AgentSGEN: Multi-Agent LLM in the Loop for Semantic Collaboration and GENeration of Synthetic Data

David Murphy; Hao Vo; Hoang D. Nguyen; Vu Dinh Xuan

REVIEW 3 major objections 4 minor 1 cited by

AgentSGEN: Multi-Agent LLM in the Loop for Semantic Collaboration and GENeration of Synthetic Data

T0 review · 3 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A two-agent LLM loop (Evaluator plans, Editor executes) edits 3D scenes to satisfy safety goals like blocked doors, beating the Holodeck baseline 38 to 6 in human preference.

desk verdict AgentSGEN is a practically useful multi-agent scene-editing pipeline whose core result—edited scenes beat the Holodeck baseline on blocking-door goals—holds, but the paper overclaims data-utility without a single downstream training experiment. read the letter →

arxiv 2505.13466 v1 pith:UQEL7ACP submitted 2025-05-07 cs.AI

Vu Dinh Xuan , Hao Vo , David Murphy , Hoang D. Nguyen This is my paper · ORCID

classification cs.AI

keywords syntheticdatagenerationmulti-agentLLMscenegrapheditingsafety-criticalscenes3DhumanevaluationLLM-as-judgeconstructionsafety

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that a multi-agent loop of two large language models—an Evaluator that plans scene edits and an Editor that executes them—can turn generic 3D room scenes into safety-critical scenes, such as a bedroom where doors are blocked by large objects, more reliably than a one-shot scene generator. This matters because dangerous situations like blocked fire exits are ethically and practically difficult to photograph, and synthetic scenes that respect safety constraints could fill the training-data gap for safety-critical AI. The supporting evidence is a preference study across 53 indoor scene types: both human annotators and GPT-4.1 chose the collision-aware edited scene over the unedited Holodeck baseline in 38 of 53 comparisons, versus 6 for the baseline. If this holds, iterative LLM-guided scene editing becomes a viable, controllable route to synthetic safety data rather than relying on manual design or procedural rules.

What carries the argument

The central object is the scene graph—a symbolic encoding of objects, positions, orientations, and attributes—treated as an interactive environment rather than a static description. The central mechanism is the Evaluator–Editor loop: a reasoning-heavy LLM (System 2) plans atomic edits from the scene graph and 2D/3D renderings, a fast LLM (System 1) executes them, and an axis-aligned bounding-box collision checker reports validity back to the Editor, with the Evaluator re-entering at termination to validate constraints. SGRender supplies top-down 2D projections and overlay renderings as visual feedback, which is what makes the loop resource-efficient and keeps it free of domain-specific rules.

What would settle it

Compute, for all 53 rooms, the fraction of each exit's free area occluded by blocking objects in the final scene graphs, and check whether every scene that human raters and GPT-4.1 preferred shows a materially higher occlusion fraction than its Holodeck baseline; if many preferred scenes show no greater geometric obstruction, the perceptual preference claim is not evidence of actual safety alignment.

Watch

Extended reading notes

Core claim

On its own terms, AgentSGEN's central discovery is that separating planning from execution in an LLM-driven scene-editing loop produces scenes that satisfy a stated safety goal—'doors are blocked with large objects'—where the one-shot generator fails. The Evaluator agent reasons over a scene graph and multi-modal renderings to produce an action plan; the Editor agent executes atomic move, rotate, and delete operations; and a feedback loop with axis-aligned bounding-box collision checking validates each step. In 53 randomized pairwise comparisons, the collision-aware variant was preferred 38 times by both human annotators and GPT-4.1 (Cohen's kappa 0.406), while Holodeck was preferred only 6 times. The authors conclude that semantic and perceptual objectives are not inherently in conflict and that the pipeline yields reproducible, annotated synthetic data for downstream vision systems.

Load-bearing premise

The load-bearing premise is that human and LLM preferences between side-by-side renderings measure whether the scenes are actually useful as training data for safety-critical AI, but the evaluation never trains or tests a downstream detector on the generated scenes.

