Adversarial Category Alignment Network for Cross-domain Sentiment Classification

Qu, Xiaoye; Zou, Zhikang; Yang, Yang; Zhou, Pan

doi:10.18653/v1/n19-1258

Cited by 48 publications

(50 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For each domain, there are 2,000 labeled reviews and approximately 4000 unlabeled reviews. Following the convention of previous works (Ziser and Reichart, 2018;Ganin et al, 2016;Qu et al, 2019), we construct 12 cross-domain sentiment analysis tasks.…”

Section: Datasets and Experimental Settingmentioning

confidence: 99%

Adversarial and Domain-Aware BERT for Cross-Domain Sentiment Analysis

Sun

Wang

et al. 2020

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics

142

122

View full text Add to dashboard Cite

Cross-domain sentiment classification aims to address the lack of massive amounts of labeled data. It demands to predict sentiment polarity on a target domain utilizing a classifier learned from a source domain. In this paper, we investigate how to efficiently apply the pre-training language model BERT on the unsupervised domain adaptation. Due to the pre-training task and corpus, BERT is taskagnostic, which lacks domain awareness and can not distinguish the characteristic of source and target domain when transferring knowledge. To tackle these problems, we design a post-training procedure, which contains the target domain masked language model task and a novel domain-distinguish pre-training task. The post-training procedure will encourage BERT to be domain-aware and distill the domain-specific features in a self-supervised way. Based on this, we could then conduct the adversarial training to derive the enhanced domain-invariant features. Extensive experiments on Amazon dataset show that our model outperforms state-of-the-art methods by a large margin. The ablation study demonstrates that the remarkable improvement is not only from BERT but also from our method.

show abstract

Section: Datasets and Experimental Settingmentioning

confidence: 99%

Adversarial and Domain-Aware BERT for Cross-Domain Sentiment Analysis

Sun

Wang

et al. 2020

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics

142

122

View full text Add to dashboard Cite

show abstract

“…Sentiment Analysis To verify the effectiveness of SENTIX, We evaluate on the widely used crossdomain sentiment dataset (Blitzer et al, 2007), containing four domains: Books (B), DVD (D), Electronic (E), and Kitchen & Housewares (K). Following the setting of previous works (Ziser and Reichart, 2018;Qu et al, 2019), we test on 12 cross-domain tasks. The model is trained on the source domain and tested on the target domains.…”

Section: Datasetsmentioning

confidence: 99%

SentiX: A Sentiment-Aware Pre-Trained Model for Cross-Domain Sentiment Analysis

Zhou¹,

Tian²,

Wang³

et al. 2020

Proceedings of the 28th International Conference on Computational Linguistics

View full text Add to dashboard Cite

Pre-trained language models have been widely applied to cross-domain NLP tasks like sentiment analysis, achieving state-of-the-art performance. However, due to the variety of users' emotional expressions across domains, fine-tuning the pre-trained models on the source domain tends to overfit, leading to inferior results on the target domain. In this paper, we pre-train a sentimentaware language model (SENTIX) via domain-invariant sentiment knowledge from large-scale review datasets, and utilize it for cross-domain sentiment analysis task without fine-tuning. We propose several pre-training tasks based on existing lexicons and annotations at both token and sentence levels, such as emoticons, sentiment words, and ratings, without human interference. A series of experiments are conducted and the results indicate the great advantages of our model. We obtain new state-of-the-art results in all the cross-domain sentiment analysis tasks, and our proposed SENTIX can be trained with only 1% samples (18 samples) and it achieves better performance than BERT with 90% samples.

show abstract

“…Due to the nature that the first stage of the above methods is not guided by sentiment labels, there is a potential to improve them by learning in an end-to-end manner. Some recent approaches (Ganin et al 2016;He et al 2018;Qu et al 2019) employ adversarial learning or MMD to fulfill the end-to-end learning. However, as discussed previously, they inevitably suffer from the limitations of ignoring sentiment polarity lying in target domains or not leveraging them in an effective manner.…”

Section: Related Workmentioning

confidence: 99%

“…Existing relevant studies could be attributed into two categories: two-stage approaches (Blitzer, Dredze, and Pereira 2007;Glorot, Bordes, and Bengio 2011;Ziser and Reichart 2018; and end-to-end models (Ganin et al 2016;He et al 2018;Qu et al 2019). The two-stage approaches typically construct unsupervised feature extractors or manually select pivot features across domains in the first stage.…”

Section: Introductionmentioning

confidence: 99%

Improving Domain-Adapted Sentiment Classification by Deep Adversarial Mutual Learning

Xue

Zhang

Zha

2020

AAAI

View full text Add to dashboard Cite

Domain-adapted sentiment classification refers to training on a labeled source domain to well infer document-level sentiment on an unlabeled target domain. Most existing relevant models involve a feature extractor and a sentiment classifier, where the feature extractor works towards learning domain-invariant features from both domains, and the sentiment classifier is trained only on the source domain to guide the feature extractor. As such, they lack a mechanism to use sentiment polarity lying in the target domain. To improve domain-adapted sentiment classification by learning sentiment from the target domain as well, we devise a novel deep adversarial mutual learning approach involving two groups of feature extractors, domain discriminators, sentiment classifiers, and label probers. The domain discriminators enable the feature extractors to obtain domain-invariant features. Meanwhile, the label prober in each group explores document sentiment polarity of the target domain through the sentiment prediction generated by the classifier in the peer group, and guides the learning of the feature extractor in its own group. The proposed approach achieves the mutual learning of the two groups in an end-to-end manner. Experiments on multiple public datasets indicate our method obtains the state-of-the-art performance, validating the effectiveness of mutual learning through label probers.

show abstract

Adversarial Category Alignment Network for Cross-domain Sentiment Classification

Cited by 48 publications

References 21 publications

Adversarial and Domain-Aware BERT for Cross-Domain Sentiment Analysis

Adversarial and Domain-Aware BERT for Cross-Domain Sentiment Analysis

SentiX: A Sentiment-Aware Pre-Trained Model for Cross-Domain Sentiment Analysis

Improving Domain-Adapted Sentiment Classification by Deep Adversarial Mutual Learning

Contact Info

Product

Resources

About