# Boomda: Balanced Multi-objective Optimization for Multimodal Domain Adaptation

Source: [Boomda: Balanced Multi-objective Optimization for Multimodal Domain Adaptation](<https://www.sota2.com/research/paper/691414812828b7cc6937eaa0>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-11-13T05:03:13.274Z

Authors: Jun Sun, Xinxin Zhang, Simin Hong, Jian Zhu, Xiang Gao

Published: 2025-11-11T12:03:45.000Z

## Abstract

Multimodal learning, while contributing to numerous success stories across various fields, faces the challenge of prohibitively expensive manual annotation. To address the scarcity of annotated data, a popular solution is unsupervised domain adaptation, which has been extensively studied in unimodal settings yet remains less explored in multimodal settings. In this paper, we investigate heterogeneous multimodal domain adaptation, where the primary challenge is the varying domain shifts of different modalities from the source to the target domain. We first introduce the information bottleneck method to learn representations for each modality independently, and then match the source and target domains in the representation space with correlation alignment. To balance the domain alignment of all modalities, we formulate the problem as a multi-objective task, aiming for a Pareto optimal solution. By exploiting the properties specific to our model, the problem can be simplified to a quadratic programming problem. Further approximation yields a closed-form solution, leading to an efficient modality-balanced multimodal domain adaptation algorithm. The proposed method features \\textbf{B}alanced multi-\\textbf{o}bjective \\textbf{o}ptimization for \\textbf{m}ultimodal \\textbf{d}omain \\textbf{a}daptation, termed \\textbf{Boomda}. Extensive empirical results showcase the effectiveness of the proposed approach and demonstrate that Boomda outperforms the competing schemes. The code is is available at: https://github.com/sunjunaimer/Boomda.git.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Multimodal Emotion Recognition in Conversations on IEMOCAP -\> MELD noise rate 40% (test)](<https://www.sota2.com/research/sota/multimodal-emotion-recognition-in-conversations-on-iemocap-meld-noise-rate-40-percent-test.md>): Cross-scenario Multimodal Emotion Recognition in Conversations; IEMOCAP -\> MELD noise rate 40% (test); Joy Accuracy: 41.29.

- [Cross-scenario Multimodal Emotion Recognition on IEMOCAP -\> MELD 20% Noise (test)](<https://www.sota2.com/research/sota/cross-scenario-multimodal-emotion-recognition-on-iemocap-meld-20-percent-noise-test.md>): Cross-scenario Multimodal Emotion Recognition; IEMOCAP -\> MELD 20% Noise (test); Joy Score: 49.27.

- [Cross-scenario Multimodal Emotion Recognition on MELD -\> IEMOCAP 20% Noise (test)](<https://www.sota2.com/research/sota/cross-scenario-multimodal-emotion-recognition-on-meld-iemocap-20-percent-noise-test.md>): Cross-scenario Multimodal Emotion Recognition; MELD -\> IEMOCAP 20% Noise (test); Joy Accuracy: 54.63.

- [Multimodal Emotion Recognition in Conversations on IEMOCAP → MELD (target)](<https://www.sota2.com/research/sota/multimodal-emotion-recognition-in-conversations-on-iemocap-meld-target.md>): Multimodal Emotion Recognition in Conversations; IEMOCAP → MELD (target); Joy Score: 45.69.

- [Multimodal Emotion Recognition in Conversations on MELD (Subset Metrics)](<https://www.sota2.com/research/sota/multimodal-emotion-recognition-in-conversations-on-meld-subset-metrics.md>): Multimodal Emotion Recognition in Conversations; MELD → IEMOCAP target; Joy Accuracy: 60.19.

- [Cross-scenario Multimodal Emotion Recognition in Conversations on MELD -\> IEMOCAP noise rate 40% (test)](<https://www.sota2.com/research/sota/cross-scenario-multimodal-emotion-recognition-in-conversations-on-meld-iemocap-noise-rate-40-percent-test.md>): Cross-scenario Multimodal Emotion Recognition in Conversations; MELD -\> IEMOCAP noise rate 40% (test); Joy Accuracy: 36.07.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2511.08152>)

[PDF](<https://arxiv.org/pdf/2511.08152.pdf>)
