Pith. sign in

REVIEW 4 major objections 5 minor 49 references

GOBench: Benchmarking Geometric Optics Generation and Understanding of MLLMs

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A new benchmark finds that state-of-the-art multimodal models score barely above chance at judging shadows, reflections, and refraction, and still err when generating them.

desk verdict A useful first benchmark for geometric optics in MLLMs; the qualitative finding holds, but the 37.35% headline is an SPS proximity score, not accuracy, and expert-agreement data are missing. read the letter →

arxiv 2506.00991 v2 pith:7IQ3NNX6 submitted 2025-06-01 cs.CV

Xiaorong Zhu , Ziheng Jia , Jiarui Wang , Xiangyu Zhao , Haodong Duan , Xiongkuo Min , Jia Wang , Zicheng Zhang
show 1 more author
Guangtao Zhai
This is my paper
classification cs.CV
keywords geometricopticsmultimodallargelanguagemodelsopticalauthenticityimagegenerationevaluationvisualunderstandingshadowrenderingreflectionandrefractionScaledProximityScore
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

GOBench is the first benchmark to test multimodal large language models (MLLMs) on geometric optics, both as image generators and as image understanders. The paper builds GOBench-Gen-1k, a dataset of 1,000 images generated by GPT-4o-Image, Seedream 3.0, and Imagen 3 across 340 scenarios covering direct light, reflection, and refraction, and has six experts score each image for Optical Authenticity, Aesthetic Quality, and Instruction Fidelity. On the generation side, the paper reports that even the top model, GPT-4o-Image, does not perfectly complete all generation tasks and shows flaws that violate optical principles. On the understanding side, eleven MLLMs are asked the same questions as the experts, and the best model, Gemini-2.5Pro, reaches only 37.35% Scaled Proximity Score on Optical Authenticity, which the paper reads as evidence that current models lack robust physical understanding of optics.

What carries the argument

The load-bearing object is the GOBench-Gen-1k dataset plus its evaluation protocol. The dataset is built from 340 curated prompts (108 direct-light, 121 reflection, 111 refraction), each with a tailored five-question Optical Authenticity rubric (Yes/No/Cannot be determined), and each scenario is rendered by GPT-4o-Image, Seedream 3.0, and Imagen 3. The metric that carries the understanding claim is the Scaled Proximity Score (SPS), which linearly credits an MLLM's rating for being within a maximum permitted difference $\delta_{\max} = 0.5$ of the averaged expert score per case, then averages across cases.

What would settle it

Re-run the expert panel with a fresh set of six physicists on the same 1k images and compute inter-rater agreement; independently ray-trace the 340 prompts to produce physical ground truth. If expert scores disagree with each other or with the simulated optics, the SPS numbers and the claim that MLLMs fail at optical understanding are not established.

Watch

Extended reading notes

Core claim

The paper's central claim is that state-of-the-art MLLMs can generate visually appealing images yet fail to respect geometric optics, and that they cannot evaluate optical phenomena the way trained human experts do. The evidence is a benchmark in which 1k images from three generative models are scored by six human experts using five per-scenario Optical Authenticity questions based on rectilinear propagation, reflection, and Snell's law, and in which 11 MLLMs are given the same questions as a test of understanding. The reported numbers—GPT-4o-Image's authenticity score of 4.09/5, Gemini-2.5Pro's 37.35% SPS on authenticity against expert ground truth—are intended to quantify the gap between current capability and physical correctness.

Load-bearing premise

The whole benchmark depends on the assumption that six experts' answers to the authors' Optical Authenticity questions are a valid and reliable ground truth for physical correctness, yet no inter-rater reliability measure is reported and the questions were not validated against an external physical ground truth such as ray-traced renders.

Editorial extensions

If this is right

  • Until models are trained with explicit physical supervision, optical generation errors such as double shadows, impossible shadow angles, and discontinuous refracted beams will persist in high-fidelity content creation.
  • The 37% SPS ceiling implies current MLLMs are not reliable judges of their own or others' optical realism, so automated filtering of AI-generated images for physical plausibility cannot yet be trusted.
  • The low correlation between Aesthetics and Authenticity (PLCC 0.2213) implies that looking good and physically right are separate axes, so improving visual appeal will not by itself fix physical realism.
  • Benchmark scores on GOBench can serve as a target metric for future physically grounded training objectives, giving a concrete number to optimize beyond human preference ratings.

