Morphology-Aware Multimodal Representation Learning for Insect Phylogenetic Reconstruction

2026-06-20

AI Summary

This work proposes a morphology-aware multimodal alignment framework that explicitly integrates insect specimen images with expert-authored morphological descriptions during representation learning—a capability absent in traditional image-based phylogenetic reconstruction methods that rely solely on visual data. By employing parameter-efficient fine-tuning of a Vision Transformer coupled with supervised contrastive learning, the framework aligns image and text modalities within a shared latent space to generate image embeddings enriched with morphological semantics. These embeddings are subsequently used as continuous traits in a Bayesian phylogenetic model. Experiments on the Rove-Tree-11 dataset demonstrate that this approach significantly improves topological congruence between inferred and reference phylogenies, underscoring the critical role of multimodal alignment in extracting biologically informative morphological characters for systematics.

Abstract

Morphological traits provide important evidence for phylogenetic reconstruction and evolutionary relationship analysis. Recent image-based approaches have introduced deep learning, particularly convolutional models, to derive morphological features from specimen images, but these methods generally rely on single-modality visual representations and do not explicitly incorporate morphological semantics. This study proposes a morphology-aware multimodal alignment framework for insect phylogenetic reconstruction. The framework combines specimen images with curated morphological descriptions by adapting a vision transformer through parameter-efficient fine-tuning and supervised contrastive learning, followed by image-text alignment in a shared latent space. The learned image embeddings are then used as continuous traits for Bayesian phylogenetic reconstruction. On the public Rove-Tree-11 dataset, comparative and ablation experiments across multiple visual backbones and feature adaptation strategies demonstrate that multimodal alignment improves topological agreement with the reference phylogeny. The results indicate that the proposed framework can derive morphology-aware visual traits for computational phylogenetic reconstruction.

Problem

• phylogenetic reconstruction
• morphological traits
• multimodal representation
• insect evolution
• image-text alignment

Innovation

• multimodal alignment
• morphology-aware representation
• vision transformer
• supervised contrastive learning
• phylogenetic reconstruction