@inproceedings{sun-etal-2026-memobench,
title = "{MEMOB}ench: A Process Level Memory Benchmark for Robotic Manipulation",
author = "Sun, Haiyang and
Wang, Haoxiao and
Chen, Junming and
Fang, Weicheng and
Su, Zihao and
Yi, Jingkun and
Yi, Wenyou and
Chen, Hao and
Zhao, Zhou",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.333/",
pages = "7025--7040",
ISBN = "979-8-89176-462-0",
abstract = "Robotic manipulation often requires actingon information that is no longer visible, yet Vision-Language-Action policies are usually evaluated when the current observation largely determines the next action. Existing robotic memory benchmarks expose this gap, but they still rely mainly on final task success and there-fore conflate forgetting with manipulation failure. We present MEMOBench, a benchmark for process level memory evaluation in robotic manipulation. MEMOBench includes 30 history dependent tasks, 1,500 expert demonstrations, and 4,200 executable checkpoint instances from 84 templates. Each checkpointpairs coarse to fine language with a simulator predicate and labels one memory operation:Storage, Update, or Compression. These annotations define Memory Storage Rate, MemoryUpdate Rate, and Memory Compression Rate,which measure memory fidelity alongside tasksuccess. Across standard and memory augmented VLA policies, the strongest memory module baseline reaches only 31.9{\%}o average success rate, and high storage often coexists with weak update and compression. Checkpoint language also supervises semantic, contrastive, and framewise memory alignment objectives, yielding modest gains across differentmemory operations. MEMOBench provides a diagnostic evaluation suite and training supervision for memory grounded robotic policies. The project page is available at https://github.com/Collab-Gen/MEMoBench."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="sun-etal-2026-memobench">
<titleInfo>
<title>MEMOBench: A Process Level Memory Benchmark for Robotic Manipulation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Haiyang</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haoxiao</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junming</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weicheng</namePart>
<namePart type="family">Fang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zihao</namePart>
<namePart type="family">Su</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jingkun</namePart>
<namePart type="family">Yi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenyou</namePart>
<namePart type="family">Yi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hao</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhou</namePart>
<namePart type="family">Zhao</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Robotic manipulation often requires actingon information that is no longer visible, yet Vision-Language-Action policies are usually evaluated when the current observation largely determines the next action. Existing robotic memory benchmarks expose this gap, but they still rely mainly on final task success and there-fore conflate forgetting with manipulation failure. We present MEMOBench, a benchmark for process level memory evaluation in robotic manipulation. MEMOBench includes 30 history dependent tasks, 1,500 expert demonstrations, and 4,200 executable checkpoint instances from 84 templates. Each checkpointpairs coarse to fine language with a simulator predicate and labels one memory operation:Storage, Update, or Compression. These annotations define Memory Storage Rate, MemoryUpdate Rate, and Memory Compression Rate,which measure memory fidelity alongside tasksuccess. Across standard and memory augmented VLA policies, the strongest memory module baseline reaches only 31.9%o average success rate, and high storage often coexists with weak update and compression. Checkpoint language also supervises semantic, contrastive, and framewise memory alignment objectives, yielding modest gains across differentmemory operations. MEMOBench provides a diagnostic evaluation suite and training supervision for memory grounded robotic policies. The project page is available at https://github.com/Collab-Gen/MEMoBench.</abstract>
<identifier type="citekey">sun-etal-2026-memobench</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.333/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>7025</start>
<end>7040</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MEMOBench: A Process Level Memory Benchmark for Robotic Manipulation
%A Sun, Haiyang
%A Wang, Haoxiao
%A Chen, Junming
%A Fang, Weicheng
%A Su, Zihao
%A Yi, Jingkun
%A Yi, Wenyou
%A Chen, Hao
%A Zhao, Zhou
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F sun-etal-2026-memobench
%X Robotic manipulation often requires actingon information that is no longer visible, yet Vision-Language-Action policies are usually evaluated when the current observation largely determines the next action. Existing robotic memory benchmarks expose this gap, but they still rely mainly on final task success and there-fore conflate forgetting with manipulation failure. We present MEMOBench, a benchmark for process level memory evaluation in robotic manipulation. MEMOBench includes 30 history dependent tasks, 1,500 expert demonstrations, and 4,200 executable checkpoint instances from 84 templates. Each checkpointpairs coarse to fine language with a simulator predicate and labels one memory operation:Storage, Update, or Compression. These annotations define Memory Storage Rate, MemoryUpdate Rate, and Memory Compression Rate,which measure memory fidelity alongside tasksuccess. Across standard and memory augmented VLA policies, the strongest memory module baseline reaches only 31.9%o average success rate, and high storage often coexists with weak update and compression. Checkpoint language also supervises semantic, contrastive, and framewise memory alignment objectives, yielding modest gains across differentmemory operations. MEMOBench provides a diagnostic evaluation suite and training supervision for memory grounded robotic policies. The project page is available at https://github.com/Collab-Gen/MEMoBench.
%U https://aclanthology.org/2026.findings-emnlp.333/
%P 7025-7040
Markdown (Informal)
[MEMOBench: A Process Level Memory Benchmark for Robotic Manipulation](https://aclanthology.org/2026.findings-emnlp.333/) (Sun et al., Findings 2026)
ACL
- Haiyang Sun, Haoxiao Wang, Junming Chen, Weicheng Fang, Zihao Su, Jingkun Yi, Wenyou Yi, Hao Chen, and Zhou Zhao. 2026. MEMOBench: A Process Level Memory Benchmark for Robotic Manipulation. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 7025–7040, Budapest, Hungary. Association for Computational Linguistics.