@inproceedings{wang-etal-2026-surgagents,
title = "{S}urg{A}gents: A Multi-agent Framework for Embodied Cataract Surgical Intent Prediction",
author = "Wang, Shengyuan and
Ji, Yanzhe and
Yao, Yue and
Qin, Zhenyue and
Liu, Yang",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the {A}sia-Pacific Chapter of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.aacl-long.135/",
pages = "2958--2973",
ISBN = "979-8-89176-439-2",
abstract = "Predicting next surgical phase, viewed as a form of surgical procedural intent, with high accuracy is essential for robotic real-time surgical systems. Previous approaches struggle with short-term dynamics, occlusions, and phase persistence; end-to-end large vision-language models (LVLMs) can reason from visual inputs but often fail to exploit structured temporal observations. Thus, we introduce SurgAgents, a multi-agent framework for embodied surgical intent prediction. SurgAgents decomposes prediction into specialized agents: (1) a ResNet-based single-frame classifier agent extracts per-frame phase information; (2) a YOLO-based instrument specialist agent extracts per-frame instrument sets; (3) a history descriptor agent composes these facts into frame-level descriptions; (4) a history summarizer agent compresses long histories into a compact temporal narrative; (5) a phase predictor agent assembles the narrative and the current frame into a formatted LVLM prompt; (6) a retrieval agent with a confidence-gated Retrieval-augmented generation (RAG) module selectively retrieves similar case patterns when prediction confidence is low. On Cataract-1K, SurgAgents attains 46.54{\%} macro-F1 with GPT-4o, outperforming the baseline, with systematic ablation studies showing that each agent materially contributes to performance. Latency measurements demonstrate feasibility for robotic real-time surgical support during surgery procedures."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2026-surgagents">
<titleInfo>
<title>SurgAgents: A Multi-agent Framework for Embodied Cataract Surgical Intent Prediction</title>
</titleInfo>
<name type="personal">
<namePart type="given">Shengyuan</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yanzhe</namePart>
<namePart type="family">Ji</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Yao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhenyue</namePart>
<namePart type="family">Qin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yang</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-439-2</identifier>
</relatedItem>
<abstract>Predicting next surgical phase, viewed as a form of surgical procedural intent, with high accuracy is essential for robotic real-time surgical systems. Previous approaches struggle with short-term dynamics, occlusions, and phase persistence; end-to-end large vision-language models (LVLMs) can reason from visual inputs but often fail to exploit structured temporal observations. Thus, we introduce SurgAgents, a multi-agent framework for embodied surgical intent prediction. SurgAgents decomposes prediction into specialized agents: (1) a ResNet-based single-frame classifier agent extracts per-frame phase information; (2) a YOLO-based instrument specialist agent extracts per-frame instrument sets; (3) a history descriptor agent composes these facts into frame-level descriptions; (4) a history summarizer agent compresses long histories into a compact temporal narrative; (5) a phase predictor agent assembles the narrative and the current frame into a formatted LVLM prompt; (6) a retrieval agent with a confidence-gated Retrieval-augmented generation (RAG) module selectively retrieves similar case patterns when prediction confidence is low. On Cataract-1K, SurgAgents attains 46.54% macro-F1 with GPT-4o, outperforming the baseline, with systematic ablation studies showing that each agent materially contributes to performance. Latency measurements demonstrate feasibility for robotic real-time surgical support during surgery procedures.</abstract>
<identifier type="citekey">wang-etal-2026-surgagents</identifier>
<location>
<url>https://aclanthology.org/2026.aacl-long.135/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>2958</start>
<end>2973</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T SurgAgents: A Multi-agent Framework for Embodied Cataract Surgical Intent Prediction
%A Wang, Shengyuan
%A Ji, Yanzhe
%A Yao, Yue
%A Qin, Zhenyue
%A Liu, Yang
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-439-2
%F wang-etal-2026-surgagents
%X Predicting next surgical phase, viewed as a form of surgical procedural intent, with high accuracy is essential for robotic real-time surgical systems. Previous approaches struggle with short-term dynamics, occlusions, and phase persistence; end-to-end large vision-language models (LVLMs) can reason from visual inputs but often fail to exploit structured temporal observations. Thus, we introduce SurgAgents, a multi-agent framework for embodied surgical intent prediction. SurgAgents decomposes prediction into specialized agents: (1) a ResNet-based single-frame classifier agent extracts per-frame phase information; (2) a YOLO-based instrument specialist agent extracts per-frame instrument sets; (3) a history descriptor agent composes these facts into frame-level descriptions; (4) a history summarizer agent compresses long histories into a compact temporal narrative; (5) a phase predictor agent assembles the narrative and the current frame into a formatted LVLM prompt; (6) a retrieval agent with a confidence-gated Retrieval-augmented generation (RAG) module selectively retrieves similar case patterns when prediction confidence is low. On Cataract-1K, SurgAgents attains 46.54% macro-F1 with GPT-4o, outperforming the baseline, with systematic ablation studies showing that each agent materially contributes to performance. Latency measurements demonstrate feasibility for robotic real-time surgical support during surgery procedures.
%U https://aclanthology.org/2026.aacl-long.135/
%P 2958-2973
Markdown (Informal)
[SurgAgents: A Multi-agent Framework for Embodied Cataract Surgical Intent Prediction](https://aclanthology.org/2026.aacl-long.135/) (Wang et al., AACL-IJCNLP 2026)
ACL
- Shengyuan Wang, Yanzhe Ji, Yue Yao, Zhenyue Qin, and Yang Liu. 2026. SurgAgents: A Multi-agent Framework for Embodied Cataract Surgical Intent Prediction. In Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2958–2973, Hengqin, China. Association for Computational Linguistics.