Search NASA⌕ Search

SEARCH · Search NASA

Results for “Foundation models”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

OReole-FM: successes and challenges toward billion-parameter foundation models for high-resolution satellite imagery

While the pretraining of Foundation Models (FMs) for remote sensing (RS) imagery is on the rise, models remain restricted to a few hundred million parameters. Scaling models to billions of parameters has been shown to yield unprecedented benefits including emergent abilities, but requires data scaling and computing resources typically not available outside industry R&D labs. In this work, we pair high-performance computing resources including Frontier supercomputer, America's first exascale system, and high-resolution optical RS data to pretrain billion-scale FMs. Our study assesses performance of different pretrained variants of vision Transformers across image classification, semantic segmentation and object detection benchmarks, which highlight the importance of data scaling for effective model scaling. Moreover, we discuss construction of a novel TIU pretraining dataset, model initialization, with data and pretrained models intended for public release. By discussing technical challenges and details often lacking in the related literature, this work is intended to offer best practices to the geospatial community toward efficient training and benchmarking of larger FMs.

Ambrozio Dias, Philipe↗
DOE OSTI2025

Yes, No, Maybe So: Human Factors Considerations for Fostering Calibrated Trust in Foundation Models Under Uncertainty

High-stakes analytical environments require analysts to evaluate evidence and generate conclusions to inform critical decisions often under conditions of uncertainty. Probabilistic decision-making based on incomplete or inaccurate information can reduce productivity, compromise national interests, and endanger public safety. Researchers are developing expert systems built on foundation models (FMs) to support analysts’ decision-making processes by enabling human-artificial intelligence (AI) teaming, in part through the quantification and expression of uncertainty information. As FMs continue to mature, it is imperative to correspondingly consider analysts’ needs for appropriately interpreting and using uncertainty information. However, prior research indicates that it remains unclear how analysts engage with FM-generated uncertainty information and the extent to which these interactions influence trust in, and reliance on, expert systems. We plan to review the state of the science and conduct an exploratory, qualitative study to (a) understand how properly communicated uncertainty can foster calibrated trust and appropriate reliance and (b) identify approaches for effectively conveying FM-generated uncertainty information during analytical workflows. We will administer semi-structured interviews with analysts from a specific high-stakes analytical environment to collect their current experiences with job-related uncertainty and their impressions when viewing FM-generated uncertainty information. During the interview protocol, participants will be presented with several different FM outputs and invited to discuss their thoughts and beliefs about the uncertainty information displayed. Participants may provide insights into how trust and reliance may be influenced by uncertainty. The results of this study will help us to better understand how analysts currently interpret and use uncertainty information. Our findings may inform human factors recommendations for effectively conveying uncertainty information to foster calibrated trust in, and appropriate reliance on, expert systems. Interaction designers and FM developers can use this knowledge to enhance human-AI teaming and ensure the responsible deployment of FM-based expert systems in analytical workflows.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

FM4NPP: A Scaling Foundation Model for Nuclear and Particle Physics

Large language models have revolutionized artificial intelligence by enabling large, generalizable models trained through self-supervision. This paradigm has inspired the development of scientific foundation models (FMs). However, applying this capability to experimental particle physics is challenging due to the sparse, spatially distributed nature of detector data, which differs dramatically from natural language. This work addresses if an FM for particle physics can scale and generalize across diverse tasks. We introduce a new dataset with more than 11 million particle collision events and a suite of downstream tasks and labeled data for evaluation. We propose a novel self-supervised training method for detector data and demonstrate its neural scalability with models that feature up to 188 million parameters. With frozen weights and task-specific adapters, this FM consistently outperforms baseline models across all downstream tasks. The performance also exhibits robust data-efficient adaptation. Further analysis reveals that the representations extracted by the FM are task-agnostic but can be specialized via a single linear mapping for different downstream tasks.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

