---
title: "An Interactive Multi-Task Learning Network for End-to-End"
description: "The authors propose IMN, an interactive multi-task learning network built from CNN encoders. 23 references, cited by 71 papers here."
authors: ["Ruidan He", "Wee Sun Lee", "Hwee Tou Ng", "Daniel Dahlmeier"]
year: 2019
venue: "Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics"
source: "https://aclanthology.org/P19-1048/"
doi: "10.18653/v1/p19-1048"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/he-2019-interactive-multi-task-learning-network/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/he-2019-interactive-multi-task-learning-network/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2019](https://md.research.elamur.ai/free/en/papers/2019/page/2/) / An Interactive Multi-Task Learning Network for End-to-End Aspect-Based Sentiment Analysis

# An Interactive Multi-Task Learning Network for End-to-End Aspect-Based Sentiment Analysis

**Authors:** [Ruidan He](https://md.research.elamur.ai/free/en/authors/ruidan-he/), [Wee Sun Lee](https://md.research.elamur.ai/free/en/authors/wee-sun-lee/), [Hwee Tou Ng](https://md.research.elamur.ai/free/en/authors/hwee-tou-ng/) and [Daniel Dahlmeier](https://md.research.elamur.ai/free/en/authors/daniel-dahlmeier/) · **Year:** 2019 · **Venue:** [Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2019](https://md.research.elamur.ai/free/en/venues/acl/2019/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/P19-1048/) · [PDF](https://aclanthology.org/P19-1048.pdf) · [DOI 10.18653/v1/p19-1048](https://doi.org/10.18653/v1/p19-1048) · [OpenAlex](https://openalex.org/W2950488390)

**Citations:** 277 citations (OpenAlex, 2026-10-14) · [71](https://md.research.elamur.ai/free/en/papers/he-2019-interactive-multi-task-learning-network/cited-by/) in our library, of the 141,840 papers whose bibliographies we read

## Summary
Aspect-based sentiment analysis (ABSA) involves finding the aspect terms in a review sentence and deciding which sentiment is expressed toward each one. Most systems handle the two subtasks, AE and AS, one after the other. As a result, errors carry over and the connection between the subtasks goes unused. Recent integrated models link the subtasks only through a merged tag set, and they train solely on small aspect-level data. The authors propose IMN, an interactive multi-task learning network built from CNN encoders. A shared sequence of latent vectors feeds several task-specific components:
- aspect and opinion term co-extraction
- token-level sentiment tagging
- two auxiliary document-level tasks: predicting a document's sentiment (DS) and identifying its domain (DD)

IMN goes beyond feature sharing. Over several rounds, it returns task outputs and internal signals to the shared representation. It also passes opinion predictions from AE into the self-attention layer of AS. On three benchmark datasets, IMN outperforms both pipeline and integrated baselines. It does so even without document-level data. The work shows that explicit, iterative interaction between tasks, combined with easily obtained document-level corpora, can strengthen end-to-end ABSA.

## Key points
- AE is framed as BIO tagging with five labels (BA, IA, BP, IP, O), so opinion terms are extracted alongside aspects. AS tags tokens as pos/neg/neu. Only tokens inside aspect terms contribute to the AS training loss.
- DS and DD are used only during training. At inference time, predictions come from the AE and AS outputs.
- Table 6 reports F1-I for IMN−d, averaged over 5 runs:

  | Dataset | No message passing (T=0) | Best result |
  |---|---|---|
  | D1 | 66.98 | 68.32 (T=2) |
  | D2 | 56.03 | 57.82 (T=3) |
  | D3 | 56.65 | 57.91 (T=2) |

  Gains level off after 2–3 iterations.
- Case studies against PIPELINE and INABSA show IMN doing better in several situations:
  - it catches rare aspects such as "build"
  - it skips aspect mentions with no expressed opinion, such as "Pizza"
  - it detects unusual opinion phrases such as "scratches easily"
  - it handles negation and minority sentiment classes better
- Source code is published at github.com/ruidan/IMN-E2E-ABSA.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [National University of Singapore](https://md.research.elamur.ai/free/en/universities/national-university-of-singapore/) — University, Singapore, Singapore · 84 works · 32,204 citations
- [Systems, Applications & Products in Data Processing](https://md.research.elamur.ai/free/en/companies/systems-applications-products-in-data-processing/) — Company, London, United Kingdom · 4 works · 1,483 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 9,086 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 8,889 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 9,296 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 9,206 works

