Prototype-Regularized Federated Learning for Cross-Domain Aspect Sentiment Triplet Extraction

Zongming Cai; Jianhang Tang; Zhenyong Zhang; Jinghui Qin; Kebing Jin; Hankz Hankui Zhuo

cs.CLApr 10, 2026

Prototype-Regularized Federated Learning for Cross-Domain Aspect Sentiment Triplet Extraction

Authors: Zongming Cai, Jianhang Tang, Zhenyong Zhang, Jinghui Qin, Kebing Jin, Hankz Hankui Zhuo

Organizations: State Key Laboratory of Public Big Data, Guizhou University · College of Computer Science and Technology, Guizhou University · School of Information Engineering, Guangdong University of Technology · School of Artificial Intelligence, Nanjing University

Abstract

Aspect Sentiment Triplet Extraction (ASTE) aims to extract all sentiment triplets of aspect terms, opinion terms, and sentiment polarities from a sentence. Existing methods are typically trained on individual datasets in isolation, failing to jointly capture the common feature representations shared across domains. Moreover, data privacy constraints prevent centralized data aggregation. To address these challenges, we propose Prototype-based Cross-Domain Span Prototype extraction (PCD-SpanProto), a prototype-regularized federated learning framework to enable distributed clients to exchange class-level prototypes instead of full model parameters. Specifically, we design a weighted performance-aware aggregation strategy and a contrastive regularization module to improve the global prototype under domain heterogeneity and the promotion between intra-class compactness and inter-class separability across clients. Extensive experiments on four ASTE datasets demonstrate that our method outperforms baselines and reduces communication costs, validating the effectiveness of prototype-based cross-domain knowledge transfer.

Figures & tables

Figure 1: Overview of the Prototype-Regularized Federated Learning Framework for Cross-Domain ASTE.

Label	Meaning
A-O-S	span contains both an aspect and an opinion term, expressing a sentiment
N-O-S	span contains an opinion term, expressing a sentiment
A-N-S	span contains an aspect term, expressing a sentiment
N-N-S	span expresses a sentiment without containing aspect or opinion terms
N-N-N	span has no semantic role and carries no sentiment

Table 1: Prototype labels derived from the three-dimensional span tagging scheme, where

t_{s}\neq N

indicates the span expresses a sentiment.

Figure 2: Illustration of the span tagging process.

Dataset	Before			After
Dataset	Train	Val	Test	Train	Val	Test
14Lap	906	219	328	906	219	328
14Res	1266	310	492	1265	310	492
15Res	605	148	322	605	147	322
16Res	857	210	326	599	145	326

Table 2: Statistics before and after deduplication.

Model	14Lap			14Res			15Res			16Res
	$P$	$R$	$F_{1}$	$P$	$R$	$F_{1}$	$P$	$R$	$F_{1}$	$P$	$R$	$F_{1}$
Peng-two-stage †	37.38	50.38	42.87	43.24	63.66	51.46	48.07	57.51	52.32	46.96	54.21	50.41
GTS-BERT †	58.62	52.35	55.29	68.14	68.77	68.45	62.37	59.71	60.98	66.16	68.81	67.44
BMRC †	70.55	48.98	57.82	75.61	61.77	67.99	68.51	53.40	60.02	71.20	61.08	65.75
ASTE-RL †	64.80	54.99	59.50	70.60	68.65	69.71	65.45	60.29	62.72	67.21	69.69	68.41
Span-ASTE †	62.74	56.75	59.58	74.17	68.27	71.00	64.15	62.10	63.05	69.31	71.32	70.29

Table 3: Performance comparison of different models on ASTE datasets.

P

R

, and

F_{1}

denote Precision, Recall, and F1-score. The second best F1 results are underlined. The symbol “

\dagger

” indicates results retrieved from Liang et al. (2023) .The symbol “

\diamond

” indicates results retrieved from their respective original papers.

Type	Number of Parameters
Full STAGE Model	110,298,760
STAGE Classifier	3,216
Prototypes	3,200

Table 4: Comparison of parameter counts for different components and prototype-based communication.