Reading between the lines

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

  • If the expert gold standard were replaced by a ray-traced physical simulation of each prompt, the SPS numbers could be recomputed as a direct measure of physical accuracy, which might shift model rankings if expert judgment disagrees with the renderer.
  • Because the benchmark's questions are binary (Yes/No/Cannot determine), the 37% SPS may underestimate models' latent optical knowledge; a free-response or part-scored rubric might reveal partial understanding.
  • The same scenario design could be extended to video frames, where dynamic light behavior (e.g., moving shadows, caustics) would test whether MLLMs understand optics as a process rather than a static pictorial feature.
  • A testable extension is to fine-tune an open MLLM on GOBench-Gen-1k with the expert labels as supervision; if scores jump after fine-tuning, the optics gap is at least partly learnable from visual data alone.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper introduces GOBench, a benchmark for evaluating multi-modal large language models (MLLMs) on two tasks: generating images that obey geometric optics principles and understanding optical phenomena in images. The authors construct GOBench-Gen-1k, a dataset of 1,000 images generated by three MLLMs (GPT-4o-Image, Seedream 3.0, Imagen 3) from 340 curated prompts across three optical categories (direct light, reflection, refraction). A panel of six human experts scores each image on Optical Authenticity, Aesthetic Quality, and Instruction Fidelity, and the averaged expert scores serve as ground truth. Eleven MLLMs are then prompted to score the same images on the same dimensions, and their outputs are compared against the human ground truth using a Scaled Proximity Score (SPS) with a hand-set tolerance delta_max = 0.5. The main reported findings are that GPT-4o-Image achieves the highest generation Authenticity score (4.09/5) but still makes optical errors, and that the best understanding model, Gemini-2.5Pro, attains only 37.35% SPS on Optical Authenticity, leading the authors to conclude that current MLLMs face significant challenges in both optical generation and understanding.

Significance. If the measurement is trustworthy, GOBench would be a useful first systematic benchmark for geometric optics in MLLMs, an underexplored but visually important domain. The paper has concrete strengths: the scenario taxonomy is sensible, the dataset and code are promised to be public, the evaluation protocol is described in enough detail to be partially reproducible, and the qualitative conclusion that MLLMs struggle with fine-grained optical phenomena is plausible and consistent with related perception benchmarks. However, the quantitative headline results are fragile because they rest entirely on the reliability of a six-expert subjective ground truth and on the chosen SPS tolerance; neither is validated or characterized. The abstract's characterization of SPS as 'accuracy' overstates what the metric measures. The significance of the numerical rankings in Table 3 therefore cannot currently be assessed.

