SEARCH · Search NASA
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.
A Flash Attention Triton kernel with support for second-order derivatives, such as Jacobian-Vector Products (JVPs) and Hessian-Vector Products (HVPs).
Explore the source record for details and available documents.
With the increasing complexity and data availability in modern power systems, learning-based approaches to AC Optimal Power Flow (AC OPF) have garnered significant attention. In particular, the structure of smart grids lends itself naturally to graph-based representations, where Graph Neural Networks (GNNs) can capture spatial and relational dependencies. This paper investigates attention-based GNN architectures tailored to heterogeneous graph representations of electric grids. We evaluate two major paradigms: relational attention, which distinguishes between edge types during message passing, and meta-path attention, which captures high-level semantics through multi-hop, typed paths. Using a large corpus of public AC OPF scenarios, we benchmark representative models of each type of attention. Our results demonstrate the benefits of heterogeneous attention-based models in accurately capturing grid dynamics; heterogeneous attention models achieve superior performance in both standard and perturbed settings. The findings highlight the importance of semantic-aware architectures for improving prediction robustness and interpretability in power system applications.
Detection of unknown waveforms in mission-critical communications is a crucial area of interest for the Department of Energy (DoE). Traditional methods and recent deep learning-based approaches often assume that the training set includes all possible classes, which is impractical for detecting new waveforms. This limitation gives rise to the problem of open-set recognition (OSR), which involves correctly identifying known classes while detecting and rejecting unknown or unseen classes. To address this limitation, we propose a novel dual-domain complex-valued neural architecture that jointly processes time-domain and frequency-domain signal representations using transformer mechanisms. A transformer model is a deep learning architecture that uses self-attention mechanisms to process and learn relationships in sequential data. Our model employs a cosine similarity loss to extract domain-specific features and incorporates a transformer architecture in the latent space to weigh the importance of different features from the time and frequency domains. The transformer layer includes stacked self-attention and cross-attention modules to learn intra-domain and inter-domain dependencies, creating a more holistic signal representation. An attention-based fusion module intelligently combines the time and frequency-domain features using multi-head attention, enabling the network to learn the optimal feature for each domain in each input signal. Quantitative results demonstrate the impact of these architectural choices on overall performance, showing significant improvement after incorporating self and cross-attention modules and using complex attention fusion over simple weighted fusion. Our ongoing work will focus on addressing the limitations of threshold-based OSR methods by developing a novel generative framework that integrates a conditional diffusion probabilistic model (DPM). DPM is a generative framework that learns to synthesize complex data by reversing a gradual noising process using a neural network trained to denoise step-by-step. Our goal is to leverage the inherent strengths of DPMs for identifying unknown signals more robustly. One primary advantage of using a DPM is its ability to provide a more reliable anomaly score based on the model's reconstruction error, rather than relying solely on classifier confidence. Additionally, the iterative denoising process of DPMs makes this approach naturally resilient to low Signal-to-Noise Ratio (SNR) conditions, where traditional methods often fail. By implementing this generative framework, we aim to enhance the model's capability to accurately detect unknown waveforms and maintain performance in challenging environments.
Objective: Improving the AI-driven automated medical encoding of clinical text plays a vital role in gathering information on the occurrence of diseases to improve population-level health. This work presents a novel attention mechanism designed to enhance text classification models and ensure appropriate classification of medical concepts in unstructured electronic health records. Materials and Methods: We developed a deformable, phrase-level attention mechanism to identify important lexical word-level and contextual phrase-level information from clinical text documents. We evaluated conventional and transformer-based deep learning models that we extended with our attention mechanism on the extraction of critical cancer information (e.g., site, subsite, laterality, histology, behavior) from 629,908 electronic pathology reports and on the automated medical encoding of 52,722 hospital discharge summaries. Results: Transformer-based models with the deformable, phrase-level attention mechanism achieved the best performance on the extraction of critical cancer information from pathology reports. Conventional- and transformer-based models show similar or better performance than their baseline counterparts on the automated medical encoding of clinical documents. Discussion: The addition of phrase-level information allowed models extended with our proposed method to outperform standard word-level attention. Our method showed favorable properties for the real-world application in terms of model robustness and phenotyping. These results indicate that our method is promising for automated data harmonization for common data models. Conclusion: This work proposes a novel deformable, phrase-level attention mechanism that enhances text classification models in the extraction of medical concepts from clinical text documents. We demonstrate strong performances on two clinical text datasets and showcase real-world deployability of our method.
The attention mechanism has become foundational for remarkable AI breakthroughs since the introduction of the Transformer, driving the demand for increasingly longer context to power frontier models such as large-scale reasoning language models and high-resolution image/video generators. However, its quadratic computational and memory complexities present substantial challenges. Current state-of-the-art parallel attention methods, such as ring attention, are widely adopted for long-context training but utilize a point-to-point communication strategy that fails to fully exploit the capabilities of modern HPC network architectures. In this work, we propose ringX, a scalable family of parallel attention methods optimized explicitly for HPC systems. By enhancing workload partitioning, refining communication patterns, and improving load balancing, ringX achieves up to 3.4 × speedup compared to conventional ring attention on the Frontier supercomputer. Optimized for both bi-directional and causal attention mechanisms, ringX demonstrates its effectiveness through training benchmarks of a Vision Transformer (ViT) on a climate dataset and a Generative Pre-Trained Transformer (GPT) model, Llama3 8B. Our method attains an end-to-end training speedup of approximately 1.5 × in both scenarios. To our knowledge, the achieved 38% model FLOPs utilization (MFU) for training Llama3 8B with a 1M-token sequence length on 4,096 GPUs represents one of the highest training efficiencies reported for long-context learning on HPC systems. Our code implementation is available at https://github.com/jqyin/ringX-attention.
