Mini-JEPA Foundation Model Fleet Enables Agentic Hydrologic Intelligence

• URL: http://arxiv.org/abs/2605.14120v1
• Date: Wed, 13 May 2026 21:11:14 GMT
• Title: Mini-JEPA Foundation Model Fleet Enables Agentic Hydrologic Intelligence
• Abstract summary: A single planetary-scale model, e.g. Google AlphaEarth, handles broad characterization well but may compromise on specialized hydrologic signals.<n>We propose Mini-JEPAs: a fleet of small sensor-specialized Joint Embedding Predictive Architecture (JEPA) foundation models consulted by a routing agent for specialized questions.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Geospatial foundation models compress multispectral observations into dense embeddings increasingly used in natural-language environmental reasoning systems. A single planetary-scale model, e.g. Google AlphaEarth, handles broad characterization well but may compromise on specialized hydrologic signals. Such generalist models are also often inaccessible, expensive, and require large-scale compute. We propose Mini-JEPAs: a fleet of small sensor-specialized Joint Embedding Predictive Architecture (JEPA) foundation models consulted by a routing agent for specialized questions. We pretrained five 22M-parameter Mini-JEPAs sharing an identical Vision Transformer backbone, JEPA recipe, and 64-d output space, using Sentinel-2 optical, Sentinel-1 SAR, MODIS thermal, multi-temporal Sentinel-2 phenology, and a topography-soil stack. Each Mini-JEPA reconstructs the variable matched to its sensor, with cross-validated $R^2$ reaching 0.97 for elevation, 0.97 for temperature, and 0.81 for precipitation. The five manifolds differ in geometric structure, with global participation ratios from 8.9 to 20.2 and local intrinsic dimensionalities from 2.3 to 9.0. Joint topography-soil and phenology models add predictive value beyond AlphaEarth alone for soil moisture, aridity, and precipitation ($ΔR^2$ up to 0.031). A router LLM reads per-modality references and selects appropriate sensors with a perfect hit rate over a curated question set. In paired LLM-as-Judge evaluation, dual retrieval over AlphaEarth and the routed fleet outperforms AlphaEarth alone on physics-matched questions (Cohen's $d = 1.10$, $p = 0.031$). Locally-trained Mini-JEPAs can be operationalized for hydrologic intelligence with modest compute.

Related papers

• FLORO: A Multimodal Geospatial Foundation Model for Ecological Remote Sensing Across Sensors and Scales [65.4821703903285]
  FLORO is a multimodal foundation model designed to learn transferable representations from a small but highly diverse remote sensing corpus.<n> FLORO is pretrained using masked autoencoding on a heterogeneous combination of Sentinel-1, Sentinel-2, SkySAT imagery, elevation, and UAV-derived data.<n>We evaluated FLORO on the PANGAEA benchmark under a frozen-encoder protocol across scene classification, segmentation, and regression tasks.
  arXiv  Detail & Related papers  (2026-05-27T08:55:54Z)
• HydroAgent: Closing the Gap Between Frontier LLMs and Human Experts in Hydrologic Model Calibration via Simulator-Grounded RL [6.680830184790049]
  Calibrating distributed hydrologic models is a critical bottleneck across operational water resources management.<n>We ask: can frontier large language model (LLM) agents replace the human hydrologic modeler, and if not, what would it take?<n>We benchmark nine frontier LLM agents on the operational CREST distributed hydrologic model used by the U.S. National Weather Service for flash-flood forecasting.
  arXiv  Detail & Related papers  (2026-05-18T03:17:38Z)
• LEPA: Learning Geometric Equivariance in Satellite Remote Sensing Data with a Predictive Architecture [2.618539394367099]
  We propose a new architecture for predicting embeddings for large-scale satellite remote sensing data.<n>Instead of averaging, LEPA conditions a latent on embeddings to directly predict on NASA/USGS Landsat-Sentinel (HLS) imagery and LNet Image 0.8k.
  arXiv  Detail & Related papers  (2026-03-07T14:56:15Z)
• Physically Interpretable AlphaEarth Foundation Model Embeddings Enable LLM-Based Land Surface Intelligence [0.0]
  We present a comprehensive interpretability analysis of Google AlphaEarth's 64-dimensional embeddings against 26 environmental variables.<n>We then developed a Land Surface Intelligence system that implements retrieval-augmented generation over a FAISS-indexed embedding database of 12.1 million vectors.
  arXiv  Detail & Related papers  (2026-02-10T22:58:50Z)
• LiM-YOLO: Less is More with Pyramid Level Shift and Normalized Auxiliary Branch for Ship Detection in Optical Remote Sensing Imagery [3.2425852744760184]
  LiM-YOLO is a specialized detector designed to resolve domain-specific conflicts.<n>Based on a statistical analysis of ship scales, we introduce a Pyramid Level Shift Strategy that mitigates the detection head to P2-P4.<n>LiM-YOLO demonstrates superior detection accuracy and efficiency compared to state-of-the-art models.
  arXiv  Detail & Related papers  (2025-12-10T14:48:58Z)
• LeJEPA: Provable and Scalable Self-Supervised Learning Without the Heuristics [53.247652209132376]
  Joint-Embedding Predictive Architectures (JEPAs) offer a promising blueprint, but lack of practical guidance and theory has led to ad-hoc R&D.<n>We present a comprehensive theory of JEPAs and instantiate it in bf LeJEPA, a lean, scalable, and theoretically grounded training objective.
  arXiv  Detail & Related papers  (2025-11-11T18:21:55Z)
• Gaussian Embeddings: How JEPAs Secretly Learn Your Data Density [51.15085346971361]
  Joint Embedding Predictive Architectures (JEPAs) learn representations able to solve numerous downstream tasks out-of-the-box.<n>JEPAs combine two objectives: (i) a latent-space prediction term, i.e., the representation of a slightly perturbed sample must be predictable from the original sample's representation, and (ii) an anti-collapse term, i.e., not all samples should have the same representation.
  arXiv  Detail & Related papers  (2025-10-07T14:06:30Z)
• Mercury: Ultra-Fast Language Models Based on Diffusion [58.52391675075641]
  We present Mercury, a new generation of commercial-scale large language models (LLMs) based on diffusion.<n>Mercury Coder comes in two sizes: Mini and Small.<n>Based on independent evaluations, Mercury Coder Mini and Mercury Coder Small achieve state-of-the-art throughputs of 1109 tokens/sec and 737 tokens/sec, respectively.
  arXiv  Detail & Related papers  (2025-06-17T17:06:18Z)
• Deep Learning for Real Time Satellite Pose Estimation on Low Power Edge TPU [58.720142291102135]
  In this paper we propose a pose estimation software exploiting neural network architectures. We show how low power machine learning accelerators could enable Artificial Intelligence exploitation in space.
  arXiv  Detail & Related papers  (2022-04-07T08:53:18Z)
• Searching Collaborative Agents for Multi-plane Localization in 3D Ultrasound [59.97366727654676]
  3D ultrasound (US) is widely used due to its rich diagnostic information, portability and low cost. Standard plane (SP) localization in US volume not only improves efficiency and reduces user-dependence, but also boosts 3D US interpretation. We propose a novel Multi-Agent Reinforcement Learning framework to localize multiple uterine SPs in 3D US simultaneously.
  arXiv  Detail & Related papers  (2020-07-30T07:23:55Z)

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.