major comments (4)
  1. [Section 2.2 and Section 3.3] The central quantitative claim, that Gemini-2.5Pro attains only 37.35% on optical understanding, depends on treating the six-expert average as ground truth, but the paper reports no inter-rater reliability statistic (e.g., Fleiss' kappa, ICC, or per-image expert variance). Since the SPS gives zero credit whenever an MLLM's score differs from the averaged expert score by more than delta_max = 0.5, large expert disagreement directly deflates all model scores. If the six experts themselves disagree by more than 0.5 points on a meaningful fraction of images, then low SPS values may reflect expert noise rather than an optical understanding deficit. Please report per-image expert score distributions and an inter-rater reliability measure, and show how the conclusions change under alternative ground-truth aggregation rules (e.g., median or per-expert agreement).
  2. [Equation (1), Abstract, Section 3.3] The Scaled Proximity Score is not accuracy, but the abstract and Section 3.3 describe the 37.35% figure as 'accuracy.' SPS is a tolerance-weighted proximity to a subjective mean, and its scale depends on the arbitrarily chosen delta_max = 0.5. A model that exactly matches one expert but differs from the averaged ground truth by more than 0.5 receives zero for that case, so the numerical value is not interpretable as percent correct. Please report standard per-question accuracy on the Yes/No/Cannot-determine answers as a separate metric, provide a sensitivity analysis over delta_max, and state the chance-level baseline for SPS given the three-option response format.
  3. [Section 2.1 and Section 2.2] The ground truth for physical optical correctness is entirely based on the judgments of six experts viewing MLLM-generated images that the authors manually filtered. There is no external validation against physically grounded references, such as ray-traced renders of the same prompts or known-correct photographs. Without such validation, the benchmark risks measuring 'agreement with six experts' rather than 'agreement with optical physics.' At minimum, please describe the manual filtering criteria quantitatively, report how many images were removed per model, and validate a random subset of expert judgments against an external physical ground truth or against a second independent panel.
  4. [Section 2.2, 'Cannot be determined' and Section 3.3] The scoring of 'Cannot be determined' as 0.5 points is arbitrary and may systematically affect both the human ground truth and the MLLM comparison. For instance, if experts use this option for genuinely ambiguous images, the averaged ground-truth score will be pulled toward 0.5, while MLLMs that commit to Yes or No will be penalized under SPS. Please report the frequency of 'Cannot be determined' responses for both experts and each MLLM, and analyze whether the SPS ranking is robust to excluding such cases or to treating them as a separate response category.
minor comments (5)
  1. [Throughout] The manuscript contains several typographical and grammatical errors, e.g., 'We curates high-quality prompts' (abstract), 'This involve 340 unique scenarios' (Section 1), and 'generated by by GPT-4o-Image' (Section 2.2). A careful proofreading pass is needed.
  2. [Section 2.1 and Table 2] The generative model name is written inconsistently as 'SeeDream3.0' and 'Seedream 3.0.' Please use one consistent name.
  3. [References] Some citations do not match their referents: reference [24] is cited for Grok-2-Vision but appears to be a paper on AI adoption, and reference [17] is cited for GLM4V but points to CogVLM2. Please verify all model-specific citations.
  4. [Table 3] The differences between adjacent models in the Authenticity column are small (e.g., 31.53% vs. 31.93%), and no confidence intervals or significance tests are reported. Please add error bars or statistical comparisons before claiming a strict ranking.
  5. [Section 3.1] The statement that an inter-dimension correlation of 0.70 indicates the dimensions are 'largely independent' is overstated; a correlation of 0.7 is substantial and suggests Aesthetic Quality and Instruction Fidelity share considerable variance. Please soften the interpretation or provide a statistical test for independence.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: GOBench is an empirical benchmark whose reported scores are measurements against human ground truth, not fitted or self-referential derivations.

full rationale

The paper's central claims are empirical benchmark results. The derivation chain is: (1) six experts score MLLM-generated images on Optical Authenticity, Aesthetic Quality, and Instruction Fidelity; (2) the averaged expert scores become ground truth; (3) eleven MLLMs are prompted to give the same scores; and (4) their outputs are compared to the human ground truth via the Scaled Proximity Score (Eq. 1). No parameter is fitted to the MLLM outputs, no model is trained or calibrated on the target data, and no prior result by the authors is invoked to define or justify the measured 37.35% SPS for Gemini-2.5Pro. The SPS is a fixed scoring rule with delta_max = 0.5, applied to both human and model ratings, so the reported accuracies are direct measurements rather than consequences of an imposed relation. The benchmark design does have validity concerns — expert agreement is unreported, the Optical Authenticity questions are author-authored rather than validated against external physical ground truth, and the understanding test images are themselves MLLM-generated and author-filtered — but these concern whether the measurement is meaningful, not whether it is circular. There is a self-citation (AIBench, [46]) in the reference list, but it is not load-bearing for GOBench's own construction or findings. The paper does not rename a known result, smuggle an ansatz through a citation, or reduce any prediction to its inputs by construction. Thus no circular step meeting the evidentiary standard can be identified.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The benchmark relies on hand-authored evaluation rubrics and hand-chosen scoring constants. There are no new physical entities. The central claim depends on the validity of the human ground truth and the correctness of the question design.

free parameters (2)
  • delta_max (SPS tolerance) = 0.5
    Hand-chosen in Equation 1 (Section 2.3) as the maximum allowed absolute difference for partial credit; all SPS scores and model rankings depend on this value.
  • Cannot-determine score weight = 0.5
    Section 2.2 assigns 0.5 points to 'Cannot be determined' answers on Optical Authenticity questions; this directly shifts both human ground truth and model scores.
assumptions (3)
  • domain assumption Human expert consensus on author-written authenticity questions is ground truth for optical correctness.
    Section 2.2 uses the average of six experts as the gold standard for both generation and understanding evaluation; no external physical ground truth such as ray-traced renders is used.
  • domain assumption The five questions per scenario correctly and sufficiently instantiate geometric optics laws (rectilinear propagation, reflection, Snell's law).
    Section 2.2 and Figure 4; if the questions are ambiguous, leading, or incomplete, all authenticity scores are affected.
  • domain assumption The taxonomy of Direct Light, Reflection, and Refraction, with three subcategories each, covers the relevant space of geometric optics phenomena.
    Section 2.1; scenarios are designed by the authors, not sampled from a documented distribution of real optical scenes.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GOBench: Benchmarking Geometric Optics Generation and Understanding of MLLMs." pith.science (2026). https://pith.science/paper/7IQ3NNX6

