., Lam, W.: Towards generative aspect-based sentiment analysis. In: Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 2:...

    2021 doi

  38. [46]

    IEEE/ACM Transactions on Audio, Speech, and Language Processing 23(4), 619--630 (2015)

    Zhou, X., Wan, X., Xiao, J.: Clopinionminer: Opinion target extraction in a cross-language scenario. IEEE/ACM Transactions on Audio, Speech, and Language Processing 23(4), 619--630 (2015)

    2015

  39. [47]

    , " * write output.state after.block = add.period write

    ENTRY address author booktitle chapter doi edition editor eid howpublished institution journal key month note number organization pages publisher school series title type url volume year label INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION in...

  40. [48]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.