Method	14Lap	14Res	15Res	16Res
Merged Single-Model	61.42	72.05	61.35	72.23
Base STAGE (14Lap)	61.58	52.49	40.05	49.25
Base STAGE (14Res)	40.10	73.76	58.58	64.62
Base STAGE (15Res)	30.85	65.91	64.79	69.75
Base STAGE (16Res)	32.98	65.26	62.85	73.24
Federated Client (14Lap)	63.66	54.33	56.96	66.24

Table 5: Cross-domain evaluation results (triplet-level F1 scores). Bold values indicate in-domain performance for each model.

Aggregation	14Lap	14Res	15Res	16Res
Uniform	61.89	75.12	68.71	73.84
F1-weighted	63.66	75.49	69.24	74.13

Table 6: Ablation study on prototype aggregation strategy (F1 scores).

Figure 3: Cross-Domain prototype similarity among federated clients on four ASTE benchmarks

Figure 4: Triplet-level F1 scores of each client.

Figure 5: t-SNE visualization of 32 class-level prototype vectors (8 classes $\times$ 4 clients).

Figure 6: Parameter sensitivity analysis of alignment weight $\alpha$ (left, $\beta$ fixed at $2.5\times 10^{-4}$ ) and separation weight $\beta$ (right, $\alpha$ fixed at $2\times 10^{-3}$ ).

Aspect-Based Sentiment Analysis

Aspect-Based Sentiment Analysis (ABSA) identifies what people express opinions about and the sentiment or emotion attached to each aspect.

See all 52 papers

MASTE: A Multi-Agent Pipeline for Zero-Shot Aspect Sentiment Triplet Extraction
Jul 9, 2026Ao Hong, Lehang Wang, Zhirun Yue +3Zero-Shot Learning Aspect-Based Sentiment Analysis
Tsinghua University · Wuhan University
FedDAP: Domain-Aware Prototype Learning for Federated Learning under Domain Shift
Apr 8, 2026Huy Q. Le, Loc X. Nguyen, Yu Qiao +3Federated Learning Distribution Shift
G-LAMP NEXUS Institute, Kyung Hee University · Kyung Hee University
Fine-grained Verification via Diagnostic Reasoning Supervision for Aspect Sentiment Triplet Extraction
May 29, 2026Wenna Lai, Haoran Xie, Guandong Xu +2Aspect-Based Sentiment Analysis
The Hong Kong Polytechnic University, China · Lingnan University, China · Education University of Hong Kong, China

Jul 9, 2026J/K move · Enter open · S save

MASTE: A Multi-Agent Pipeline for Zero-Shot Aspect Sentiment Triplet Extraction

Ao Hong, Lehang Wang, Zhirun Yue, Mingxin Wang+2

Tsinghua University · Wuhan University

Zero-Shot Learning Aspect-Based Sentiment Analysis Multi-Agent LLMs

Open paper AskPDF arXiv

Aspect Sentiment Triplet Extraction (ASTE) requires jointly identifying (aspect, opinion, sentiment) triples from a given review sentence. While large language models (LLMs) achieve strong zero-shot performance on many NLP benchmarks, their effectiveness on ASTE remains limited, as single-pass generation forces the model to determine span boundaries, opinion grouping, and sentiment polarity in a single decoding step. Common remedies, such as few-shot in-context learning and chain-of-thought prompting, offer only marginal improvements and rely heavily on either in-domain demonstrations sampled from labeled training data or carefully engineered reasoning prompts, neither of which is broadly available in zero-shot deployment. Inspired by the classical agent paradigm, we propose MASTE, a multi-agent pipeline for zero-shot Aspect Sentiment Triplet Extraction. MASTE decomposes ASTE into four sequential stages, where specialized agents handle different compositional subtasks with explicit conditioning on prior outputs. This design enables entirely training-free zero-shot ASTE and generalizes across different backbones and datasets. Extensive experiments on four ASTE benchmarks show that MASTE substantially outperforms zero-shot and chain-of-thought LLM baselines under the same backbone, narrowing the gap to fully supervised methods without using any labeled triplets. Code is available at https://github.com/Hankerlove/MASTE.

Abstract

Aspect-Based Sentiment Analysis

Explore similar work

MASTE: A Multi-Agent Pipeline for Zero-Shot Aspect Sentiment Triplet Extraction

FedDAP: Domain-Aware Prototype Learning for Federated Learning under Domain Shift

Fine-grained Verification via Diagnostic Reasoning Supervision for Aspect Sentiment Triplet Extraction