Editorial extensions

If this is right

Safety-critical synthetic datasets—blocked-exit images with segmentation, depth, and object metadata—can be generated at scale without photographing real hazards.
The same two-agent loop should generalize to other safety constraints, such as unobstructed evacuation routes or equipment exclusion zones, by changing the goal prompt and constraint set.
LLM judges can serve as auxiliary evaluators for structured scene editing, since GPT-4.1 and Gemini preferences aligned with human preferences in pairwise choice.
Enabling collision checking is a key ingredient of the result: it raised human agreement (kappa 0.406 vs 0.151 without it) and produced the largest preference margin.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A direct test the paper leaves undone: train a downstream object detector or safety monitor on AgentSGEN renders versus Holodeck renders and measure performance on real images; preference judgments alone do not prove transfer, and this is the claim's weakest link.
Because planning and execution are decoupled, the Evaluator could be repurposed for other high-level goals (autonomous driving, warehouse safety) without changing the Editor's atomic operations; this modularity is our inference, not the authors' claim.
A cheaper and more objective evaluation could replace pairwise preference with a geometric obstruction metric: the fraction of a door's free area occluded by blocking objects; if that metric stops matching preferences, the perception-based evidence may overstate true safety alignment.
The authors' own observation that LLM Likert ratings are unstable suggests that future LLM-as-judge designs should stick to discrete-choice protocols, a practical lesson that extends beyond scene generation.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: if you need to generate synthetic indoor scenes where doors are blocked, this is the most straightforward recipe I've seen. The new bit is an Evaluator–Editor loop over a scene graph: one LLM plans (move/rotate/delete) with collision and safety constraints, another executes, and the loop iterates until the scene satisfies the goal. That is a real combination, not just a thin application of Holodeck. The paper is honest about Holodeck's limits, and the Dual Process Theory framing is mostly harmless scaffolding; the architecture is what matters.

What the paper does well: the 53-scene comparison with humans and two LLM judges gives a clear preference signal. Collision-aware edits were preferred by a wide margin (38 vs 6 over the baseline in the binary task), and Cohen's kappa (0.406) is reported for human agreement. That supports the claim that iterative editing improves task satisfaction, at least on static renderings. I'd trust the central preference result; it is strong enough to survive the missing error bars.

Where it is soft: - The data-utility claim is not tested. Section 3.9 says the output is ready for downstream training, but no detector or monitoring model is trained. Whether human/LLM preference transfers to better safety-critical models is unmeasured, so the abstract's bridge-the-gap claim is supported only by implication. - Evaluation reporting is thin: no annotator count, no confidence intervals, one baseline, and no ablations except the collision on/off toggle. The toggle is useful; the rest is thin but fixable. - Section 3.5 has an unresolved "[?]" citation for the collision-checking mechanism. Minor but sloppy. - The internal Evaluator is an LLM and two of the external judges are LLMs; the human anchor keeps this from being circular, but LLM-vs-LLM agreement is weak evidence on its own. The paper mostly acknowledges this in Section 4.4.

The central claim—that an iterative LLM editing loop can produce scenes satisfying a specific safety goal better than one-shot generation—holds up. What does not hold up is the jump to "synthetic data for safety-critical AI." That is a missing experiment, not a contradiction.

Who it's for: anyone building scene-generation pipelines for safety or embodied AI, and anyone refereeing synthetic-data papers. It deserves a serious referee; with one downstream-task experiment (e.g., blocked-door detection or segmentation) and complete reporting it would be much stronger. I'd engage with it.

Referee Report

3 major / 4 minor

Summary. The paper proposes AgentSGEN, a two-agent LLM system for editing 3D indoor scenes generated by Holodeck to satisfy safety goals such as blocking doors with large objects. An Evaluator Agent reasons over a scene graph and visual projections to produce an edit plan; an Editor Agent executes move, rotate, and delete operations; and an axis-aligned bounding box collision checker provides feedback. The evaluation covers 53 indoor scene types from the MIT Indoor Scenes dataset, comparing the edited scenes against the unedited Holodeck baseline through binary preference judgments and goal-oriented Likert ratings from human annotators and from GPT-4.1 and Gemini 2.5 Pro. The reported results show that the collision-aware version is preferred in 38 of 53 pairwise comparisons by both human judges and GPT-4.1, with moderate human inter-annotator agreement, and the paper concludes that the framework can generate synthetic safety-critical data.