🧭 Topics (1): [😊 Sentiment & emotion](https://md.research.elamur.ai/free/en/topics/sentiment-emotion/) · 608 works

## Cite

Cite as: Ruidan He, Wee Sun Lee, Hwee Tou Ng and Daniel Dahlmeier. 2019. An Interactive Multi-Task Learning Network for End-to-End Aspect-Based Sentiment Analysis. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{he2019interactive,
  title = {{An Interactive Multi-Task Learning Network for End-to-End Aspect-Based Sentiment Analysis}},
  author = {Ruidan He, Wee Sun Lee, Hwee Tou Ng and Daniel Dahlmeier},
  booktitle = {Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics},
  year = {2019},
  doi = {10.18653/v1/p19-1048},
  url = {https://doi.org/10.18653/v1/p19-1048}
}
```

- [← previous STACL: Simultaneous Translation with Implicit Anticipation and Controllable Latency using Prefix-to-Prefix Framework](https://md.research.elamur.ai/free/en/papers/ma-2019-stacl-simultaneous-translation-implicit-anticipation/)
- [next → Learning to Navigate Unseen Environments: Back Translation with Environmental Dropout](https://md.research.elamur.ai/free/en/papers/tan-2019-learning-navigate-unseen-environments-back/)
- [↑ back to papers of 2019](https://md.research.elamur.ai/free/en/papers/2019/page/2/)

## References

23 references, 11 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Xin Li, Lidong Bing, Piji Li, and Wai Lam. 2019. A unified model for opinion target extraction and target sentiment prediction. In AAAI Conference on Artificial Intelligence. [find](https://scholar.google.com/scholar?q=Xin+Li%2C+Lidong+Bing%2C+Piji+Li%2C+and+Wai+Lam%2E+2019%2E+A+unified+model+for+opinion+target+extraction+and+target+sentiment+prediction%2E+In+AAAI+Conference+on+Artificial+Intelligence%2E)
2. Stefanos Angelidis and Mirella Lapata. 2018. [Summarizing opinions: Aspect extraction meets sentiment prediction and they are both weakly supervised](https://md.research.elamur.ai/free/en/papers/angelidis-2018-summarizing-opinions-aspect-extraction-meets/). In Conference on Empirical Methods in Natural Language Processing.
3. Xin Li, Lidong Bing, Piji Li, Wai Lam, and Zhimou Yang. 2018b. Aspect term extraction with history attention and selective transformation. In International Joint Conference on Artificial Intelligence. [find](https://scholar.google.com/scholar?q=Xin+Li%2C+Lidong+Bing%2C+Piji+Li%2C+Wai+Lam%2C+and+Zhimou+Yang%2E+2018b%2E+Aspect+term+extraction+with+history+attention+and+selective+transformation%2E+In+International+Joint+Conference+on+Artificial+Intelligence%2E)
4. Anurag Arnab, Shuai Zheng, Sadeep Jayasumana, Bernardino Romera-Paredes, Måns Larsson, Alexander Kirillov, Bogdan Savchynskyy, Carsten Rother, Fredrik Kahl, and Philip HS Torr. 2018. Conditional random fields meet deep neural networks for semantic segmentation: Combining probabilistic graphical mod… [find](https://scholar.google.com/scholar?q=Anurag+Arnab%2C+Shuai+Zheng%2C+Sadeep+Jayasumana%2C+Bernardino+Romera%2DParedes%2C+Ma%CC%8Ans+Larsson%2C+Alexander+Kirillov%2C+Bogdan+Savchynskyy%2C+Carsten+Rother%2C+Fredrik+Kahl%2C+and+Philip+HS+Torr%2E+2018%2E+Conditional+rand)
5. Xin Li and Wai Lam. 2017. Deep multi-task learning for aspect term extraction with memory interaction. In Conference on Empirical Methods in Natural Language Processing. [find](https://scholar.google.com/scholar?q=Xin+Li+and+Wai+Lam%2E+2017%2E+Deep+multi%2Dtask+learning+for+aspect+term+extraction+with+memory+interaction%2E+In+Conference+on+Empirical+Methods+in+Natural+Language+Processing%2E)
