AgriChain dataset with expert-verified CoT rationales enables fine-tuned AgriChain-VL3B to reach 73.1% top-1 accuracy on plant disease diagnosis while producing explanations aligned with agricultural experts.
Collectively, these stud- ies demonstrate the growing potential of CoT for multimodal reasoning tasks, including fine-grained agricultural diagnostics
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
AgriChain Visually Grounded Expert Verified Reasoning for Interpretable Agricultural Vision Language Models
AgriChain dataset with expert-verified CoT rationales enables fine-tuned AgriChain-VL3B to reach 73.1% top-1 accuracy on plant disease diagnosis while producing explanations aligned with agricultural experts.