# Learning to Share by Masking the Non-shared for Multi-domain Sentiment Classification

arXiv: https://arxiv.org/abs/2104.08480 · HTML: https://mlsanity.com/paper/2104.08480/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 35 references that are in the map.
- First posted on arXiv 17 Apr 2021.

## Builds on (most cited first, 16 of 16)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Domain-Adversarial Training of Neural Networks](https://mlsanity.com/paper/1505.07818/) (2015)
- [Categorical Reparameterization with Gumbel-Softmax](https://mlsanity.com/paper/1611.01144/) (2016)
- [Deep contextualized word representations](https://mlsanity.com/paper/1802.05365/) (2018)
- [Adversarial Examples Are Not Bugs, They Are Features](https://mlsanity.com/paper/1905.02175/) (2019)
- [Domain Separation Networks](https://mlsanity.com/paper/1608.06019/) (2016)
- [Utilizing BERT for Aspect-Based Sentiment Analysis via Constructing Auxiliary Sentence](https://mlsanity.com/paper/1903.09588/) (2019)
- [Learned in Translation: Contextualized Word Vectors](https://mlsanity.com/paper/1708.00107/) (2017)
- [Adversarial Multi-task Learning for Text Classification](https://mlsanity.com/paper/1704.05742/) (2017)
- [Multi-Source Domain Adaptation with Mixture of Experts](https://mlsanity.com/paper/1809.02256/) (2018)
- [SKEP: Sentiment Knowledge Enhanced Pre-training for Sentiment Analysis](https://mlsanity.com/paper/2005.05635/) (2020)
- [Multinomial Adversarial Networks for Multi-Domain Text Classification](https://mlsanity.com/paper/1802.05694/) (2018)
- [Conditional Adversarial Networks for Multi-Domain Text Classification](https://mlsanity.com/paper/2102.10176/) (2021)
- [Dual Adversarial Co-Learning for Multi-Domain Text Classification](https://mlsanity.com/paper/1909.08203/) (2019)
- [Same Representation, Different Attentions: Shareable Sentence Representation Learning from Multiple Tasks](https://mlsanity.com/paper/1804.08139/) (2018)
- [Mixup Regularized Adversarial Networks for Multi-Domain Text Classification](https://mlsanity.com/paper/2102.00467/) (2021)

## Built on by (most cited first, 3 of 3)

- [Graph Adaptive Semantic Transfer for Cross-domain Sentiment Classification](https://mlsanity.com/paper/2205.08772/) (2022)
- [Explainable Misinformation Detection Across Multiple Social Media Platforms](https://mlsanity.com/paper/2203.11724/) (2022)
- [Multi-Attribute Multi-Grained Adaptation of Pre-Trained Language Models for Text Understanding from Bayesian Perspective](https://mlsanity.com/paper/2503.06085/) (2025)

## Abstract (arXiv)

Multi-domain sentiment classification deals with the scenario where labeled data exists for multiple domains but insufficient for training effective sentiment classifiers that work across domains. Thus, fully exploiting sentiment knowledge shared across domains is crucial for real world applications. While many existing works try to extract domain-invariant features in high-dimensional space, such models fail to explicitly distinguish between shared and private features at text-level, which to some extent lacks interpretablity. Based on the assumption that removing domain-related tokens from texts would help improve their domain-invariance, we instead first transform original sentences to be domain-agnostic. To this end, we propose the BertMasker network which explicitly masks domain-related words from texts, learns domain-invariant sentiment features from these domain-agnostic texts, and uses those masked words to form domain-aware sentence representations. Empirical experiments on a well-adopted multiple domain sentiment classification dataset demonstrate the effectiveness of our proposed model on both multi-domain sentiment classification and cross-domain settings, by increasing the accuracy by 0.94% and 1.8% respectively. Further analysis on masking proves that removing those domain-related and sentiment irrelevant tokens decreases texts' domain distinction, resulting in the performance degradation of a BERT-based domain classifier by over 12%.