Pretraining Billion-Scale Geospatial Foundational Models on Frontier

As AI workloads increase in scope, generalization capability becomes challenging for small task-specific models and their demand for large amounts of labeled training samples increases. On the contrary, Foundation Models (FMs) are trained with internet-scale unlabeled data via self-supervised learning and have been shown to adapt to various tasks with minimal fine-tuning. Although large FMs have demonstrated significant impact in natural language processing and computer vision, efforts toward FMs for geospatial applications have been restricted to smaller size models, as pretraining larger models requires very large computing resources equipped with state-of-the-art hardware accelerators. Current satellite constellations collect 100+TBs of data a day, resulting in images that are billions of pixels and multimodal in nature. Such geospatial data poses unique challenges opening up new opportunities to develop FMs. We investigate billion scale FMs and HPC training profiles for geospatial applications by pretraining on publicly available data. We studied from end-to-end the performance and impact in the solution by scaling the model size. Our larger 3B parameter size model achieves up to 30% improvement in top1 scene classification accuracy when comparing a 100M parameter model. Moreover, we detail performance experiments on the Frontier supercomputer, America's first exascale system, where we study different model and data parallel approaches using PyTorch's Fully Sharded Data Parallel library. Specifically, we study variants of the Vision Transformer architecture (ViT), conducting performance analysis for ViT models with size up to 15B parameters. By discussing throughput and performance bottlenecks under different parallelism configurations, we offer insights on how to leverage such leadership-class HPC resources when developing large models for geospatial imagery applications.

Tsaris, Aristeidis (aris)↗
DOE OSTI2024

AI Foundation Models for Science [Slides]

This presentation deals with three main points: why answer science questions with AI, trusting AI for science, and foundation models for science.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Learning together: Towards foundation models for machine learning interatomic potentials with meta-learning

Abstract The development of machine learning models has led to an abundance of datasets containing quantum mechanical (QM) calculations for molecular and material systems. However, traditional training methods for machine learning models are unable to leverage the plethora of data available as they require that each dataset be generated using the same QM method. Taking machine learning interatomic potentials (MLIPs) as an example, we show that meta-learning techniques, a recent advancement from the machine learning community, can be used to fit multiple levels of QM theory in the same training process. Meta-learning changes the training procedure to learn a representation that can be easily re-trained to new tasks with small amounts of data. We then demonstrate that meta-learning enables simultaneously training to multiple large organic molecule datasets. As a proof of concept, we examine the performance of a MLIP refit to a small drug-like molecule and show that pre-training potentials to multiple levels of theory with meta-learning improves performance. This difference in performance can be seen both in the reduced error and in the improved smoothness of the potential energy surface produced. We therefore show that meta-learning can utilize existing datasets with inconsistent QM levels of theory to produce models that are better at specializing to new datasets. This opens new routes for creating pre-trained, foundation models for interatomic potentials.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2023

An Agenda for Multimodal Foundation Models for Earth Observation

Archives of remote sensing (RS) data are increasing swiftly as new sensing modalities with enhanced spatiotemporal resolution become operational. While promising new breakthroughs, the sheer volume of RS archives stretches the limits of human analysts and existing AI tools, as most models are: i) limited to single data modalities; ii) task-specific; iii) heavily reliant on labeled data. The emerging Foundation Models (FMs) have the potential to address these limitations. Trained on vast unlabeled datasets through self-supervised learning, FMs enable generic feature extraction that facilitate specialization to a wide variety of downstream tasks. This paper describes a vision towards an FM for multimodal Earth Observation data (FM4EO), discussing key building blocks and open challenges. We put particular emphasis on multimodal reasoning, a topic underexplored in EO. Our ultimate goal is a practical path toward FM4EO with capacity to unlock breakthroughs in few-shot learning scenarios, multimodal geographic knowledge integration, synthesis, and hypothesis generation.

Ambrozio Dias, Philipe↗
DOE OSTI2025

Constrained GAN-Generated X-Ray CT Data For Self-Supervised And Foundation-Model Segmentation Of Concrete Microstructures

Three-dimensional characterization of materials using X-ray computed tomography (XCT) is challenging due to the complexity of internal structures, noise, and variations in resolution. Traditional computer vision models often struggle to accurately segment these images, particularly in domain-specific applications like materials science. While supervised deep learning approaches have been developed to address the limitations of conventional algorithms, they typically require large amounts of labeled training data and often fail to generalize across different datasets. Self-supervised, few-and zero-shot learning methods have gained prominence in natural image processing and segmentation tasks, but their application to scientific imaging remains limited due to the unique structural complexity, noise, and textural artifacts present in materials science data. In this work, we investigate how domain adaptation, leveraging physics-based and GAN-generated synthetic data, impacts segmentation performance. We introduce a modified Contrastive Unpaired Translation (CUT) model designed to generate realistic labeled data, which can be used for training, pre-training, and fine-tuning segmentation models for real XCT microstructure data. We evaluate the performance of two segmentation approaches: a self-supervised network (SSL-ALPNet) and a foundation model (Segment Anything Model), assessing their improvements when pre-trained and/or fine-tuned on the synthesized data. Our results demonstrate that leveraging synthetic data significantly enhances segmentation performance, particularly in challenging materials science applications.