Significance. If the reported results hold, the system offers a practical approach to goal-driven, fine-grained editing of synthetic 3D scenes, addressing a genuine bottleneck in safety-critical data generation. The separation of planning and execution is a sensible architectural choice, and the collision-aware variant's strong preference over the Holodeck baseline is a meaningful signal that interactive scene-graph editing can shift perceived semantic alignment. The paper also contributes a reproducible pipeline and reports encouraging human-LLM agreement in preference judgments. However, the significance of the contribution for downstream AI training is not yet established because the evaluation stops at perceptual preference on static renderings: no downstream object detector, safety monitor, or other vision model is trained or evaluated on the generated scenes. The abstract's framing as a solution to data scarcity therefore remains unsupported by the presented evidence.

major comments (3)

[Section 4 (Evaluation) and Section 3.9] The central utility claim is not supported by the evaluation. Section 3.9 states that the generated outputs are 'ready for training or evaluation of downstream computer vision or robotics models,' and the abstract frames the work as a way to bridge the data-scarcity gap. Yet Section 4 evaluates only binary preference and Likert ratings on static renderings; no downstream task is performed. It is therefore unknown whether annotator-preferred and LLM-preferred scenes lead to better performance in, for example, detecting blocked exits or recognizing safety violations. To support the stated contribution, the paper should either add a downstream experiment that trains or evaluates a model on the generated scenes, or substantially narrow the claims to perceptual scene quality rather than synthetic-data utility.
[Section 4.2 and Figure 8] The claim that the system 'significantly improves task satisfaction in human perception' is statistically under-supported. The main text never reports the number of human annotators, and the Figure 8 caption says 'Each entry indicates how often a model was preferred by two annotators,' implying that only two annotators took part in the binary task. No confidence intervals, exact tests, or per-annotator breakdowns are provided for the 38-versus-6 preference count, and the Cohen's kappa values are reported as point estimates without uncertainty. Please report the annotator count, the per-annotator preference data, and at least a binomial confidence interval or exact test; otherwise the 'significantly improves' wording is not justified.
[Section 4.4 and Figure 11] The LLM-based evaluation has a self-referentiality concern that is not fully addressed. The same class of model (LLM) is used inside the system as the Evaluator agent and externally as the automatic judge. Section 4.4 also reports that GPT-4.1 and Gemini preferred the Holodeck baseline over the method on the arrangement and scale criteria, with an attribution to visual priors from web-scale data. This is plausible but not verified. To make the automatic evaluation more convincing, the paper should report the exact prompts and temperatures used for the LLM judges, show the stability of the DCE results across repetitions, and provide a more detailed analysis of where human and LLM judgments diverge, rather than attributing the divergence to unmeasured visual priors.

minor comments (4)

[Section 3.5] There is an unresolved citation placeholder in the sentence describing the 'AABB collision checking mechanism [?].' Please replace it with a proper reference or remove the placeholder.
[Section 5] There are several typographical errors, including 'ann advancement' and 'lastests' in the conclusion; these should be corrected.
[Section 4.1] The annotation protocol is described at a high level but does not state whether the side-by-side renderings were order-balanced, whether annotators saw all 53 room types, or whether the same annotators judged both the binary and Likert tasks. Please specify these details.
[Section 4.4 and Figure 11] Section 4.4 says LLM Likert-style ratings were unstable and were therefore deprioritized, yet Figure 11 is titled 'DCE results' and reports three goal-oriented scores from GPT-4.1 and Gemini. Please clarify how the scores in Figure 11 were obtained and how they relate to the statement about unstable ratings.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the AgentSGEN evaluation is a preference study against an external baseline, not a derived prediction from fitted inputs.

