mlsanityData to 14 Oct 2026

IchthyoNoma: Nomenclature and Context Sensitivity of Zero-Shot Biological Vision--Language Models for Bangladeshi Freshwater Fish Recognition

arXiv 2609.03985 ↗First posted 3 Sep 2026cs.CV cs.CLcited by 0

On the map: Multimodal LLMs / vision-language models → Agricultural vision-language models

Abstract (arXiv)

Zero-shot vision-language models (VLMs) are increasingly used as training-free species recognizers, but reported accuracy can reflect more than visual species knowledge. We audit CLIP, BioCLIP, BioCLIP2, and a multilingual Jina CLIP v2 control on seven freshwater-fish categories from two Bangladeshi sources (10,321 images). BioCLIP2 reaches 72.36% on BFF-15 with English common names and 68.91% on SylFishBD with scientific names, versus 25.15% and 14.40% for generic CLIP. BioCLIP2 Bengali prompts are near chance in balanced accuracy (14.22-14.29%); Jina partially recovers Bengali discrimination to 21.89% and 16.36%, but bare Bengali names return to 14.29% on both sources. Paired SylFishBD interventions show no significant weak-blur effect, modest losses from stronger blur/gray masking, a larger white-mask artifact, and strong species dependence. Zero-shot biological VLM scores therefore jointly reflect biological specialization, multilingual alignment, nomenclature, prompt formulation, and context.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Agricultural vision-language models, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 11 of its 23 references that are in the map. First posted on arXiv 3 Sep 2026.

Builds on

11 of 23 references.

References · 11

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2020Shortcut Learning in Deep Neural Networkscited by 1,969
  3. 2020Noise or Signal: The Role of Image Backgrounds in Object Recognitioncited by 343
  4. 2023BioCLIP: A Vision Foundation Model for the Tree of Lifecited by 161
  5. 2025BioCLIP 2: Emergent Properties from Scaling Hierarchical Contrastive Learningcited by 50
  6. 2024jina-clip-v2: Multilingual Multimodal Embeddings for Text and Imagescited by 37
  7. 2024TaxaBind: A Unified Embedding Space for Ecological Applicationscited by 35
  8. 2022Improving deep neural network generalization and robustness to background bias via layer-wise relevance propagation optimizationcited by 25
  9. 2023Prompting Scientific Names for Zero-Shot Species Recognitioncited by 14
  10. 2023Babel-ImageNet: Massively Multilingual Evaluation of Vision-and-Language Representationscited by 11
  11. 2025uCLIP: Parameter-Efficient Multilingual Extension of Vision-Language Models with Unpaired Datacited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025FishNet++: Analyzing the capabilities of Multimodal Large Language Models in marine biology
  2. 2026Can Edge-Deployable Vision-Language Models Identify Species?
  3. 2025FishAI 2.0: Marine Fish Image Classification with Multi-modal Few-shot Learning
  4. 2023LiT Tuned Models for Efficient Species Detection
  5. 2024Improved Zero-Shot Classification by Adapting VLMs with Text Descriptions
  6. 2026Teach a Molmo2Fish: Towards interactive fish tracking with natural language guidance
  7. 2025Visual Species Recognition with Large Multimodal Models as Post-Hoc Correctors
  8. 2025Are vision-language models ready to zero-shot replace supervised classification models in agriculture?