arXiv cs.AI· S M Rayeed, Qian Ma, Charles V. Stewart, Yao Ma·· 11 小时前AI 评分38
Bio-VQA：视觉语言模型在生物学视觉理解上表现如何？
How Finely Can Vision-Language Models See Biology? Bio-VQA: A Benchmark for Taxonomic, Morphological, Spatial and Individual Reasoning
AI 导读
研究者发布 Bio-VQA，这是目前最大的生物视觉问答基准，含 1.86M 问题、217K 图像，覆盖分类学、形态属性、空间计数与个体重识别四类任务。评测 18 个 3.75B 至 38.4B 参数的开源模型发现，通用视觉能力难以迁移到生物对象：最佳模型在 2 至 10 个个体的图像上精确计数仅 41.7%，分类学准确率从门级 0.72 降至学名级 0.12。
正文 · 原文
Abstract:In biodiversity monitoring, a single image raises several questions: which taxon, how many, which trait, and whether the individual has been seen before. Vision-language models (VLMs) could answer all of these from the image, but their ability is largely untested in biology, as existing benchmarks focus mostly on species identification within a few taxonomic groups. We present Bio-VQA, the largest biological visual question answering (VQA) benchmark and the only one to span four aspects of biological visual understanding: hierarchical taxonomic reasoning, fine-grained attribute grounding, quantitative spatial reasoning, and individual re-identification. Bio-VQA contains 1.86M questions over 217K images from 11 public datasets, and extends right-or-wrong scoring to graded scoring, so a near-miss counts for more than a distant one. Evaluating 18 open-weight models from 3.75B to 38.4B parameters, we find that general visual capabilities do not transfer to biological organisms: the best model counts animals exactly on 41.7% of images with 2 to 10 individuals, less than half its accuracy on everyday objects at the same counts. Taxonomic accuracy falls to one sixth, from 0.72 at phylum to 0.12 at scientific name, and scaling helps mostly at coarse ranks. The gap to a domain specialist widens accordingly, from 0.731 against BioCLIP-2's 0.874 at class (1.2x) to 0.269 against 0.544 at scientific name (2x). On trait size measurement, 15 of the 18 models do no better than an image-blind baseline that predicts one fixed number for every specimen, and on re-identification, 17 of the 18 models fall below chance under adversarial conditions. These failures persist across a tenfold range of model size. We hope Bio-VQA becomes a standard test of how well VLM capabilities transfer to the biological domain.
| Comments: | Accepted to WACV 2027, 8 pages main paper; 4 figures, 4 tables |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.05559 [cs.CV] |
| (or arXiv:2611.05559v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2611.05559 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: S M Rayeed [view email]
[v1]
Tue, 3 Nov 2026 19:51:00 UTC (10,868 KB)
来源：arXiv cs.AI · arxiv.org