full rationale

The paper does not contain any derivation chain in which a claimed prediction is equivalent to its inputs by construction, nor any fitted parameter that is later renamed as a prediction. The system converts a user goal G into constraints C, uses an Evaluator agent to plan edits and an Editor agent to execute them, then validates the final scene; the evaluation compares the edited scenes against the unedited Holodeck baseline using both human and LLM preference. No parameter is fit to a subset and then 'predicted' on a related subset, and no uniqueness theorem or ansatz is imported from the authors' prior work in a load-bearing way. Reference [22] is a self-citation (the survey includes the corresponding author), but it supports only background claims about multi-agent collaboration modes and is not load-bearing for the central result. The in-loop validation is performed by the same Evaluator that planned the edits, which is a mild self-referentiality, but the paper's central claims rest on external side-by-side comparisons against Holodeck, with human annotators providing an independent anchor; the automatic LLM evaluation is presented as complementary, and the paper notes its limitations (unstable Likert ratings, preference for visually regular layouts). The absence of downstream training or testing of a safety-critical model is a real validity limitation, but it is a gap in the utility claim, not circularity. The evaluation measures task completion as defined by the user goal, which is a direct performance measure rather than a prediction derived from the goal. Therefore the circularity score is 0.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical entities, forces, or mathematical constants. It does introduce an algorithmic structure (Evaluator and Editor agents, scene graph loop) whose behavior depends on hand-chosen thresholds, model choices, and untested assumptions about scene validity and preference-as-proxy. The unresolved citation for the collision checker adds a verification gap.

free parameters (2)

Maximum iteration threshold T = not reported
Section 3.8 introduces a loop bound that determines when scene editing stops. The value is chosen by hand, not reported, and no ablation is given, yet it directly affects the final scenes and all evaluation outcomes.
LLM model choices per agent = Evaluator: O1/O3/o4-mini; Editor: GPT-4o/GPT-4.1
The selection of proprietary LLMs is a design choice not varied or ablated. Plan quality and edit reliability plausibly depend on these choices, so they are effective free parameters of the reported system.

assumptions (4)

domain assumption Holodeck-generated initial scenes are valid, reproducible, and representative of the 53 MIT Indoor Scenes room types.
Section 4 states that each room type's initial scene was generated by Holodeck, but no verification of scene validity, diversity, or reproducibility is provided.
domain assumption Human preference on static renderings is a valid proxy for the usefulness of synthetic data in downstream safety AI training.
The evaluation in Section 4 measures perception only; no downstream detector or monitoring model is trained. The abstract, however, frames the contribution as solving data scarcity for safety-critical AI.
ad hoc to paper Scene graph edit operations (move, rotate, delete) with AABB collision checking are sufficient to represent door-blocking hazards realistically.
Sections 3.4 and 3.5 restrict the Editor to these operations. Whether this action space covers realistic blocked-exit scenarios is not validated against physical scenes or alternative editing methods.
standard math Standard geometry of AABB collision checking.
Section 3.5 relies on 'lightweight AABB collision checking', but the citation marker is unresolved ([?]), so the exact implementation and its correctness are not verifiable from the paper.

how reviews work

0 comments

Cite this review

Pith. "Pith review of AgentSGEN: Multi-Agent LLM in the Loop for Semantic Collaboration and GENeration of Synthetic Data." pith.science (2026). https://pith.science/paper/UQEL7ACP

@misc{pith2026250513466,
  author       = {Pith},
  title        = {Pith review of: AgentSGEN: Multi-Agent LLM in the Loop for Semantic Collaboration and GENeration of Synthetic Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UQEL7ACP}},
  note         = {Machine review of arXiv:2505.13466}
}

read the original abstract

The scarcity of data depicting dangerous situations presents a major obstacle to training AI systems for safety-critical applications, such as construction safety, where ethical and logistical barriers hinder real-world data collection. This creates an urgent need for an end-to-end framework to generate synthetic data that can bridge this gap. While existing methods can produce synthetic scenes, they often lack the semantic depth required for scene simulations, limiting their effectiveness. To address this, we propose a novel multi-agent framework that employs an iterative, in-the-loop collaboration between two agents: an Evaluator Agent, acting as an LLM-based judge to enforce semantic consistency and safety-specific constraints, and an Editor Agent, which generates and refines scenes based on this guidance. Powered by LLM's capabilities to reasoning and common-sense knowledge, this collaborative design produces synthetic images tailored to safety-critical scenarios. Our experiments suggest this design can generate useful scenes based on realistic specifications that address the shortcomings of prior approaches, balancing safety requirements with visual semantics. This iterative process holds promise for delivering robust, aesthetically sound simulations, offering a potential solution to the data scarcity challenge in multimedia safety applications.

Figures

Figures reproduced from arXiv: 2505.13466 by the authors.