Ziabari, Amir [ORNL] (ORCID:000000034776457X)↗
DOE OSTI2021

A probabilistic graphical model foundation for enabling predictive digital twins at scale

A unifying mathematical formulation is needed to move from one-off digital twins built through custom implementations to robust digital twin implementations at scale. This work proposes a probabilistic graphical model as a formal mathematical representation of a digital twin and its associated physical asset. We create an abstraction of the asset–twin system as a set of coupled dynamical systems, evolving over time through their respective state spaces and interacting via observed data and control inputs. The formal definition of this coupled system as a probabilistic graphical model enables us to draw upon well-established theory and methods from Bayesian statistics, dynamical systems and control theory. The declarative and general nature of the proposed digital twin model make it rigorous yet flexible, enabling its application at scale in a diverse range of application areas. Here, we demonstrate how the model is instantiated to enable a structural digital twin of an unmanned aerial vehicle (UAV). The digital twin is calibrated using experimental data from a physical UAV asset. Its use in dynamic decision-making is then illustrated in a synthetic example where the UAV undergoes an in-flight damage event and the digital twin is dynamically updated using sensor data. The graphical model foundation ensures that the digital twin calibration and updating process is principled, unified and able to scale to an entire fleet of digital twins.

42 ENGINEERING↗
DOE OSTI2025

genomeocean: a pretrained microbial genome foundational model (genomeoceanLLM) v1.0

We present Genomeocean, a foundational genome language model that represents the microbial genome sequences from complex environmental samples. By training on a large, diverse metagenomic dataset, Genomeocean learns species-specific sequence composition and can generate long, realistic open reading frames (ORFs). Our model employs a Byte-pair-encoding (BPE) tokenization strategy, allowing it to efficiently process large genomic datasets and generate long sequences up to 50kb. We demonstrate that fine-tuning Genomeocean can generate novel gene clusters encoding biosynthetic pathways, showcasing its ability to model both fundamental and complex biological processes. Our work establishes Genomeocean as a powerful tool for understanding microbial genome biology and paves the way for its application in a range of fields, from synthetic biology to microbiome research.

Wang, Zhong [Lawrence Berkeley National Laboratory↗
DOE OSTI2026

Scalable Federated Learning for Scientific Foundation Models on Leadership-Class Systems

Federated learning (FL) at leadership-class HPC systems remains largely unexplored, despite growing interest in deploying federated workflows on modern HPC systems. This paper provides the first system-level empirical characterization of federated fine-tuning of pretrained foundation models on an exascale supercomputer under a multi-node deployment. Using up to 96 concurrent FL clients deployed across Frontier nodes, we study the impact of client scale, model size, data heterogeneity, partial participation, and differential privacy on runtime, communication overhead, and convergence stability. Our results show that pretrained transformer models remain robust to heterogeneity, client dropout, and privacy noise, while system efficiency degrades rapidly with scale as synchronizat and orchestration dominate runtime. We further demonstrate that system-aware execution strategies, including intra-node aggregation and early aggregation, significantly reduce wall-clock time without degrading model quality. These findings establish a practical performance baseline and inform the design of communication-efficient FL systems on leadership-class HPC platforms.

Kotevska, Olivera [ORNL] (ORCID:0000000316772243)↗
DOE OSTI2023

Towards an astronomical foundation model for stars with a transformer-based model

ABSTRACT Rapid strides are currently being made in the field of artificial intelligence using transformer-based models like Large Language Models (LLMs). The potential of these methods for creating a single, large, versatile model in astronomy has not yet been explored. In this work, we propose a framework for data-driven astronomy that uses the same core techniques and architecture as used by LLMs. Using a variety of observations and labels of stars as an example, we build a transformer-based model and train it in a self-supervised manner with cross-survey data sets to perform a variety of inference tasks. In particular, we demonstrate that a single model can perform both discriminative and generative tasks even if the model was not trained or fine-tuned to do any specific task. For example, on the discriminative task of deriving stellar parameters from Gaia XP spectra, we achieve an accuracy of 47 K in Teff, 0.11 dex in log g, and 0.07 dex in [M/H], outperforming an expert XGBoost model in the same setting. But the same model can also generate XP spectra from stellar parameters, inpaint unobserved spectral regions, extract empirical stellar loci, and even determine the interstellar extinction curve. Our framework demonstrates that building and training a single foundation model without fine-tuning using data and parameters from multiple surveys to predict unmeasured observations and parameters is well within reach. Such ‘Large Astronomy Models’ trained on large quantities of observational data will play a large role in the analysis of current and future large surveys.