@misc{pith2026250600991,
  author       = {Pith},
  title        = {Pith review of: GOBench: Benchmarking Geometric Optics Generation and Understanding of MLLMs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7IQ3NNX6}},
  note         = {Machine review of arXiv:2506.00991}
}
read the original abstract

The rapid evolution of Multi-modality Large Language Models (MLLMs) is driving significant advancements in visual understanding and generation. Nevertheless, a comprehensive assessment of their capabilities, concerning the fine-grained physical principles especially in geometric optics, remains underexplored. To address this gap, we introduce GOBench, the first benchmark to systematically evaluate MLLMs' ability across two tasks: 1) Generating Optically Authentic Imagery and 2) Understanding Underlying Optical Phenomena. We curates high-quality prompts of geometric optical scenarios and use MLLMs to construct GOBench-Gen-1k dataset.We then organize subjective experiments to assess the generated imagery based on Optical Authenticity, Aesthetic Quality, and Instruction Fidelity, revealing MLLMs' generation flaws that violate optical principles. For the understanding task, we apply crafted evaluation instructions to test optical understanding ability of eleven prominent MLLMs. The experimental results demonstrate that current models face significant challenges in both optical generation and understanding. The top-performing generative model, GPT-4o-Image, cannot perfectly complete all generation tasks, and the best-performing MLLM model, Gemini-2.5Pro, attains a mere 37.35\% accuracy in optical understanding. Database and codes are publicly available at https://github.com/aiben-ch/GOBench.

Figures

Figures reproduced from arXiv: 2506.00991 by the authors.

Figure 1
Figure 1. We propose GOBench, the first benchmark on emerging abilities of MLLMs on geometric optical generation and [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Task Distribution of GOBench-Gen-1K, involves [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Examples of GOBench-Gen-1k that show cases of designed scenarios, including direct light scenario, reflect light scenario and refracted light scenario. Each case includes the input prompt, and the output image generated by state-of-art MLLM. The red words represent the obvious flaws of the MLLM’s generations that violating optical or basic physical principles. geometric optical scenarios in three foundational catego… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Optical Authenticity questions of GOBench. The [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Examples of several different models’ answers to the optical authentic questions.GT represents ground truth result; [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

49 extracted references · 19 canonical work pages

  1. [1]

    Danial Abshari, Chenglong Fu, and Meera Sridhar. 2024. LLM-assisted Physical Invariant Extraction for Cyber-Physical Systems Anomaly Detection. arXiv preprint arXiv:2411.10918 (2024)

    arXiv 2024

  2. [2]

    Stanislaw Antol, Aishwarya Agrawal, Jiasen Lu, Margaret Mitchell, Dhruv Batra, C Lawrence Zitnick, and Devi Parikh. 2015. Vqa: Visual question answering. In Proceedings of the IEEE international conference on computer vision . 2425–2433

    2015

  3. [3]

    Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xiaodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, et al. 2023. Qwen technical report. arXiv preprint arXiv:2309.16609 (2023)

    arXiv 2023

  4. [4]

    Jonathan T Barron and Jitendra Malik. 2012. Shape, albedo, and illumination from a single image of an unknown object. In 2012 IEEE Conference on Computer Vision and Pattern Recognition . IEEE, 334–341

    work page 2012

  5. [5]

    James Betker, Gabriel Goh, Li Jing, Tim Brooks, Jianfeng Wang, Linjie Li, Long Ouyang, Juntang Zhuang, Joyce Lee, Yufei Guo, et al . 2023. Improving im- age generation with better captions. Computer Science. https://cdn. openai. com/papers/dall-e-3. pdf 2, 3 (2023), 8

    2023

  6. [6]