6. Jiangming Liu and Yue Zhang. 2017. Attention modeling for target sentiment. In Conference of the European Chapter of the Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Jiangming+Liu+and+Yue+Zhang%2E+2017%2E+Attention+modeling+for+target+sentiment%2E+In+Conference+of+the+European+Chapter+of+the+Association+for+Computational+Linguistics%2E)
7. Peng Chen, Zhongqian Sun, Lidong Bing, and Wei Yang. 2017. [Recurrent attention network on memory for aspect sentiment analysis](https://md.research.elamur.ai/free/en/papers/chen-2017-recurrent-attention-network-memory-aspect/). In Conference on Empirical Methods in Natural Language Processing.
8. Yang Liu, Sujian Li, Xiaodong Zhang, and Zhifang Sui. 2016. [Implicit discourse relation classification via multi-task neural networks](https://md.research.elamur.ai/free/en/papers/liu-2016-implicit-discourse-relation-classification-multi/). In AAAI Conference on Artificial Intelligence.
9. Jiajun Cheng, Shenglin Zhao, Jiani Zhang, Irwin King, Xin Zhang, and Hui Wang. 2017. Aspect-level sentiment classification with heat (hierarchical attention) network. In ACM on Conference on Information and Knowledge Management. [find](https://scholar.google.com/scholar?q=Jiajun+Cheng%2C+Shenglin+Zhao%2C+Jiani+Zhang%2C+Irwin+King%2C+Xin+Zhang%2C+and+Hui+Wang%2E+2017%2E+Aspect%2Dlevel+sentiment+classification+with+heat+%28hierarchical+attention%29+network%2E+In+ACM+on+Conference+on+Informati)
10. Minh-Thang Luong, Quoc V. Le, Ilya Sutskever, Oriol Vinyals, and Lukasz Kaiser. 2015a. [Multi-task sequence to sequence learning](https://md.research.elamur.ai/free/en/papers/luong-2015-multi-task-sequence-sequence-learning/). In International Conference on Learning Representation.
11. Ronan Collobert and Jason Weston. 2008. A unified architecture for natural language processing: Deep neural networks with multitask learning. In International Conference on Machine Learning. [find](https://scholar.google.com/scholar?q=Ronan+Collobert+and+Jason+Weston%2E+2008%2E+A+unified+architecture+for+natural+language+processing%3A+Deep+neural+networks+with+multitask+learning%2E+In+International+Conference+on+Machine+Learning%2E)
12. Yukun Ma, Haiyun Peng, and Erik Cambira. 2018. Targeted aspect-based sentiment analysis via embedding commonsense knowledge into an attentive LSTM. In AAAI Conference on Artificial Intelligence. [find](https://scholar.google.com/scholar?q=Yukun+Ma%2C+Haiyun+Peng%2C+and+Erik+Cambira%2E+2018%2E+Targeted+aspect%2Dbased+sentiment+analysis+via+embedding+commonsense+knowledge+into+an+attentive+LSTM%2E+In+AAAI+Conference+on+Artificial+Intelligence%2E)
13. Li Dong, Furu Wei, Chuanqi Tan, Duyu Tang, Ming Zhou, and Ke Xu. 2014. [Adaptive recursive neural network for target-dependent Twitter sentiment classification](https://md.research.elamur.ai/free/en/papers/dong-2014-adaptive-recursive-neural-network-target/). In Annual Meeting of the Association for Computational Linguistics.
14. Thien Hai Nguyen and Kiyoaki Shirai. 2015. PhraseRNN: Phrase recursive neural network for aspect-based sentiment analysis. In Conference on Empirical Methods in Natural Language Processing. [find](https://scholar.google.com/scholar?q=Thien+Hai+Nguyen+and+Kiyoaki+Shirai%2E+2015%2E+PhraseRNN%3A+Phrase+recursive+neural+network+for+aspect%2Dbased+sentiment+analysis%2E+In+Conference+on+Empirical+Methods+in+Natural+Language+Processing%2E)