**Figure 1.** Comparison of 3D scene generation methods. • Comprehensive evaluations with human raters on randomized images with insights on human and human/AI agreement, demonstrating superior performance in task completion and effectiveness. 2 Related Work 2.1 Synthetic data generation In safety-critical AI systems, synthetic data generation has become indispensable due to the scarcity of real-world examples involving rare o… view at source ↗

**Figure 3.** Multi-modal context components passed to both agents. The context bundle consists of (i) goal-specific constraint requirements, (ii) symbolic scene graph, and (iii) 2D/3D renderings from SGRender. 1. Aligns semantically with G; 2. Satisfies domain-specific safety and spatial constraints C; 3. Meets visual and structural coherence; 4. Supports downstream rendering for synthetic dataset generation. The process begins… view at source ↗

**Figure 2.** Initial scene from Holodeck with corresponding scene graph and SGRender views. enforcement capabilities within a dedicated Evaluator agent and assigning scene modification responsibilities to an Editor agent, our framework mirrors human-like collaborative workflows. This design enables precise, constraint-aware synthetic data generation, advancing the development of reliable and interpretable AI systems for high-s… view at source ↗

Figures from the paper (7 more)

**Figure 4.** Figure 4: AgentSGEN architecture showcasing semantic planning by the Evaluator Agent and iterative scene editing by the Editor Agent [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

**Figure 5.** Figure 5: Examples of Holodeck’s Limitation We applied our editing pipeline under two operational configurations: • With Collision Checking: Enforces physical feasibility and realism by preventing object interpenetration. • Without Collision Checking: Allows free-form object p…

**Figure 6.** Figure 6: Final rendering and synthetic dataset generation using AI2-THOR. Output includes RGB, depth, segmentation, and annotated metadata [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]

**Figure 7.** Figure 7: Binary Task Completion for both human evaluation and automatic evaluation setup 4.4 Automatic Evaluation To complement human judgments, we conducted an automatic evaluation using two latest LLM models: GPT-4.1 and Gemini 2.5 Pro. Each model was presented with side-by…

**Figure 8.** Figure 8: Confusion matrix of binary preference judgments under two collision settings. Each entry indicates how often a model was preferred by two annotators. With collision checking enabled, Cohen’s kappa 0.406 indicates moderate agreement, while disabling collision checking …

**Figure 9.** Figure 9: Average Likert scores (1–7 scale) for three goal-oriented evaluation questions. Our collision-aware method outperforms both the collisiondisabled version and the Holodeck baseline across all metrics [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]

**Figure 10.** Figure 10: Confusion matrix of LLM judgments under two collision settings. 4.5 Discussion Our evaluation reveals that AgentSGEN meaningfully advances the state of synthetic scene editing by introducing fine-grained semantic control while preserving visual plausibility. Unlike …

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

TalkPlayData 2: An Agentic Synthetic Data Pipeline for Multimodal Conversational Music Recommendation
cs.IR 2025-08 conditional novelty 5.0 of 10

TalkPlayData 2 is a new synthetic conversational music recommendation dataset created by four cooperating multimodal LLM agents, featuring user profiles, conversation goals, chain-of-thought annotations, and cold-star...

Reference graph

Works this paper leans on

28 extracted references · 18 canonical work pages · cited by 1 Pith paper

[1]

S. C. Bellini-Leite. Dual process theory: Embodied and predictive; sym- bolic and classical. Frontiers in Psychology, 13:805386, 2022. doi: 10.3389/fpsyg.2022.805386. URL https://www.frontiersin.org/articles/ 10.3389/fpsyg.2022.805386/full

work page arXiv 2022
[2]

Blattmann, T

A. Blattmann, T. Dockhorn, S. Kulal, D. Mendelevitch, M. Kilian, D. Lorenz, Y . Levi, Z. English, V . V oleti, A. Letts, et al. Stable video diffusion: Scaling latent video diffusion models to large datasets. arXiv preprint arXiv:2311.15127, 2023

arXiv 2023
[3]

Bochkovskiy, C.-Y

A. Bochkovskiy, C.-Y . Wang, and H.-Y . M. Liao. Yolov4: Op- timal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934, 2020

arXiv 2004
[4]

W. Ding, C. Xu, M. Arief, H. Lin, B. Li, and D. Zhao. A survey on safety-critical driving scenario generation—a methodological perspec- tive. IEEE Transactions on Intelligent Transportation Systems , 24(7): 6971–6988, 2023

work page 2023
[5]

