@inproceedings{jahin-parvez-2026-gui,
title = "{GUI}-Primitives: Diagnosing Spatial Reasoning Failures in Vision-Language {GUI} Grounding",
author = "Jahin, Md Abrar and
Parvez, Md Rizwan",
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.2620/",
pages = "59530--59564",
ISBN = "979-8-89176-459-0",
abstract = "Computer-use agents ground natural-language instructions in screenshots to locate interface elements, yet existing benchmarks do not isolate whether models bind relational language to the correct element. We introduce GUI-Primitives, a 994-item benchmark of contrastive instruction pairs over seven spatial relations in graphical user interfaces (left/right, above/below, containment, alignment, proximity, list ordinal, occlusion). Each pair holds the screenshot and anchor fixed while changing the relation expression, so the correct target moves between two designated candidates. Five annotators validate a 196-item subset ($\kappa = 0.94$ well-formedness; $\kappa = 0.79$ target selection). Nineteen vision-language models reach at most 32{\%} strict point-in-box accuracy. Because models emit unconstrained coordinates, we classify each prediction by the candidate region in which it falls. Predictions fall outside both candidates on 60{--}92{\%} of items. Conditional on falling within a candidate region, target selection reaches 0.82{--}0.90 for horizontal position, vertical position, proximity, and list ordinal, but does not differ significantly from 0.50 for containment and occlusion: most failures reflect candidate localization rather than relation understanding. Across ten models, benchmark accuracy correlates with ScreenSpot-Pro accuracy (Spearman $\rho = +0.74$), an exploratory association at this sample size. Marking the two designated candidates raises selection accuracy by 35{--}57 percentage points, an oracle diagnostic that supplies the candidate set rather than a deployable method. We release the benchmark, predictions, and code."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="jahin-parvez-2026-gui">
<titleInfo>
<title>GUI-Primitives: Diagnosing Spatial Reasoning Failures in Vision-Language GUI Grounding</title>
</titleInfo>
<name type="personal">
<namePart type="given">Md</namePart>
<namePart type="given">Abrar</namePart>
<namePart type="family">Jahin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Md</namePart>
<namePart type="given">Rizwan</namePart>
<namePart type="family">Parvez</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>Computer-use agents ground natural-language instructions in screenshots to locate interface elements, yet existing benchmarks do not isolate whether models bind relational language to the correct element. We introduce GUI-Primitives, a 994-item benchmark of contrastive instruction pairs over seven spatial relations in graphical user interfaces (left/right, above/below, containment, alignment, proximity, list ordinal, occlusion). Each pair holds the screenshot and anchor fixed while changing the relation expression, so the correct target moves between two designated candidates. Five annotators validate a 196-item subset (ąppa = 0.94 well-formedness; ąppa = 0.79 target selection). Nineteen vision-language models reach at most 32% strict point-in-box accuracy. Because models emit unconstrained coordinates, we classify each prediction by the candidate region in which it falls. Predictions fall outside both candidates on 60–92% of items. Conditional on falling within a candidate region, target selection reaches 0.82–0.90 for horizontal position, vertical position, proximity, and list ordinal, but does not differ significantly from 0.50 for containment and occlusion: most failures reflect candidate localization rather than relation understanding. Across ten models, benchmark accuracy correlates with ScreenSpot-Pro accuracy (Spearman ρ = +0.74), an exploratory association at this sample size. Marking the two designated candidates raises selection accuracy by 35–57 percentage points, an oracle diagnostic that supplies the candidate set rather than a deployable method. We release the benchmark, predictions, and code.</abstract>
<identifier type="citekey">jahin-parvez-2026-gui</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2620/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>59530</start>
<end>59564</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T GUI-Primitives: Diagnosing Spatial Reasoning Failures in Vision-Language GUI Grounding
%A Jahin, Md Abrar
%A Parvez, Md Rizwan
%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 jahin-parvez-2026-gui
%X Computer-use agents ground natural-language instructions in screenshots to locate interface elements, yet existing benchmarks do not isolate whether models bind relational language to the correct element. We introduce GUI-Primitives, a 994-item benchmark of contrastive instruction pairs over seven spatial relations in graphical user interfaces (left/right, above/below, containment, alignment, proximity, list ordinal, occlusion). Each pair holds the screenshot and anchor fixed while changing the relation expression, so the correct target moves between two designated candidates. Five annotators validate a 196-item subset (ąppa = 0.94 well-formedness; ąppa = 0.79 target selection). Nineteen vision-language models reach at most 32% strict point-in-box accuracy. Because models emit unconstrained coordinates, we classify each prediction by the candidate region in which it falls. Predictions fall outside both candidates on 60–92% of items. Conditional on falling within a candidate region, target selection reaches 0.82–0.90 for horizontal position, vertical position, proximity, and list ordinal, but does not differ significantly from 0.50 for containment and occlusion: most failures reflect candidate localization rather than relation understanding. Across ten models, benchmark accuracy correlates with ScreenSpot-Pro accuracy (Spearman ρ = +0.74), an exploratory association at this sample size. Marking the two designated candidates raises selection accuracy by 35–57 percentage points, an oracle diagnostic that supplies the candidate set rather than a deployable method. We release the benchmark, predictions, and code.
%U https://aclanthology.org/2026.emnlp-main.2620/
%P 59530-59564
Markdown (Informal)
[GUI-Primitives: Diagnosing Spatial Reasoning Failures in Vision-Language GUI Grounding](https://aclanthology.org/2026.emnlp-main.2620/) (Jahin & Parvez, EMNLP 2026)
ACL