Tighnari v2: Mitigating Label Noise and Distribution Shift in Multimodal Plant Distribution Prediction via Mixture of Experts and Weakly Supervised Learning

TL;DR

This paper introduces a multimodal fusion framework with mixture of experts and weakly supervised learning to improve plant distribution prediction by mitigating label noise and distribution shift between presence-absence and presence-only data.

Contribution

It proposes a novel pseudo-label aggregation strategy, a multimodal model architecture, and a mixture-of-experts approach to handle geographic distribution shifts and label noise in plant distribution modeling.

Findings

Achieves superior performance on GeoLifeCLEF 2025 dataset

Effectively mitigates label noise in PO data

Handles distribution shifts between training and testing data

Abstract

Large-scale, cross-species plant distribution prediction plays a crucial role in biodiversity conservation, yet modeling efforts in this area still face significant challenges due to the sparsity and bias of observational data. Presence-Absence (PA) data provide accurate and noise-free labels, but are costly to obtain and limited in quantity; Presence-Only (PO) data, by contrast, offer broad spatial coverage and rich spatiotemporal distribution, but suffer from severe label noise in negative samples. To address these real-world constraints, this paper proposes a multimodal fusion framework that fully leverages the strengths of both PA and PO data. We introduce an innovative pseudo-label aggregation strategy for PO data based on the geographic coverage of satellite imagery, enabling geographic alignment between the label space and remote sensing feature space. In terms of model architecture, we adopt Swin Transformer Base as the backbone for satellite imagery, utilize the TabM network for tabular feature extraction, retain the Temporal Swin Transformer for time-series modeling, and employ a stackable serial tri-modal cross-attention mechanism to optimize the fusion of heterogeneous modalities. Furthermore, empirical analysis reveals significant geographic distribution shifts between PA training and test samples, and models trained by directly mixing PO and PA data tend to experience performance degradation due to label noise in PO data. To address this, we draw on the mixture-of-experts paradigm: test samples are partitioned according to their spatial proximity to PA samples, and different models trained on distinct datasets are used for inference and post-processing within each partition. Experiments on the GeoLifeCLEF 2025 dataset demonstrate that our approach achieves superior predictive performance in scenarios with limited PA coverage and pronounced distribution shifts.