Search for: All records

Abstract 1. Species distribution models (SDMs) are crucial tools for understanding and predicting biodiversity patterns, yet they often struggle with limited data, biased sampling, and complex species-environment relationships. Here I present NicheFlow, a novel foundation model for SDMs that leverages generative AI to address these challenges and advance our ability to model and predict species distributions across taxa and environments. 2. NicheFlow employs a two-stage generative approach, combining species embeddings with two chained generative models, one to generate a distribution in environmental space, and a second to generate a distribution in geographic space. This architecture allows for the sharing of information across species and captures complex, non-linear relationships in environmental space. I trained NicheFlow on a comprehensive dataset of reptile distributions and evaluated its performance using both standard SDM metrics and zero-shot prediction tasks. 3. NicheFlow demonstrates good predictive performance, particularly for rare and data-deficient species. The model successfully generated plausible distributions for species not seen during training, showcasing its potential for zero-shot prediction. The learned species embeddings captured meaningful ecological information, revealing patterns in niche structure across taxa, latitude and range sizes. 4. As a proof-of-principle foundation model, NicheFlow represents a significant advance in species distribution modeling, offering a powerful tool for addressing pressing questions in ecology, evolution, and conservation biology. Its ability to model joint species distributions and generate hypothetical niches opens new avenues for exploring ecological and evolutionary questions, including ancestral niche reconstruction and community assembly processes. This approach has the potential to transform our understanding of biodiversity patterns and improve our capacity to predict and manage species distributions in the face of global change. 

Data on the three dimensional shape of organismal morphology is becoming increasingly available, and forms part of a new revolution in high-throughput phenomics that promises to help understand ecological and evolutionary processes that influence phenotypes at unprecedented scales. However, in order to meet the potential of this revolution we need new data analysis tools to deal with the complexity and heterogeneity of large-scale phenotypic data such as 3D shapes. In this study we explore the potential of generative Artificial Intelligence to help organize and extract meaning from complex 3D data. Specifically, we train a deep representational learning method known as DeepSDF on a dataset of 3D scans of the bills of 2,020 bird species. The model is designed to learn a continuous vector representation of 3D shapes, along with a ’decoder’ function, that allows the transformation from this vector space to the original 3D morphological space. We find that approach successfully learns coherent representations: particular directions in latent space are associated with discernible morphological meaning (such as elongation, flattening, etc.). More importantly, learned latent vectors have ecological meaning as shown by their ability to predict the trophic niche of the bird each bill belongs to with a high degree of accuracy. Unlike existing 3D morphometric techniques, this method has very little requirements for human supervised tasks such as landmark placement, increasing it accessibility to labs with fewer labour resources. It has fewer strong assumptions than alternative dimension reduction techniques such as PCA. Once trained, 3D morphology predictions can be made from latent vectors very computationally cheaply. The trained model has been made publicly available and can be used by the community, including for finetuning on new data, representing an early step toward developing shared, reusable AI models for analyzing organismal morphology.