Leung, Henry W. (ORCID:0000000200362752)↗
DOE OSTI2025

Active Learning Meets Foundation Models: Fast Remote Sensing Data Annotation for Object Detection

Object detection in remote sensing demands extensive, high-quality annotations—a process that is both labor-intensive and time-consuming. In this work, we introduce a real-time active learning and semi-automated labeling framework that leverages foundation models to streamline dataset annotation for object detection in remote sensing imagery. For example, by integrating a Segment Anything Model (SAM), our approach generates mask-based bounding boxes that serve as the basis for dual sampling: (a) uncertainty estimation to pinpoint challenging samples, and (b) diversity assessment to ensure broad data coverage. Furthermore, our Dynamic Box Switching Module (DBS) addresses the well-known cold start problem for object detection models by replacing its suboptimal initial predictions with SAM-derived masks, thereby enhancing early-stage localization accuracy. Extensive evaluations on multiple remote sensing datasets plus a real-world user study, demonstrate that our framework not only reduces annotation effort, but also significantly boosts detection performance compared to traditional active learning sampling methods. The code for training and the user interface will be made available.

Burges, Marvin [ORNL] (ORCID:0000000312690769)↗
DOE OSTI2025

ORBIT-2: Scaling Exascale Vision Foundation Models for Weather and Climate Downscaling

Sparse observations and coarse-resolution climate models limit effective regional decision-making, underscoring the need for robust downscaling. However, existing AI methods struggle with generalization across variables and geographies and are constrained by the quadratic complexity of Vision Transformer (ViT) self-attention. We introduce ORBIT-2, a scalable foundation model for global, hyper-resolution climate downscaling. ORBIT-2 incorporates two key innovations: (1) Residual Slim ViT (Reslim), a lightweight architecture with residual learning and Bayesian regularization for efficient, robust prediction; and (2) TILES, a tile-wise sequence scaling algorithm that reduces self-attention complexity from quadratic to linear, enabling long-sequence processing and massive parallelism. ORBIT-2 scales to 10 billion parameters across 65,536 GPUs, achieving up to 4.1 ExaFLOPS sustained throughput and 74–98% strong scaling efficiency. It supports downscaling to 0.9 km global resolution and processes sequences up to 4.2 billion tokens. On 7 km resolution benchmarks, ORBIT-2 achieves high accuracy with R2 scores in range of 0.98–0.99 against observation data.

Wang, Xiao [ORNL] (ORCID:0000000165451943)↗
DOE OSTI2024

Maven: a multimodal foundation model for supernova science

Abstract A common setting in astronomy is the availability of a small number of high-quality observations, and larger amounts of either lower-quality observations or synthetic data from simplified models. Time-domain astrophysics is a canonical example of this imbalance, with the number of supernovae observed photometrically outpacing the number observed spectroscopically by multiple orders of magnitude. At the same time, no data-driven models exist to understand these photometric and spectroscopic observables in a common context. Contrastive learning objectives, which have grown in popularity for aligning distinct data modalities in a shared embedding space, provide a potential solution to extract information from these modalities. We present Maven, the first foundation model for supernova science. To construct Maven, we first pre-train our model to align photometry and spectroscopy from 0.5 M synthetic supernovae using a contrastive objective. We then fine-tune the model on 4702 observed supernovae from the Zwicky transient facility. Maven reaches state-of-the-art performance on both classification and redshift estimation, despite the embeddings not being explicitly optimized for these tasks. Through ablation studies, we show that pre-training with synthetic data improves overall performance. In the upcoming era of the Vera C. Rubin observatory, Maven will serve as a valuable tool for leveraging large, unlabeled and multimodal time-domain datasets.

Zhang, Gemma (ORCID:0000000280198082)↗