@inproceedings{chen-etal-2026-agentic,
title = "Agentic Rubrics: Interfacing Agents with Heterogeneous Environments using Executable Rubrics",
author = "Chen, Shunian and
Yu, Rui and
Tang, Zhengyang and
Du, Yuhao and
Xie, Zejian and
Zhang, Songxin and
Wang, Benyou",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.2690/",
pages = "61198--61226",
ISBN = "979-8-89176-459-0",
abstract = "Language-model agents increasingly act through tools, APIs, databases, and other stateful environments, where task success depends on external evidence rather than final text alone. Existing reward paradigms face a trade-off: preference and rubric-based feedback are flexible but often weakly grounded, while verifier-based rewards are grounded but typically tied to narrow task-specific checkers. We argue that agentic reinforcement learning needs reward interfaces: mechanisms that explicitly connect task requirements to environment evidence, provenance, and reward aggregation. We introduce **Agentic Rubrics**, an executable reward-interface view with four components: criteria, evidence acquisition, provenance, and aggregation. This view separates what should be evaluated from how evidence is obtained, audited, and converted into reward. We also specify a minimal environment contract for auditable reward construction and give an illustrative instantiation in stateful tool-use environments. In this case study, Agentic Rubrics augment final environment feedback with process-aware rewards derived from tool traces and state records, yielding modest improvements in task performance and process reliability while revealing domain-dependent limitations. These results support the position that grounded agent learning should treat reward construction as an interface problem, not only as scalar reward modeling."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chen-etal-2026-agentic">
<titleInfo>
<title>Agentic Rubrics: Interfacing Agents with Heterogeneous Environments using Executable Rubrics</title>
</titleInfo>
<name type="personal">
<namePart type="given">Shunian</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rui</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhengyang</namePart>
<namePart type="family">Tang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuhao</namePart>
<namePart type="family">Du</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zejian</namePart>
<namePart type="family">Xie</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Songxin</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Benyou</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Language-model agents increasingly act through tools, APIs, databases, and other stateful environments, where task success depends on external evidence rather than final text alone. Existing reward paradigms face a trade-off: preference and rubric-based feedback are flexible but often weakly grounded, while verifier-based rewards are grounded but typically tied to narrow task-specific checkers. We argue that agentic reinforcement learning needs reward interfaces: mechanisms that explicitly connect task requirements to environment evidence, provenance, and reward aggregation. We introduce **Agentic Rubrics**, an executable reward-interface view with four components: criteria, evidence acquisition, provenance, and aggregation. This view separates what should be evaluated from how evidence is obtained, audited, and converted into reward. We also specify a minimal environment contract for auditable reward construction and give an illustrative instantiation in stateful tool-use environments. In this case study, Agentic Rubrics augment final environment feedback with process-aware rewards derived from tool traces and state records, yielding modest improvements in task performance and process reliability while revealing domain-dependent limitations. These results support the position that grounded agent learning should treat reward construction as an interface problem, not only as scalar reward modeling.</abstract>
<identifier type="citekey">chen-etal-2026-agentic</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2690/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>61198</start>
<end>61226</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Agentic Rubrics: Interfacing Agents with Heterogeneous Environments using Executable Rubrics
%A Chen, Shunian
%A Yu, Rui
%A Tang, Zhengyang
%A Du, Yuhao
%A Xie, Zejian
%A Zhang, Songxin
%A Wang, Benyou
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F chen-etal-2026-agentic
%X Language-model agents increasingly act through tools, APIs, databases, and other stateful environments, where task success depends on external evidence rather than final text alone. Existing reward paradigms face a trade-off: preference and rubric-based feedback are flexible but often weakly grounded, while verifier-based rewards are grounded but typically tied to narrow task-specific checkers. We argue that agentic reinforcement learning needs reward interfaces: mechanisms that explicitly connect task requirements to environment evidence, provenance, and reward aggregation. We introduce **Agentic Rubrics**, an executable reward-interface view with four components: criteria, evidence acquisition, provenance, and aggregation. This view separates what should be evaluated from how evidence is obtained, audited, and converted into reward. We also specify a minimal environment contract for auditable reward construction and give an illustrative instantiation in stateful tool-use environments. In this case study, Agentic Rubrics augment final environment feedback with process-aware rewards derived from tool traces and state records, yielding modest improvements in task performance and process reliability while revealing domain-dependent limitations. These results support the position that grounded agent learning should treat reward construction as an interface problem, not only as scalar reward modeling.
%U https://aclanthology.org/2026.emnlp-main.2690/
%P 61198-61226
Markdown (Informal)
[Agentic Rubrics: Interfacing Agents with Heterogeneous Environments using Executable Rubrics](https://aclanthology.org/2026.emnlp-main.2690/) (Chen et al., EMNLP 2026)
ACL