Esser, R

P. Esser, R. Rombach, and B. Ommer. Taming transformers for high- resolution image synthesis. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12873–12883, 2021

work page 2021
[6]

Esser, S

P. Esser, S. Kulal, A. Blattmann, R. Entezari, J. Müller, H. Saini, Y . Levi, D. Lorenz, A. Sauer, F. Boesel, et al. Scaling rectified flow transformers for high-resolution image synthesis. In Forty-first international confer- ence on machine learning, 2024

work page 2024
[7]

Hagendorff, S

T. Hagendorff, S. Fabi, and M. Kosinski. Human-like intuitive behavior and reasoning biases emerged in large language models but disappeared in chatgpt. Nature Computational Science, 3(10):833–838, 2023

work page 2023
[8]

S. Hong, X. Zheng, J. Chen, Y . Cheng, J. Wang, C. Zhang, Z. Wang, S. K. S. Yau, Z. Lin, L. Zhou, et al. Metagpt: Meta programming for multi-agent collaborative framework.arXiv preprint arXiv:2308.00352, 3(4):6, 2023

arXiv 2023

Show all 28 references

[9]

Kawar, S

B. Kawar, S. Zada, O. Lang, O. Tov, H. Chang, T. Dekel, I. Mosseri, and M. Irani. Imagic: Text-based real image editing with diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 6007–6017, 2023

2023
[10]

Kim and J.-S

H. Kim and J.-S. Yi. Image generation of hazardous situations in con- struction sites using text-to-image generative model for training deep neural networks. Automation in Construction, 166:105615, 2024

2024
[11]

H. Lee, J. Jeon, D. Lee, C. Park, J. Kim, and D. Lee. Game engine- driven synthetic data generation for computer vision-based safety mon- itoring of construction workers. Automation in Construction , 155: 105060, 2023

2023
[12]

Z. Liu, Y . Zhang, P. Li, Y . Liu, and D. Yang. Dynamic llm-agent net- work: An llm-agent collaboration framework with agent team optimiza- tion. arXiv preprint arXiv:2310.02170, 2023

2023 arXiv
[13]

C. Mou, X. Wang, L. Xie, Y . Wu, J. Zhang, Z. Qi, and Y . Shan. T2i- adapter: Learning adapters to dig out more controllable ability for text- to-image diffusion models. In Proceedings of the AAAI conference on artificial intelligence, volume 38, pages 4296–4304, 2024

2024
[14]

Neuhausen, P

M. Neuhausen, P. Herbers, and M. König. Synthetic data for evaluating the visual tracking of construction workers. In Construction Research Congress 2020, pages 354–361. American Society of Civil Engineers Reston, V A, 2020

2020
[15]

Podell, Z

D. Podell, Z. English, K. Lacey, A. Blattmann, T. Dockhorn, J. Müller, J. Penna, and R. Rombach. Sdxl: Improving latent diffusion models for high-resolution image synthesis. arXiv preprint arXiv:2307.01952, 2023

2023 arXiv
[16]

Quattoni and A

A. Quattoni and A. Torralba. Recognizing indoor scenes. In 2009 IEEE Conference on Computer Vision and Pattern Recognition , pages 413–

2009
[17]

Raistrick, L

A. Raistrick, L. Mei, K. Kayan, D. Yan, Y . Zuo, B. Han, H. Wen, M. Parakh, S. Alexandropoulos, L. Lipson, et al. Infinigen indoors: Pho- torealistic indoor scenes using procedural generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- ti...

2024
[18]

Rombach, A

R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer. High- resolution image synthesis with latent diffusion models. InProceedings of the IEEE/CVF conference on computer vision and pattern recogni- tion, pages 10684–10695, 2022

2022
[19]

Sauer, D

A. Sauer, D. Lorenz, A. Blattmann, and R. Rombach. Adversarial dif- fusion distillation. In European Conference on Computer Vision, pages 87–103. Springer, 2024

2024
[20]

Shorten and T

C. Shorten and T. M. Khoshgoftaar. A survey on image data augmenta- tion for deep learning. Journal of big data, 6(1):1–48, 2019

2019
[21]

