mlsanityData to 14 Oct 2026

VaseVQA: Multimodal Agent and Benchmark for Ancient Greek Pottery

arXiv 2509.17191 ↗First posted 21 Sep 2025cs.CV cs.CLcited by 6

On the map: LLM agents and tool use → Multimodal and embodied agents

Abstract (arXiv)

Understanding cultural heritage artifacts such as ancient Greek pottery requires expert-level reasoning that remains challenging for current MLLMs due to limited domain-specific data. We introduce VaseVQA, a benchmark of 31,773 images and 67,614 question-answer pairs across seven expert-defined categories, enabling systematic evaluation of expert-level cultural heritage understanding. Using this dataset, we explore effective training strategies for domain-specific reasoning. While supervised fine-tuning improves adaptation to domain knowledge, it struggles with deeper reasoning tasks. We propose VaseVL, which augments SFT with reinforcement learning using verifiable rewards. Experiments show that VaseVL consistently outperforms supervised baselines, especially on reasoning-intensive questions, highlighting the value of targeted reinforcement learning for cultural heritage visual question answering. Our code and dataset will be released at https://github.com/AIGeeksGroup/VaseVQA.

Key facts. Cited by 6 papers in the map (data through 14 Oct 2026). Placed in Multimodal and embodied agents, one of 16 directions in LLM agents and tool use. Builds on 30 of its 91 references that are in the map. First posted on arXiv 21 Sep 2025.

Builds on

30 of 91 references.

References · 30

  1. 2023Visual Instruction Tuningcited by 7,869
  2. 2023BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Modelscited by 6,137
  3. 2024Gemini 1.5: Unlocking multimodal understanding across millions of tokens of contextcited by 3,224
  4. 2023MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Modelscited by 2,711
  5. 2013Fine-Grained Visual Classification of Aircraftcited by 1,726
  6. 2015Microsoft COCO Captions: Data Collection and Evaluation Servercited by 1,462
  7. 2019VisualBERT: A Simple and Performant Baseline for Vision and Languagecited by 1,296
  8. 2024MiniCPM-V: A GPT-4V Level MLLM on Your Phonecited by 1,079
  9. 2024How Far Are We to GPT-4V? Closing the Gap to Commercial Multimodal Models with Open-Source Suitescited by 1,050
  10. 2025Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Modelscited by 665
  11. 2020PathVQA: 30000+ Questions for Medical Visual Question Answeringcited by 475
  12. 2015Neural Module Networkscited by 376
18 more
  1. 2019A Short Note on the Kinetics-700 Human Action Datasetcited by 310
  2. 2021Multi-Grained Vision Language Pre-Training: Aligning Texts with Visual Conceptscited by 256
  3. 2022ELEVATER: A Benchmark and Toolkit for Evaluating Language-Augmented Visual Modelscited by 141
  4. 2015Exploring Models and Data for Image Question Answeringcited by 139
  5. 2015Explicit Knowledge-based Reasoning for Visual Question Answeringcited by 85
  6. 2015Are You Talking to a Machine? Dataset and Methods for Multilingual Image Question Answeringcited by 80
  7. 20253D-R1: Enhancing Reasoning in 3D VLMs for Unified Scene Understandingcited by 47
  8. 2025Nav-R1: Reasoning and Navigation in Embodied Scenescited by 38
  9. 2025Think or Not? Selective Reasoning via Reinforcement Learning for Vision-Language Modelscited by 30
  10. 2025ManipLVM-R1: Reinforcement Learning for Reasoning in Embodied Manipulation with Large Vision-Language Modelscited by 27
  11. 2017C-VQA: A Compositional Split of the Visual Question Answering (VQA) v1.0 Datasetcited by 24
  12. 2025Geolocation with Real Human Gameplay Data: A Large-Scale Dataset and Human-Like Reasoning Frameworkcited by 24
  13. 2020RareAct: A video dataset of unusual interactionscited by 23
  14. 2025Audio Jailbreak: An Open Comprehensive Benchmark for Jailbreaking Large Audio-Language Modelscited by 21
  15. 2019NLVR2 Visual Bias Analysiscited by 19
  16. 2024An open dataset for oracle bone script recognition and deciphermentcited by 19
  17. 2024Re-assembling the past: The RePAIR dataset and benchmark for real world 2D and 3D puzzle solvingcited by 19
  18. 2015Visual Madlibs: Fill in the blank Image Generation and Question Answeringcited by 14

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 4.

6 papers.

  1. 2025VLA-R1: Enhancing Reasoning in Vision-Language-Action Modelscited by 38
  2. 2025VaseVQA-3D: Benchmarking 3D VLMs on Ancient Greek Potterycited by 4
  3. 2026MechVQA: Benchmarking and Enhancing Multimodal LLMs on Comprehensive Mechanical Drawing Understandingcited by 4
  4. 2026WebCryptoAgent: Agentic Crypto Trading with Web Informaticscited by 0
  5. 2026CiQi-Agent: Aligning Vision, Tools and Aesthetics in Multimodal Agent for Cultural Reasoning on Chinese Porcelainscited by 0
  6. 2026VaseMuseum: Digital Intelligent Museum for Ancient Greek Potterycited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025VQArt-Bench: A semantically rich VQA Benchmark for Art and Cultural Heritage
  2. 2026EduArt: An educational-level benchmark for evaluating art history knowledge in large language models
  3. 2025On Path to Multimodal Historical Reasoning: HistBench and HistAgent
  4. 2024Understanding Museum Exhibits using Vision-Language Reasoning
  5. 2025Seeing Culture: A Benchmark for Visual Reasoning and Grounding
  6. 2026A-MAR: Agent-based Multimodal Art Retrieval for Fine-Grained Artwork Understanding
  7. 2026On the Cultural Anachronism and Temporal Reasoning in Vision Language Models
  8. 2024Increasing the Difficulty of Automatically Generated Questions via Reinforcement Learning with Synthetic Preference