LITE: Modeling Environmental Ecosystems with Multimodal Large Language Models

TL;DR

LITE is a multimodal large language model designed for environmental ecosystems modeling, effectively handling incomplete data and distribution shifts to improve spatial-temporal predictions.

Contribution

It introduces a unified multimodal framework that transforms environmental data into natural language and images, using specialized encoders and a sparse Mixture-of-Experts to address data issues.

Findings

Achieves 41.25% reduction in prediction error

Outperforms existing baselines in spatial-temporal prediction

Effectively handles incomplete features and distribution shifts

Abstract

The modeling of environmental ecosystems plays a pivotal role in the sustainable management of our planet. Accurate prediction of key environmental variables over space and time can aid in informed policy and decision-making, thus improving people's livelihood. Recently, deep learning-based methods have shown promise in modeling the spatial-temporal relationships for predicting environmental variables. However, these approaches often fall short in handling incomplete features and distribution shifts, which are commonly observed in environmental data due to the substantial cost of data collection and malfunctions in measuring instruments. To address these issues, we propose LITE -- a multimodal large language model for environmental ecosystems modeling. Specifically, LITE unifies different environmental variables by transforming them into natural language descriptions and line graph images. Then, LITE utilizes unified encoders to capture spatial-temporal dynamics and correlations in different modalities. During this step, the incomplete features are imputed by a sparse Mixture-of-Experts framework, and the distribution shift is handled by incorporating multi-granularity information from past observations. Finally, guided by domain instructions, a language model is employed to fuse the multimodal representations for the prediction. Our experiments demonstrate that LITE significantly enhances performance in environmental spatial-temporal prediction across different domains compared to the best baseline, with a 41.25% reduction in prediction error. This justifies its effectiveness. Our data and code are available at https://github.com/hrlics/LITE.