@inproceedings{jia-etal-2025-hierarchical,
title = "Hierarchical-Task-Aware Multi-modal Mixture of Incremental {L}o{RA} Experts for Embodied Continual Learning",
author = "Jia, Ziqi and
Wang, Anmin and
Qu, Xiaoyang and
Yang, Xiaowen and
Wang, Jianzong",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.acl-long.1379/",
doi = "10.18653/v1/2025.acl-long.1379",
pages = "28415--28427",
ISBN = "979-8-89176-251-0",
abstract = "Previous continual learning setups for embodied intelligence focused on executing low-level actions based on human commands, neglecting the ability to learn high-level planning and multi-level knowledge. To address these issues, we propose the Hierarchical Embodied Continual Learning Setups (HEC) that divide the agent{'}s continual learning process into two layers: high-level instructions and low-level actions, and define five embodied continual learning sub-setups. Building on these setups, we introduce the Task-aware $\textbf{M}$ixture $\textbf{o}$f $\textbf{I}$ncremental $\textbf{L}$oRA $\textbf{E}$xperts (Task-aware MoILE) method. This approach achieves task recognition by clustering visual-text embeddings and uses both a task-level router and a token-level router to select the appropriate LoRA experts. To effectively address the issue of catastrophic forgetting, we apply Singular Value Decomposition (SVD) to the LoRA parameters obtained from prior tasks, preserving key components while orthogonally training the remaining parts. The experimental results show that our method stands out in reducing the forgetting of old tasks compared to other methods, effectively supporting agents in retaining prior knowledge while continuously learning new tasks."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="jia-etal-2025-hierarchical">
<titleInfo>
<title>Hierarchical-Task-Aware Multi-modal Mixture of Incremental LoRA Experts for Embodied Continual Learning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ziqi</namePart>
<namePart type="family">Jia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anmin</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaoyang</namePart>
<namePart type="family">Qu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaowen</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jianzong</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-251-0</identifier>
</relatedItem>
<abstract>Previous continual learning setups for embodied intelligence focused on executing low-level actions based on human commands, neglecting the ability to learn high-level planning and multi-level knowledge. To address these issues, we propose the Hierarchical Embodied Continual Learning Setups (HEC) that divide the agent’s continual learning process into two layers: high-level instructions and low-level actions, and define five embodied continual learning sub-setups. Building on these setups, we introduce the Task-aware Mixture of Incremental LoRA Experts (Task-aware MoILE) method. This approach achieves task recognition by clustering visual-text embeddings and uses both a task-level router and a token-level router to select the appropriate LoRA experts. To effectively address the issue of catastrophic forgetting, we apply Singular Value Decomposition (SVD) to the LoRA parameters obtained from prior tasks, preserving key components while orthogonally training the remaining parts. The experimental results show that our method stands out in reducing the forgetting of old tasks compared to other methods, effectively supporting agents in retaining prior knowledge while continuously learning new tasks.</abstract>
<identifier type="citekey">jia-etal-2025-hierarchical</identifier>
<identifier type="doi">10.18653/v1/2025.acl-long.1379</identifier>
<location>
<url>https://aclanthology.org/2025.acl-long.1379/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>28415</start>
<end>28427</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Hierarchical-Task-Aware Multi-modal Mixture of Incremental LoRA Experts for Embodied Continual Learning
%A Jia, Ziqi
%A Wang, Anmin
%A Qu, Xiaoyang
%A Yang, Xiaowen
%A Wang, Jianzong
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-251-0
%F jia-etal-2025-hierarchical
%X Previous continual learning setups for embodied intelligence focused on executing low-level actions based on human commands, neglecting the ability to learn high-level planning and multi-level knowledge. To address these issues, we propose the Hierarchical Embodied Continual Learning Setups (HEC) that divide the agent’s continual learning process into two layers: high-level instructions and low-level actions, and define five embodied continual learning sub-setups. Building on these setups, we introduce the Task-aware Mixture of Incremental LoRA Experts (Task-aware MoILE) method. This approach achieves task recognition by clustering visual-text embeddings and uses both a task-level router and a token-level router to select the appropriate LoRA experts. To effectively address the issue of catastrophic forgetting, we apply Singular Value Decomposition (SVD) to the LoRA parameters obtained from prior tasks, preserving key components while orthogonally training the remaining parts. The experimental results show that our method stands out in reducing the forgetting of old tasks compared to other methods, effectively supporting agents in retaining prior knowledge while continuously learning new tasks.
%R 10.18653/v1/2025.acl-long.1379
%U https://aclanthology.org/2025.acl-long.1379/
%U https://doi.org/10.18653/v1/2025.acl-long.1379
%P 28415-28427
Markdown (Informal)
[Hierarchical-Task-Aware Multi-modal Mixture of Incremental LoRA Experts for Embodied Continual Learning](https://aclanthology.org/2025.acl-long.1379/) (Jia et al., ACL 2025)
ACL