EarthPT: a time series foundation model for Earth Observation

• URL: http://arxiv.org/abs/2309.07207v2
• Date: Thu, 11 Jan 2024 14:36:57 GMT
• Title: EarthPT: a time series foundation model for Earth Observation
• Authors: Michael J. Smith, Luke Fleming and James E. Geach
• Abstract summary: We introduce EarthPT -- an Earth Observation (EO) pretrained transformer. We demonstrate that EarthPT is an effective forecaster that can accurately predict future pixel-level surface reflectances. We also demonstrate that embeddings learnt by EarthPT hold semantically meaningful information.
• Score: 0.0
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: We introduce EarthPT -- an Earth Observation (EO) pretrained transformer. EarthPT is a 700 million parameter decoding transformer foundation model trained in an autoregressive self-supervised manner and developed specifically with EO use-cases in mind. We demonstrate that EarthPT is an effective forecaster that can accurately predict future pixel-level surface reflectances across the 400-2300 nm range well into the future. For example, forecasts of the evolution of the Normalised Difference Vegetation Index (NDVI) have a typical error of approximately 0.05 (over a natural range of -1 -> 1) at the pixel level over a five month test set horizon, out-performing simple phase-folded models based on historical averaging. We also demonstrate that embeddings learnt by EarthPT hold semantically meaningful information and could be exploited for downstream tasks such as highly granular, dynamic land use classification. Excitingly, we note that the abundance of EO data provides us with -- in theory -- quadrillions of training tokens. Therefore, if we assume that EarthPT follows neural scaling laws akin to those derived for Large Language Models (LLMs), there is currently no data-imposed limit to scaling EarthPT and other similar `Large Observation Models.'

Related papers

• SpectralEarth: Training Hyperspectral Foundation Models at Scale [47.93167977587301]
  We introduce SpectralEarth, a large-scale multi-temporal dataset designed to pretrain hyperspectral foundation models. We pretrain a series of foundation models on SpectralEarth using state-of-the-art self-supervised learning (SSL) algorithms. We construct four downstream datasets for land-cover and crop-type mapping, providing benchmarks for model evaluation.
  arXiv  Detail & Related papers  (2024-08-15T22:55:59Z)
• Aurora: A Foundation Model of the Atmosphere [56.97266186291677]
  We introduce Aurora, a large-scale foundation model of the atmosphere trained on over a million hours of diverse weather and climate data. In under a minute, Aurora produces 5-day global air pollution predictions and 10-day high-resolution weather forecasts.
  arXiv  Detail & Related papers  (2024-05-20T14:45:18Z)
• ORBIT: Oak Ridge Base Foundation Model for Earth System Predictability [10.88886669820126]
  We introduce the Oak Ridge Base Foundation Model for Earth System Predictability (ORBIT) ORBIT is the largest model of its kind and surpasses the current climate AI foundation model size by a thousandfold. Performance scaling tests on the Frontier supercomputer have demonstrated that ORBIT achieves 684 petaFLOPS to 1.6 exaFLOPS sustained throughput.
  arXiv  Detail & Related papers  (2024-04-23T03:39:57Z)
• A Geospatial Approach to Predicting Desert Locust Breeding Grounds in Africa [3.6826233660285395]
  locust swarms present a major threat to agriculture and food security. Our study develops an operationally-ready model for predicting locust breeding grounds.
  arXiv  Detail & Related papers  (2024-03-11T16:13:58Z)
• Observation-Guided Meteorological Field Downscaling at Station Scale: A Benchmark and a New Method [66.80344502790231]
  We extend meteorological downscaling to arbitrary scattered station scales and establish a new benchmark and dataset. Inspired by data assimilation techniques, we integrate observational data into the downscaling process, providing multi-scale observational priors. Our proposed method outperforms other specially designed baseline models on multiple surface variables.
  arXiv  Detail & Related papers  (2024-01-22T14:02:56Z)
• Foundation Models for Generalist Geospatial Artificial Intelligence [3.7002058945990415]
  This paper introduces a first-of-a-kind framework for the efficient pre-training and fine-tuning of foundational models on extensive data. We have utilized this framework to create Prithvi, a transformer-based foundational model pre-trained on more than 1TB of multispectral satellite imagery.
  arXiv  Detail & Related papers  (2023-10-28T10:19:55Z)
• A Comparative Study on Generative Models for High Resolution Solar Observation Imaging [59.372588316558826]
  This work investigates capabilities of current state-of-the-art generative models to accurately capture the data distribution behind observed solar activity states. Using distributed training on supercomputers, we are able to train generative models for up to 1024x1024 resolution that produce high quality samples indistinguishable to human experts.
  arXiv  Detail & Related papers  (2023-04-14T14:40:32Z)
• Predictive World Models from Real-World Partial Observations [66.80340484148931]
  We present a framework for learning a probabilistic predictive world model for real-world road environments. While prior methods require complete states as ground truth for learning, we present a novel sequential training method to allow HVAEs to learn to predict complete states from partially observed states only.
  arXiv  Detail & Related papers  (2023-01-12T02:07:26Z)
• Earthformer: Exploring Space-Time Transformers for Earth System Forecasting [27.60569643222878]
  We propose Earthformer, a space-time Transformer for Earth system forecasting. The Transformer is based on a generic, flexible and efficient space-time attention block, named Cuboid Attention. Experiments on two real-world benchmarks about precipitation nowcasting and El Nino/Southerntemporaltion show Earthformer achieves state-of-the-art performance.
  arXiv  Detail & Related papers  (2022-07-12T20:52:26Z)
• Embedding Earth: Self-supervised contrastive pre-training for dense land cover classification [61.44538721707377]
  We present Embedding Earth a self-supervised contrastive pre-training method for leveraging the large availability of satellite imagery. We observe significant improvements up to 25% absolute mIoU when pre-trained with our proposed method. We find that learnt features can generalize between disparate regions opening up the possibility of using the proposed pre-training scheme.
  arXiv  Detail & Related papers  (2022-03-11T16:14:14Z)
• EarthNet2021: A novel large-scale dataset and challenge for forecasting localized climate impacts [12.795776149170978]
  Large Earth observation datasets now enable us to create machine learning models capable of translating coarse weather information into high-resolution Earth surface forecasts. We define high-resolution Earth surface forecasting as video prediction of satellite imagery conditional on mesoscale weather forecasts. We introduce EarthNet 2021, a new curated dataset containing target-temporal Sentinel 2 satellite imagery at 20 m resolution, matched with high-resolution topography and mesoscale (1.28 km) weather variables.
  arXiv  Detail & Related papers  (2020-12-11T11:21:00Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.