V. Miele

Zero-shot animal behavior classification with vision-language foundation models

This study uses vision-language foundation models to classify animal behaviors in camera trap images without prior training, achieving over 91% accuracy in three ungulate species.

Being confident in confidence scores : calibration in deep learning models for camera trap image sequences

This study evaluates the calibration of deep learning models for camera trap image sequences, showing that temperature scaling improves confidence score interpretability for ecological applications.

Deep learning for species identification of modern and fossil rodent molars

This paper develops a deep learning method for identifying rodent species from molar teeth images, achieving performance comparable to geometric morphometrics and demonstrating applicability to fossil and archaeological specimens.

Core-periphery structure in mutualistic networks: an epitaph for nestedness?

The paper reviews 419 studies on mutualistic networks and argues that core-periphery structure is more relevant than nestedness for describing ecological network topology, providing code for its calculation in R.

Revisiting giraffe photo-identification using deep learning and network analysis

The paper develops a deep learning pipeline for giraffe re-identification using image similarity networks and CNNs, achieving 90% Top-1 accuracy in identifying known individuals from 4,000 photographs.

Core--periphery dynamics in a plant--pollinator network

The paper analyzes a plant-pollinator network over six years using dynamic stochastic block models, finding persistent core-periphery structure with dynamic species roles.

The diversity of interaction types drives the functioning of ecological communities

The paper studies how diverse interaction types in ecological communities affect species diversity and functioning, finding that they increase the slope of the diversity-functioning relationship.