    Xiao Bi, Deli Chen, Guanting Chen, Shanhuang Chen, Damai Dai, Chengqi Deng, Honghui Ding, Kai Dong, Qiushi Du, Zhe Fu, et al. 2024. Deepseek llm: Scaling open-source language models with longtermism. arXiv preprint arXiv:2401.02954 (2024)

    arXiv 2024

  7. [7]

    Hong Chen, Xin Wang, Yuwei Zhou, Bin Huang, Yipeng Zhang, Wei Feng, Houlun Chen, Zeyang Zhang, Siao Tang, and Wenwu Zhu. 2024. Multi-modal generative ai: Multi-modal llm, diffusion and beyond. arXiv preprint arXiv:2409.14993 (2024)

    arXiv 2024

  8. [8]

    Zhe Chen, Weiyun Wang, Hao Tian, Shenglong Ye, Zhangwei Gao, Erfei Cui, Wenwen Tong, Kongzhi Hu, Jiapeng Luo, Zheng Ma, Ji Ma, Jiaqi Wang, Xiaoyi Dong, Hang Yan, Hewei Guo, Conghui He, Botian Shi, Zhenjiang Jin, Chao Xu, Bin Wang, Xingjian Wei, Wei Li, Wenjian Zhang, Bo Zhang, Pinlong Cai, Licheng Wen, Xiangchao Yan, Min Dou, Lewei Lu, Xizhou Zhu, Tong Lu...

    arXiv 2024

Show all 49 references
  1. [9]

    Anoop Cherian, Radu Corcodel, Siddarth Jain, and Diego Romeres. 2024. Llmphy: Complex physical reasoning using large language models and world models. arXiv preprint arXiv:2411.08027 (2024)

    2024 arXiv

  2. [10]

    Nicholas Crafts. 2021. Artificial intelligence as a general-purpose technology: an historical perspective. University of Sussex (2021)

    2021

  3. [11]

    Luciano Floridi and Massimo Chiriatti. 2020. GPT-3: Its nature, scope, limits, and consequences. Minds and Machines 30 (2020), 681–694

    2020

  4. [12]

    Xingyu Fu, Yushi Hu, Bangzheng Li, Yu Feng, Haoyu Wang, Xudong Lin, Dan Roth, Noah A Smith, Wei-Chiu Ma, and Ranjay Krishna. 2024. Blink: Multimodal large language models can see but not perceive. In European Conference on Computer Vision. Springer, 148–166

    2024

  5. [13]

    Hanan Gani, Shariq Farooq Bhat, Muzammal Naseer, Salman Khan, and Peter Wonka. 2023. Llm blueprint: Enabling text-to-image generation with complex and detailed prompts. arXiv preprint arXiv:2310.10640 (2023)

    2023 arXiv

  6. [14]

    Yu Gu, Kai Zhang, Yuting Ning, Boyuan Zheng, Boyu Gou, Tianci Xue, Cheng Chang, Sanjari Srivastava, Yanan Xie, Peng Qi, et al. 2024. Is your llm secretly a world model of the internet? model-based planning for web agents. arXiv preprint arXiv:2411.06559 (2024)

    2024 arXiv

  7. [15]

    ANDRÉ GUIDETTI. 2019. Artificial intelligence as general purpose technology: an empirical and applied analysis of its perception. (2019)

    2019

  8. [16]

    Minjie Hong, Yan Xia, Zehan Wang, Jieming Zhu, Ye Wang, Sihang Cai, Xi- aoda Yang, Quanyu Dai, Zhenhua Dong, Zhimeng Zhang, et al. 2025. EAGER- LLM: Enhancing Large Language Models as Recommenders through Exogenous Behavior-Semantic Integration. In Proceedings of the ACM on We...

    2025

  9. [17]

    Wenyi Hong, Weihan Wang, Ming Ding, Wenmeng Yu, Qingsong Lv, Yan Wang, Yean Cheng, Shiyu Huang, Junhui Ji, Zhao Xue, et al. 2024. Cogvlm2: Visual lan- guage models for image and video understanding.arXiv preprint arXiv:2408.16500 (2024)

    2024 arXiv

  10. [18]

    Hao-Yu Hsu, Zhi-Hao Lin, Albert Zhai, Hongchi Xia, and Shenlong Wang. 2024. AutoVFX: Physically Realistic Video Editing from Natural Language Instructions. arXiv preprint arXiv:2411.02394 (2024)

    2024 arXiv

  11. [19]

