SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2606.08063 · ICML · 2026

Robust-U1: Can MLLMs Self-Recover Corrupted Visual Content for Robust Understanding?

Qifeng Chen, Runtao Liu, Mengjie Zhao, Wei Wei, Xiangyu Wu, Qingfa Xiao, Jianmin Chen, Jiaqi Tang, Youyang Zhai

arXiv · PDF · Open in the Atlas

Code that ran

7 of 13 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 6 failures of our instrument; 6 no recorded run), measured 2026-11-21; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2606.08063.json · badge: /api/ran/2606.08063.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
jqtangust/Robust-U1 canonical (classifier label)
classifier basis: census
7 ran · 6 no recorded run
FunctionStatusWhere it lives
get_1d_sincos_pos_embed_from_grid Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jqtangust/Robust-U1/modeling/modeling/bagel/modeling_utils.py
pointer only (licence: NONE) · get_code("e5947aba1d10885f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_2d_sincos_pos_embed Ran
ran, honoured the contract we drafted
jqtangust/Robust-U1/modeling/modeling/bagel/modeling_utils.py
pointer only (licence: NONE) · get_code("58f584dfd7b8fc3f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

pad_sequence Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jqtangust/Robust-U1/modeling/modeling/bagel/qwen2_navit.py
pointer only (licence: NONE) · get_code("26da3ed73c66b411")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

robust_tinyclip_score Ran
ran; our contract draft was wrong, not the code
jqtangust/Robust-U1/rewards/flow_grpo_adapter.py
pointer only (licence: NONE) · get_code("6078af656ef97528")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

run_image_edit Ran
ran; our contract draft was wrong, not the code
jqtangust/Robust-U1/inference/model_setup.py
pointer only (licence: NONE) · get_code("038b60d2c77bf8b2")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

save_reasoning_results Ran
ran; our contract draft was wrong, not the code
jqtangust/Robust-U1/inference/reasoner.py
pointer only (licence: NONE) · get_code("dddcf5ce6a1048ab")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

swish Ran
ran; our contract draft was wrong, not the code
jqtangust/Robust-U1/modeling/modeling/autoencoder.py
pointer only (licence: NONE) · get_code("5fb036462ae82f9f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_2d_sincos_pos_embed_from_grid Not yet run
unverified (harvested, no recorded run)
jqtangust/Robust-U1/modeling/modeling/bagel/modeling_utils.py
pointer only (licence: NONE) · get_code("665d8a4e8f673a4c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

load_ae Not yet run
unverified (harvested, no recorded run)
jqtangust/Robust-U1/modeling/modeling/autoencoder.py
pointer only (licence: NONE) · get_code("734993c851924b68")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

restoration_score Not yet run
unverified (harvested, no recorded run)
jqtangust/Robust-U1/rewards/flow_grpo_adapter.py
pointer only (licence: NONE) · get_code("996ec3da8f220bf2")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

robust_structure_score Not yet run
unverified (harvested, no recorded run)
jqtangust/Robust-U1/rewards/flow_grpo_adapter.py
pointer only (licence: NONE) · get_code("fc21f3bf737b62db")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

structure_score Not yet run
unverified (harvested, no recorded run)
jqtangust/Robust-U1/rewards/rewards.py
pointer only (licence: NONE) · get_code("0e9a007209e3e65c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

tinyclip_score Not yet run
unverified (harvested, no recorded run)
jqtangust/Robust-U1/rewards/rewards.py
pointer only (licence: NONE) · get_code("df4cde9b7ea87f02")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jqtangust/Robust-U1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 6 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 2 the same month, 3 later) and 7 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
get_2d_sincos_pos_embed_from_grid 190 other papers
185 dated earlier, 3 the same month, 2 later · 65 only as an assembly brick
1807.10221 (2018-07)
Unified Perceptual Parsing for Scene Understanding
2105.12723 (2021-05)
Nested Hierarchical Transformer: Towards Accurate, Data-Efficient and Interpretable Visual Understanding
2111.06377 (2021-11)
Masked Autoencoders Are Scalable Vision Learners
and 187 more not named here
get_2d_sincos_pos_embed 69 other papers
66 dated earlier, 1 the same month, 2 later · 12 only as an assembly brick
2203.10897 (2022-03)
Unified Multivariate Gaussian Mixture for Efficient Neural Image Compression
2212.09748 (2022-12)
Scalable Diffusion Models with Transformers
2303.14389 (2023-03)
Masked Diffusion Transformer is a Strong Image Synthesizer
and 66 more not named here
get_1d_sincos_pos_embed_from_grid 42 other papers
41 dated earlier, 1 later · 22 only as an assembly brick
2212.09748 (2022-12)
Scalable Diffusion Models with Transformers
2303.14389 (2023-03)
Masked Diffusion Transformer is a Strong Image Synthesizer
2305.13311 (2023-05)
VDT: General-purpose Video Diffusion Transformers via Mask Modeling
and 39 more not named here
swish 7 other papers
7 dated earlier
2409.00587 (2024-09)
FLUX that Plays Music
2412.07517 (2024-12)
FireFlow: Fast Inversion of Rectified Flow for Image Semantic Editing
2504.02160 (2025-04)
Less-to-More Generalization: Unlocking More Controllability by In-Context Generation
and 4 more not named here
pad_sequence 4 other papers
4 dated earlier
2505.14683 (2025-05)
Emerging Properties in Unified Multimodal Pretraining
2508.05606 (2025-08)
Uni-cot: Towards Unified Chain-of-Thought Reasoning Across Text and Vision
2601.11522 (2026-01)
UniX: Unifying Autoregression and Diffusion for Chest X-Ray Understanding and Generation
and 1 more not named here
load_ae 2 other papers
2 dated earlier
2508.05606 (2025-08)
Uni-cot: Towards Unified Chain-of-Thought Reasoning Across Text and Vision
2605.08724 (2026-05)
SynerMedGen: Synergizing Medical Multimodal Understanding with Generation via Task Alignment

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Multimodal Large Language Models (MLLMs) have demonstrated remarkable success in visual understanding, yet their performance degrades significantly under real-world visual corruptions. While existing robustness enhancement approaches exist, they are limited: black-box feature alignment lacks interpretability, and whitebox text-based reasoning cannot restore lost pixellevel details. This work investigates a fundamental research question: Can MLLMs recover corrupted visual content by themselves? To address this, we propose Robust-U1, a novel framework that equips MLLMs with explicit visual self-recovery capability for robust understanding. The approach comprises three core stages: supervised fine-tuning for initial reconstruction, reinforcement learning with dual rewards (pixellevel SSIM and semantic-level CLIP similarity) for aligning high visual quality, and multimodal reasoning that jointly considers both the corrupted input and the recovered image. Extensive experiments demonstrate that Robust-U1 achieves state-of-the-art robustness on the real-world corruption benchmark and maintains superior performance under adversarial corruptions on general VQA benchmarks. Analysis confirms that high-quality visual recovery directly enhances reasoning performance, establishing self-recovery as a critical mechanism for robust visual understanding. The source code is available at https: //github.com/jqtangust/Robust-U1.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2606.08063")
get_code_for_paper("2606.08063")
have("2606.08063")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2606.08063.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2606.08063.svg)](https://syntology.ai/paper/2606.08063)

Connect an agent — have() is free.