# Dynamic Domain Information Modulation Algorithm for Multi-domain Sentiment Analysis

arXiv: https://arxiv.org/abs/2505.06630 · HTML: https://mlsanity.com/paper/2505.06630/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 59 references that are in the map.
- First posted on arXiv 10 May 2025.

## Builds on (most cited first, 21 of 21)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Communication-Efficient Learning of Deep Networks from Decentralized Data](https://mlsanity.com/paper/1602.05629/) (2016)
- [Explaining and Harnessing Adversarial Examples](https://mlsanity.com/paper/1412.6572/) (2014)
- [Towards Deep Learning Models Resistant to Adversarial Attacks](https://mlsanity.com/paper/1706.06083/) (2017)
- [Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling](https://mlsanity.com/paper/1412.3555/) (2014)
- [Identity Mappings in Deep Residual Networks](https://mlsanity.com/paper/1603.05027/) (2016)
- [Gradient Surgery for Multi-Task Learning](https://mlsanity.com/paper/2001.06782/) (2020)
- [A Tutorial on Bayesian Optimization](https://mlsanity.com/paper/1807.02811/) (2018)
- [Conflict-Averse Gradient Descent for Multi-task Learning](https://mlsanity.com/paper/2110.14048/) (2021)
- [A Survey of Convolutional Neural Networks: Analysis, Applications, and Prospects](https://mlsanity.com/paper/2004.02806/) (2020)
- [xLSTM: Extended Long Short-Term Memory](https://mlsanity.com/paper/2405.04517/) (2024)
- [Adversarial Training Methods for Semi-Supervised Text Classification](https://mlsanity.com/paper/1605.07725/) (2016)
- [Recent Advances in Adversarial Training for Adversarial Robustness](https://mlsanity.com/paper/2102.01356/) (2021)
- [The Illusion of State in State-Space Models](https://mlsanity.com/paper/2404.08819/) (2024)
- [Towards Building the Federated GPT: Federated Instruction Tuning](https://mlsanity.com/paper/2305.05644/) (2023)
- [Adversarial Multi-task Learning for Text Classification](https://mlsanity.com/paper/1704.05742/) (2017)
- [RotoGrad: Gradient Homogenization in Multitask Learning](https://mlsanity.com/paper/2103.02631/) (2021)
- [Provably Efficient Online Hyperparameter Optimization with Population-Based Bandits](https://mlsanity.com/paper/2002.02518/) (2020)
- [Multi-task Learning for Multi-modal Emotion Recognition and Sentiment Analysis](https://mlsanity.com/paper/1905.05812/) (2019)
- [Scalarization for Multi-Task and Multi-Domain Learning at Scale](https://mlsanity.com/paper/2310.08910/) (2023)

## Abstract (arXiv)

Multi-domain sentiment classification aims to mitigate poor performance models due to the scarcity of labeled data in a single domain, by utilizing data labeled from various domains. A series of models that jointly train domain classifiers and sentiment classifiers have demonstrated their advantages, because domain classification helps generate necessary information for sentiment classification. Intuitively, the importance of sentiment classification tasks is the same in all domains for multi-domain sentiment classification; but domain classification tasks are different because the impact of domain information on sentiment classification varies across different fields; this can be controlled through adjustable weights or hyper parameters. However, as the number of domains increases, existing hyperparameter optimization algorithms may face the following challenges: (1) tremendous demand for computing resources, (2) convergence problems, and (3) high algorithm complexity. To efficiently generate the domain information required for sentiment classification in each domain, we propose a dynamic information modulation algorithm. Specifically, the model training process is divided into two stages. In the first stage, a shared hyperparameter, which would control the proportion of domain classification tasks across all fields, is determined. In the second stage, we introduce a novel domain-aware modulation algorithm to adjust the domain information contained in the input text, which is then calculated based on a gradient-based and loss-based method. In summary, experimental results on a public sentiment analysis dataset containing 16 domains prove the superiority of the proposed method.