    Jun Hu, Wenwen Xia, Xiaolu Zhang, Chilin Fu, Weichang Wu, Zhaoxin Huan, Ang Li, Zuoli Tang, and Jun Zhou. 2024. Enhancing sequential recommendation via llm-based semantic embedding learning. In Companion Proceedings of the ACM Web Conference 2024. 103–111

    2024

  12. [20]

    Qingyong Hu, Bo Yang, Sheikh Khalid, Wen Xiao, Niki Trigoni, and Andrew Markham. 2021. Towards semantic segmentation of urban-scale 3D point clouds: A dataset, benchmarks and challenges. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 4977–4987

    2021

  13. [21]

    Wenbo Hu, Yifan Xu, Yi Li, Weiyue Li, Zeyuan Chen, and Zhuowen Tu. 2024. Bliva: A simple multimodal llm for better handling of text-rich visual questions. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 38. 2256–2264

    2024

  14. [22]

    Maeve Hutchinson, Radu Jianu, Aidan Slingsby, and Pranava Madhyastha. 2024. LLM-Assisted Visual Analytics: Opportunities and Challenges. arXiv preprint arXiv:2409.02691 (2024)

    2024 arXiv

  15. [23]

    Michael P Keating. 1988. Geometric, physical, and visual optics . Elsevier Health Sciences

    1988

  16. [24]

    Smith Lee. 2025. Bridging the AI Adoption Gap: The Disparity Between Rapid Technological Advancements and Corporate Adaptation. (2025)

    2025

  17. [25]

    Zhihao Li, Yao Du, Yang Liu, Yan Zhang, Yufang Liu, Mengdi Zhang, and Xunliang Cai. 2024. Eagle: Elevating geometric reasoning through llm-empowered visual instruction tuning. arXiv preprint arXiv:2408.11397 (2024)

    2024

  18. [26]

    Zijing Liang, Yanjie Xu, Yifan Hong, Penghui Shang, Qi Wang, Qiang Fu, and Ke Liu. 2024. A Survey of Multimodel Large Language Models. In Proceedings of the 3rd International Conference on Computer, Artificial Intelligence and Control Engineering. 405–409

    2024

  19. [27]

    Hao Liu, Jiarui Feng, Lecheng Kong, Ningyue Liang, Dacheng Tao, Yixin Chen, and Muhan Zhang. 2023. One for all: Towards training one graph model for all classification tasks. arXiv preprint arXiv:2310.00149 (2023)

    2023 arXiv

  20. [28]

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. 2023. Visual instruc- tion tuning. arXiv preprint arXiv:2304.08485 (2023)

    2023 arXiv

  21. [29]

    David S Loshin. 2015. The geometrical optics workbook . Elsevier Health Sciences

    2015

  22. [30]

    Pan Lu, Baolin Peng, Hao Cheng, Michel Galley, Kai-Wei Chang, Ying Nian Wu, Song-Chun Zhu, and Jianfeng Gao. 2023. Chameleon: Plug-and-play composi- tional reasoning with large language models. Advances in Neural Information Processing Systems 36 (2023), 43447–43478

    2023

  23. [31]

    Gen Luo, Xue Yang, Wenhan Dou, Zhaokai Wang, Jifeng Dai, Yu Qiao, and Xizhou Zhu. 2024. Mono-internvl: Pushing the boundaries of monolithic multimodal large language models with endogenous visual pre-training. arXiv preprint arXiv:2410.08202 (2024)

    2024 arXiv

  24. [32]

    Chuofan Ma, Yi Jiang, Jiannan Wu, Zehuan Yuan, and Xiaojuan Qi. 2024. Groma: Localized visual tokenization for grounding multimodal large language models. In European Conference on Computer Vision . Springer, 417–435

    2024

  25. [33]

    AI Meta. 2024. Introducing meta llama 3: The most capable openly available llm to date. Meta AI 2, 5 (2024), 6

    2024

  26. [34]

    Muhammad Ahmed Mohsin, Ahsan Bilal, Sagnik Bhattacharya, and John M Cioffi. 2025. Retrieval augmented generation with multi-modal llm framework for wireless environments. arXiv preprint arXiv:2503.07670 (2025)

    2025 arXiv

  27. [35]

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas Müller, Joe Penna, and Robin Rombach. 2023. Sdxl: Improving latent diffusion models for high-resolution image synthesis. arXiv preprint arXiv:2307.01952 (2023)