15. Matthew R Gormley, Mark Dredze, and Jason Eisner. 2015. Approximation-aware dependency parsing by belief propagation. Transactions of the Association for Computational Linguistics, 3:489–501. [find](https://scholar.google.com/scholar?q=Matthew+R+Gormley%2C+Mark+Dredze%2C+and+Jason+Eisner%2E+2015%2E+Approximation%2Daware+dependency+parsing+by+belief+propagation%2E+Transactions+of+the+Association+for+Computational+Linguistics%2C+3%3A489%E2%80%93501%2E)
16. Jeffrey Pennington, Richard Socher, and Christopher D Manning. 2014. [GloVe: Global vectors for word representation](https://md.research.elamur.ai/free/en/papers/pennington-2014-glove-global-vectors-word-representation/). In Conference on Empirical Methods in Natural Language Processing.
17. Ruidan He, Wee Sun Lee, Hwee Tou Ng, and Daniel Dahlmeier. 2017. [An unsupervised neural attention model for aspect extraction](https://md.research.elamur.ai/free/en/papers/he-2017-unsupervised-neural-attention-model-aspect/). In Annual Meeting of the Association for Computational Linguistics.
18. Maria Pontiki, Dimitrios Galanis, Haris Papageorgiou, Suresh Manandhar, and Ion Androutsopoulos. 2015. [SemEval-2015 task 12: Aspect based sentiment analysis](https://md.research.elamur.ai/free/en/papers/pontiki-2015-semeval-2015-task-12-aspect/). In International Workshop on Semantic Evaluation.
19. Ruidan He, Wee Sun Lee, Hwee Tou Ng, and Daniel Dahlmeier. 2018a. Effective attention modeling for aspect-level sentiment classification. In International Conference on Computational Linguistics. [find](https://scholar.google.com/scholar?q=Ruidan+He%2C+Wee+Sun+Lee%2C+Hwee+Tou+Ng%2C+and+Daniel+Dahlmeier%2E+2018a%2E+Effective+attention+modeling+for+aspect%2Dlevel+sentiment+classification%2E+In+International+Conference+on+Computational+Linguistics%2E)
20. Ruidan He, Wee Sun Lee, Hwee Tou Ng, and Daniel Dahlmeier. 2018b. Exploiting document knowledge for aspect-level sentiment classification. In Annual Meeting of the Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Ruidan+He%2C+Wee+Sun+Lee%2C+Hwee+Tou+Ng%2C+and+Daniel+Dahlmeier%2E+2018b%2E+Exploiting+document+knowledge+for+aspect%2Dlevel+sentiment+classification%2E+In+Annual+Meeting+of+the+Association+for+Computational+Lingui)
21. Maria Pontiki, Dimitrios Galanis, John Pavlopoulos, Haris Papageorgiou, Ion Androutsopoulos, and Suresh Manandhar. 2014. [SemEval-2014 task 4: Aspect based sentiment analysis](https://md.research.elamur.ai/free/en/papers/pontiki-2014-semeval-2014-task-4-aspect/). In International Workshop on Semantic Evaluation.
22. Xin Li, Lidong Bing, Wai Lam, and Bei Shi. 2018a. [Transformation networks for target-oriented sentiment classification](https://md.research.elamur.ai/free/en/papers/li-2018-transformation-networks-target-oriented-sentiment/). In Annual Meeting of the Association for Computational Linguistics.
23. Guang Qiu, Bing Liu, Jiajun Bu, and Chun Chen. 2011. [Opinion word expansion and target extraction through double propagation](https://md.research.elamur.ai/free/en/papers/qiu-2011-opinion-word-expansion-target-extraction/). Computational Linguistics, 37(1):9–27.

*Updated 24 October 2026*