[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84310-en":3,"doc-seo-84310-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},84310,687197207919,"Theodora","https://ap-avatar.wpscdn.com/avatar/a000253d6f5f7c60be?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779446848396160552",8,"Research & Report","MASTE A Multi-Agent Pipeline for Zero-Shot Aspect Sentiment Triplet Extraction","Aspect Sentiment Triplet Extraction (ASTE) aims to extract complete (aspect, opinion, sentiment) triples from a single review sentence. Although large language models show strong zero-shot results on many NLP tasks, single-pass generation leaves ASTE constrained by span boundary, opinion grouping, and polarity commitments. MASTE introduces a training-free four-stage multi-agent pipeline where specialized agents sequentially handle compositional subtasks with conditioning on prior outputs. Experiments on four ASTE benchmarks show large gains over zero-shot and chain-of-thought baselines, reducing the gap to supervised methods without labeled triplets.","MASTE: A Multi-Agent Pipeline for Zero-Shot Aspect Sentiment Triplet Extraction  \nAo Hong1 , Lehang Wang2 , Zhirun Yue1 , Mingxin Wang1 , Zihan Wang1 , Houde Liu1 *  \n1Tsinghua University, 2Wuhan University,  \narXiv :2607 .08080v 1 [ cs .CL] 9 Jul 2026  \nAbstract  \nAspect Sentiment Triplet Extraction (ASTE) requires jointly identifying (aspect, opinion, sentiment) triples from a given review sentence.  \nWhile 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 commit to span boundaries, opinion grouping, and polarity in one decoding step.  \nCommon remedies—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 (Multi-Agent pipeline for zero-shot Aspect Sentiment Triplet Extraction), a four-stage framework in which specialized agents handle each compositional subtask sequentially with explicit conditioning on prior outputs. This design enables entirely trainingfree zero-shot ASTE and generalizes across different backbones and datasets. Extensive experiments on four ASTE benchmarks show that MASTE substantially outperforms zeroshot and chain-of-thought LLM baselines under the same backbone, narrowing the gap to fully supervised methods without using any labeled triplets. Our code is available at [https:](https:)//[github.com/Hankerlove/MASTE](github.com/Hankerlove/MASTE).  \n1 Introduction  \nAspect-Based Sentiment Analysis (ABSA) is a fine-grained sentiment analysis task that identifies the sentiment polarity expressed toward each specific aspect mentioned in a given text (Pontiki et al., 2014) . Among its subtasks, Aspect Sentiment Triplet Extraction (ASTE) (Penget al., 2020) is the most integrated formulation.  \n* Corresponding author.  \nFigure 1: (Top) An illustration of the ASTE task: given an input sentence, the model is required to extract the complete set of (Aspect, Opinion, Polarity) triples. Aspects (blue) and opinions (orange) arepaired by arcs whose labels denote the sentiment polarity. (Bottom) Three possible systematic failure modes that zero-shot LLM prompting exhibits on the same example; each impairs Precision (P), Recall (R), or both.  \nAs shown in Figure 1 (Top), given a review sentence, ASTE returns the complete set of (Aspect, Opinion, Polarity) triples, where the Aspect is the term being evaluated, Opinion is the expression that describes the aspect, and Polarity denotes the corresponding sentiment polarity, including positive, negative, and neutral. For example, given the review sentence “The food was great but the service was dreadful !”, an ASTE system is expected to output both (food, great, POS) and (service, dreadful, NEG) .  \nPrevious approaches to ASTE and related ABSA subtasks fall into several categories, including pipeline-based methods (Peng et al., 2020), sequence tagging (Xu et al., 2020 ; Yan et al., 2021), sequence-to-sequence generation (Zhang et al., 2021 ; Naglik and Lango, 2024), grid and table filling tagging schemes (Wu et al., 2020 ; Chen et al., 2022 ; Sun et al., 2024), and multi-domain joint training (Hou et al., 2024) . Despite their strong  \nin-domain performance, these supervised methods share two limitations: (i) they depend on datasetspecific triplet annotations that limit cross-domain transfer, and (ii) they are bound to a fixed annotation convention (e.g., span-boundary style) and require re-training when transferred to new domains or extraction granularities.  \nRecently, large language models (LLMs) have demonstrated remarkable performance and strong zero/few-shot generalization across a wide range of NLP tasks (Brown et al., 2020 ; DeepSeek-AI, 2024), including A","cbCaifmUManEa5mn","https://ap.wps.com/l/cbCaifmUManEa5mn","pdf",378364,3,1,12,"English","en",105,"# Introduction\n## Task definition: ABSA and ASTE\n## Limitations of supervised ASTE methods\n## Challenges of direct LLM zero-shot ASTE\n## Proposed approach: MASTE\n## Contributions and experimental setup","[{\"question\":\"What does ASTE extract from a review sentence?\",\"answer\":\"ASTE jointly extracts (aspect, opinion, sentiment) triples, where the aspect is the target term, the opinion describes it, and the sentiment polarity is positive, negative, or neutral.\"},{\"question\":\"Why are single-pass LLM approaches limited for ASTE in zero-shot settings?\",\"answer\":\"Single-pass generation forces the model to commit to aspect/opinion span boundaries, opinion grouping, and polarity in one decoding step, leading to systematic span-boundary errors and hallucinated spans.\"},{\"question\":\"How does MASTE achieve training-free zero-shot ASTE?\",\"answer\":\"MASTE decomposes ASTE into four sequentially conditioned agents—Aspect, Opinion, Sentiment, and Consistency—each producing structured outputs that condition downstream agents, enabling zero-shot extraction without labeled triplets.\"}]",1784194737,30,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"maste-a-multi-agent-pipeline-for-zero-shot-aspect-sentiment-triplet-extraction","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/maste-a-multi-agent-pipeline-for-zero-shot-aspect-sentiment-triplet-extraction/84310/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-24","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What does ASTE extract from a review sentence?","Question",{"text":75,"@type":76},"ASTE jointly extracts (aspect, opinion, sentiment) triples, where the aspect is the target term, the opinion describes it, and the sentiment polarity is positive, negative, or neutral.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why are single-pass LLM approaches limited for ASTE in zero-shot settings?",{"text":80,"@type":76},"Single-pass generation forces the model to commit to aspect/opinion span boundaries, opinion grouping, and polarity in one decoding step, leading to systematic span-boundary errors and hallucinated spans.",{"name":82,"@type":73,"acceptedAnswer":83},"How does MASTE achieve training-free zero-shot ASTE?",{"text":84,"@type":76},"MASTE decomposes ASTE into four sequentially conditioned agents—Aspect, Opinion, Sentiment, and Consistency—each producing structured outputs that condition downstream agents, enabling zero-shot extraction without labeled triplets.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,122,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":29,"slug":121},"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]