    2023 arXiv

  28. [36]

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. 2022. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 10684–10695

    2022

  29. [37]

    Gemini Team, Rohan Anil, Sebastian Borgeaud, Yonghui Wu, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M Dai, Anja Hauth, et al. 2023. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805 (2023)

    2023 arXiv

  30. [38]

    Ayush Tewari, Ohad Fried, Justus Thies, Vincent Sitzmann, Stephen Lombardi, Kalyan Sunkavalli, Ricardo Martin-Brualla, Tomas Simon, Jason Saragih, Matthias Nießner, et al. 2020. State of the art on neural rendering. In Computer Graphics Forum, Vol. 39. Wiley Online Library, 701–727

    2020

  31. [39]

    Shiekh Zia Uddin, Sachin Vaidya, Shrish Choudhary, Zhuo Chen, Raafat K Salib, Luke Huang, Dirk R Englund, and Marin Soljačić. 2025. AI-Driven Robotics for Free-Space Optics. arXiv preprint arXiv:2505.17985 (2025)

    2025

  32. [40]

    Peng Wang, Shuai Bai, Sinan Tan, Shijie Wang, Zhihao Fan, Jinze Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, et al. 2024. Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution.arXiv preprint arXiv:2409.12191 (2024)

    2024 arXiv

  33. [41]

    Xinlong Wang, Xiaosong Zhang, Zhengxiong Luo, Quan Sun, Yufeng Cui, Jin- sheng Wang, Fan Zhang, Yueze Wang, Zhen Li, Qiying Yu, et al. 2024. Emu3: Next-token prediction is all you need. arXiv preprint arXiv:2409.18869 (2024)

    2024 arXiv

  34. [42]

    Zhenhua Wang, Guang Xu, and Ming Ren. 2024. LLM-Generated Natural Lan- guage Meets Scaling Laws: New Explorations and Data Augmentation Methods. arXiv preprint arXiv:2407.00322 (2024)

    2024 arXiv

  35. [43]

    Changrong Xiao, Sean Xin Xu, and Kunpeng Zhang. 2023. Multimodal data augmentation for image captioning using diffusion models. In Proceedings of the 1st Workshop on Large Generative Models Meet Multimodal Applications . 23–33

    2023

  36. [44]

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. 2025. Scaling in-the-wild training for diffusion-based illumination harmonization and editing by imposing consistent light transport. In The Thirteenth International Conference on Learning Representations

    2025

  37. [45]

    Long Zhang, Meng Zhang, Wei Lin Wang, and Yu Luo. 2025. Simulation as Reality? The Effectiveness of LLM-Generated Data in Open-ended Question 7 Assessment. arXiv preprint arXiv:2502.06371 (2025)

    2025 arXiv

  38. [46]

    Zicheng Zhang, Junying Wang, Yijin Guo, Farong Wen, Zijian Chen, Hanqing Wang, Wenzhe Li, Lu Sun, Yingjie Zhou, Jianbo Zhang, Bowen Yan, Ziheng Jia, Jiahao Xiao, Yuan Tian, Xiangyang Zhu, Kaiwei Zhang, Chunyi Li, Xiaohong Liu, Xiongkuo Min, Qi Jia, and Guangtao Zhai. 2025. AIB...

    2025

  39. [47]

    Haoyu Zhao, Wenhang Ge, and Ying-cong Chen. 2024. Llm-optic: Unveiling the capabilities of large language models for universal visual grounding. arXiv preprint arXiv:2405.17104 (2024)

    2024 arXiv

  40. [48]

    Xiangyu Zhao, Peiyuan Zhang, Kexian Tang, Hao Li, Zicheng Zhang, Guangtao Zhai, Junchi Yan, Hua Yang, Xue Yang, and Haodong Duan. 2025. Envisioning beyond the pixels: Benchmarking reasoning-informed visual editing. arXiv preprint arXiv:2504.02826 (2025)

    2025 arXiv

  41. [49]

    Longwei Zheng, Fei Jiang, Xiaoqing Gu, Yuanyuan Li, Gong Wang, and Haomin Zhang. 2025. Teaching via LLM-enhanced simulations: Authenticity and barriers to suspension of disbelief. The Internet and Higher Education 65 (2025), 100990. 8

    2025

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.