Z. Song, Z. He, X. Li, Q. Ma, R. Ming, Z. Mao, H. Pei, L. Peng, J. Hu, D. Yao, et al. Synthetic datasets for autonomous driving: A survey. IEEE Transactions on Intelligent Vehicles, 9(1):1847–1864, 2023

2023
[22]

K.-T. Tran, D. Dao, M.-D. Nguyen, Q.-V . Pham, B. O’Sullivan, and H. D. Nguyen. Multi-agent collaboration mechanisms: A survey of llms. arXiv preprint arXiv:2501.06322, 2025

2025 arXiv
[23]

L. Wang, C. Ma, X. Feng, Z. Zhang, H. Yang, J. Zhang, Z. Chen, J. Tang, X. Chen, Y . Lin, et al. A survey on large language model based autonomous agents. Frontiers of Computer Science , 18(6):186345, 2024

2024
[24]

Yang, F.-Y

Y . Yang, F.-Y . Sun, L. Weihs, E. VanderBilt, A. Herrasti, W. Han, J. Wu, N. Haber, R. Krishna, L. Liu, et al. Holodeck: Language guided genera- tion of 3d embodied ai environments. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16...

2024
[25]

Zhang, M

H. Zhang, M. Cisse, Y . N. Dauphin, and D. Lopez-Paz. mixup: Beyond empirical risk minimization. arXiv preprint arXiv:1710.09412, 2017

2017 arXiv
[26]

Zhang, A

L. Zhang, A. Rao, and M. Agrawala. Adding conditional control to text-to-image diffusion models. In Proceedings of the IEEE/CVF inter- national conference on computer vision, pages 3836–3847, 2023

2023
[27]

Zhang, L

Z. Zhang, L. Han, A. Ghosh, D. N. Metaxas, and J. Ren. Sine: Single image editing with text-to-image diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, pages 6027–6037, 2023

2023
[420]

doi: 10.1109/CVPR.2009.5206537

IEEE, 2009. doi: 10.1109/CVPR.2009.5206537

2009

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

S. C. Bellini-Leite. Dual process theory: Embodied and predictive; sym- bolic and classical. Frontiers in Psychology, 13:805386, 2022. doi: 10.3389/fpsyg.2022.805386. URL https://www.frontiersin.org/articles/ 10.3389/fpsyg.2022.805386/full

work page arXiv 2022

[2] [2]

Blattmann, T

A. Blattmann, T. Dockhorn, S. Kulal, D. Mendelevitch, M. Kilian, D. Lorenz, Y . Levi, Z. English, V . V oleti, A. Letts, et al. Stable video diffusion: Scaling latent video diffusion models to large datasets. arXiv preprint arXiv:2311.15127, 2023

arXiv 2023

[3] [3]

Bochkovskiy, C.-Y

A. Bochkovskiy, C.-Y . Wang, and H.-Y . M. Liao. Yolov4: Op- timal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934, 2020

arXiv 2004

[4] [4]

W. Ding, C. Xu, M. Arief, H. Lin, B. Li, and D. Zhao. A survey on safety-critical driving scenario generation—a methodological perspec- tive. IEEE Transactions on Intelligent Transportation Systems , 24(7): 6971–6988, 2023

work page 2023

[5] [5]

Esser, R

P. Esser, R. Rombach, and B. Ommer. Taming transformers for high- resolution image synthesis. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12873–12883, 2021

work page 2021

[6] [6]

Esser, S

P. Esser, S. Kulal, A. Blattmann, R. Entezari, J. Müller, H. Saini, Y . Levi, D. Lorenz, A. Sauer, F. Boesel, et al. Scaling rectified flow transformers for high-resolution image synthesis. In Forty-first international confer- ence on machine learning, 2024

work page 2024

[7] [7]

Hagendorff, S

T. Hagendorff, S. Fabi, and M. Kosinski. Human-like intuitive behavior and reasoning biases emerged in large language models but disappeared in chatgpt. Nature Computational Science, 3(10):833–838, 2023

work page 2023

[8] [8]

S. Hong, X. Zheng, J. Chen, Y . Cheng, J. Wang, C. Zhang, Z. Wang, S. K. S. Yau, Z. Lin, L. Zhou, et al. Metagpt: Meta programming for multi-agent collaborative framework.arXiv preprint arXiv:2308.00352, 3(4):6, 2023

arXiv 2023