Search NASA⌕ Search

SEARCH · Search NASA

Results for “mixture-of-experts.”

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.

6 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Mixture-of-Experts for Multi-Domain Defect Identification in Non-Destructive Inspection

Composite materials are widely used in aircraft structures because of their superior mechanical properties. However, their complex failure modes require sophisticated inspection methods to ensure structural integrity. Ultrasonic testing (UT) is a common non-destructive inspection (NDI) technique for aircraft composites that can detect internal and external defects with high resolution and accuracy. Despite their effectiveness, traditional UT methods rely on the manual interpretation of ultrasonic signals, which is time-consuming, labor-intensive, and subjective. Furthermore, processing such large-scale data, particularly across materials of varying thicknesses, significantly increases the computational demands of deep learning model optimization. To overcome these challenges, we propose an efficient sparse mixture-of-experts (MoE) model with a multi-level loss function and introduce four novel training objectives to improve computational efficiency and accuracy in identifying surface defects in composite aircraft materials. Here, we evaluated our approach on material with multiple thicknesses or domains comprising various defects. Our experimental results demonstrate higher accuracy and F1-Score, with only 10% training epochs compared to baseline MoE.

composite materials↗
DOE OSTI2025

MoRE-Brain: Routed Mixture of Experts for Interpretable and Generalizable Cross-Subject fMRI Visual Decoding

Decoding visual experiences from fMRI offers a powerful avenue to understand human perception and develop advanced brain-computer interfaces. However, current progress often prioritizes maximizing reconstruction fidelity while overlooking interpretability, an essential aspect for deriving neuroscientific insight. To address this gap, we propose MoRE-Brain, a neuro-inspired framework designed for high-fidelity, adaptable, and interpretable visual reconstruction. MoRE-Brain uniquely employs a hierarchical Mixture-of-Experts architecture where distinct experts process fMRI signals from functionally related voxel groups, mimicking specialized brain networks. The experts are first trained to encode fMRI into the frozen CLIP space. A finetuned diffusion model then synthesizes images, guided by expert outputs through a novel dual-stage routing mechanism that dynamically weighs expert contributions across the diffusion process. MoRE-Brain offers three main advancements: First, it introduces a novel Mixture-of-Experts architecture grounded in brain network principles for neuro-decoding. Second, it achieves efficient cross-subject generalization by sharing core expert networks while adapting only subject-specific routers. Third, it provides enhanced mechanistic insight, as the explicit routing reveals precisely how different modeled brain regions shape the semantic and spatial attributes of the reconstructed image. Extensive experiments validate MoRE-Brain’s high reconstruction fidelity, with bottleneck analyses further demonstrating its effective utilization of fMRI signals, distinguishing genuine neural decoding from over-reliance on generative priors. Consequently, MoRE-Brain marks a substantial advance towards more generalizable and interpretable fMRI-based visual decoding.

Wei, Yuxiang [Georgia Institute of Technology]↗
DOE OSTI2024

Unsupervised physics-informed disentanglement of multimodal data

Here, we introduce physics-informed multimodal autoencoders (PIMA) - a variational inference framework for discovering shared information in multimodal datasets. Individual modalities are embedded into a shared latent space and fused through a product-of-experts formulation, enabling a Gaussian mixture prior to identify shared features. Sampling from clusters allows cross-modal generative modeling, with a mixture-of-experts decoder that imposes inductive biases from prior scientific knowledge and thereby imparts structured disentanglement of the latent space. This approach enables cross-modal inference and the discovery of features in high-dimensional heterogeneous datasets. Consequently, this approach provides a means to discover fingerprints in multimodal scientific datasets and to avoid traditional bottlenecks related to high-fidelity measurement and characterization of scientific datasets.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Development of a Machine-Learned Cruise Guide Indicator for Rotorcraft

This paper presents a machine-learned virtual cruise guide indicator (vCGI) for Chinook helicopters. Two temporal neural networks were trained and evaluated on measured data from 55 flight tests, one for the fore rotor and another for the aft rotor, to predict a vCGI value, which protects 23 components from fatigue damage during steady-state conditions. Three different classes of machine learning architectures were evaluated for prediction of the vCGI from time sequences: a temporal convolutional neural network with 1D dilated causal convolutions, a long short-term memory recurrent neural network, and an attention-based transformer architecture. The final average model accuracy on unseen flight data is currently greater than 93% for CGI values which could result in fatigue damage and 90% for normal operation CGI values. Model accuracy was improved through a series of advancements in:(1) selection of optimal training data using temporal collective variables and unsupervised learning, (2) dataset augmentation with maximum-entropy temporal collective variables, and (3) implementation of a mixture-of-experts classification- regression approach using an adversarial classification approach to assign maneuver labels. The results are presented for each advancement in model development along with lessons learned in training machine learning models on real- world, time-dependent rotorcraft data.

Boyer, Mathew↗
DOE OSTI2026

HARMONY: Large-Scale Architecture Search for Efficient Hybrid Language Models

As large language models scale to trillions of parameters, their computational and memory requirements present critical challenges for efficient training and deployment. While Mixture of Experts (MoE) architectures enable efficient scaling through sparse parameter activation, and state-space models like Mamba offer linear-time complexity, principled methods for combining these paradigms remain undeveloped. We introduce HARMONY (Hybrid Architecture Research for Mamba, Optimized with Neural efficiencY), a multi-objective evolutionary neural architecture search framework for discovering efficient hybrid language models that integrate Transformer attention mechanisms, Mixture-of-Experts routing, and Mamba state-space components. Through large-scale distributed search using 16,384 MI250X GPUs on the Frontier supercomputer, HARMONY explores a comprehensive design space encompassing six attention variants (MHA, MQA, GQA, MLA, SWA, and Mamba-2), variable MoE configurations with both routed and shared experts, and extensive Mamba hyperparameters. Our framework discovers heterogeneous architectures that balance training performance with computational efficiency through multi-objective optimization incorporating latency penalties and fitness-based selection. Analysis of discovered architectures reveals that optimal hybrid designs favor heterogeneous component mixing rather than homogeneous patterns, with Mamba-2 and Multi-Head Latent Attention (MLA) emerging as preferred mechanisms. Discovered architectures demonstrate superior training efficiency: our best configuration achieves a final perplexity of 1.0874 with 2.38B parameters while processing 4,320 tokens/second, outperforming significantly larger manually designed models. Full-scale evaluation shows HARMONY's top architectures achieve better loss trajectories than equivalently-sized models using state-of-the-art configurations including Mixtral, Jamba, and Samba. Additionally, we demonstrate 91% weak scaling efficiency when training discovered 36B-parameter models across 1,024 GPUs. HARMONY is released as an open framework with comprehensive tools for building and training hybrid models using expert-data-pipeline parallelism, democratizing access to automated architecture design for next-generation language models.

Herron, Emily [ORNL] (ORCID:0000000273008172)↗