Graph neural networks (GNNs) are widely used as surrogates for costly experiments and first-principles simulations to study the behavior of compounds at atomistic scale, and their architectural complexity is constantly increasing to enable the modeling of complex physics. While most recent GNNs combine more traditional message passing neural networks (MPNNs) layers to model short-range interactions with more advanced graph transformers (GTs) with global attention mechanisms to model long-range interactions, it is still unclear when global attention mechanisms provide real benefits over well-tuned MPNN layers due to inconsistent implementations, features, or hyperparameter tuning. We introduce the first unified, reproducible benchmarking framework–built on HydraGNN–that enables seamless switching among four controlled model classes: MPNN, MPNN with chemistry/topology encoders, GPS-style hybrids of MPNN with global attention, and fully fused localglobal models with encoders. Using seven diverse open-source datasets for benchmarking across regression and classification tasks, we systematically isolate the contributions of message passing, global attention, and encoder-based feature augmentation. Our study shows that encoder-augmented MPNNs form a robust baseline, while fused localglobal models yield the clearest benefits for properties governed by long-range interaction effects. We further quantify the accuracycompute trade-offs of attention, reporting its overhead in memory. Together, these results establish the first controlled evaluation of global attention in atomistic graph learning and provide a reproducible testbed for future model development.
Machine learning methods are emerging as a universal paradigm for constructing correlative structure–property relationships in materials science based on multimodal characterization. However, this necessitates the development of methods for the physical interpretability of the resulting correlative models. Here, we demonstrate the potential of attention-based neural networks for revealing structure–property relationships and the underlying physical mechanisms, using the ferroelectric properties of PbTiO3 thin films as a case study. Through the analysis of attention scores, we disentangle the influence of distinct domain patterns on the polarization switching process. The attention-based Transformer model is explored both as a direct interpretability tool and as a surrogate for explaining representations learned via unsupervised machine learning, enabling the identification of physically grounded correlations. We compare attention-derived interpretability scores with classical SHapley Additive exPlanations analysis and show that, in contrast to applications in natural language processing, attention mechanisms in materials science exhibit high efficiency in highlighting meaningful structural features.
Transformer-based models have achieved state-of-the-art performance in jet tagging at the CERN Large Hadron Collider (LHC), with the Particle Transformer (ParT) representing a leading example of such models. A striking feature of ParT is its sparse, nearly binary, attention structure, raising questions about the origin of this behavior and whether it encodes physically meaningful correlations. In this work, we investigate the source of ParT's sparse attention by comparing models trained on multiple benchmark datasets and examine the relative contributions of the attention term and the physics-inspired interaction matrix before softmax. We find that binary sparsity arises primarily from the attention mechanism itself, with the interaction matrix playing a secondary role. Moreove, we show that ParT is able to identify key jet substructure elements, such as leptons in semileptonic top decays, even without explicit particle identification inputs. These results provide new insight into the interpretability of transformer-based jet taggers and clarify the conditions under which sparse attention patterns emerge in ParT.
Current state-of-the-art (SOTA) LiDAR-only detectors perform well for 3D object detection tasks, but point cloud data are typically sparse and lacks semantic information. Detailed semantic information obtained from camera images can be added with existing LiDAR-based detectors to create a robust 3D detection pipeline. With two different data types, a major challenge in developing multi-modal sensor fusion networks is to achieve effective data fusion while managing computational resources. With separate 2D and 3D feature extraction backbones, feature fusion can become more challenging as these modes generate different gradients, leading to gradient conflicts and suboptimal convergence during network optimization. To this end, we propose a 3D object detection method, Attention-Enabled Point Fusion (AEPF). AEPF uses images and voxelized point cloud data as inputs and estimates the 3D bounding boxes of object locations as outputs. An attention mechanism is introduced to an existing feature fusion strategy to improve 3D detection accuracy and two variants are proposed. These two variants, AEPF-Small and AEPF-Large, address different needs. AEPF-Small, with a lightweight attention module and fewer parameters, offers fast inference. AEPF-Large, with a more complex attention module and increased parameters, provides higher accuracy than baseline models. Experimental results on the KITTI validation set show that AEPF-Small maintains SOTA 3D detection accuracy while inferencing at higher speeds. AEPF-Large achieves mean average precision scores of 91.13, 79.06, and 76.15 for the car class’s easy, medium, and hard targets, respectively, in the KITTI validation set. Results from ablation experiments are also presented to support the choice of model architecture.
Real-time jet tagging is critical for identifying short-lived particle decays in the high-throughput detectors of the Large Hadron Collider, where real-time trigger systems responsible for deciding which collision events to store impose strict latency and accuracy constraints. While transformer architectures achieve the highest jet tagging accuracy when compute is unconstrained, their quadratic self-attention cost makes inference restrictive on trigger budget. Existing efficient variants reduce the computational cost, but hinder the classification performance. To address this limitation, we introduce the Patch Hierarchical Attention Transformer (PHAT-JeT), which combines two mechanisms: a physics-inspired geometric message-passing module that encodes local detector-plane structure, and a hierarchical patch-based attention scheme that computes exact attention within small particle groups while preserving global context through lightweight patch-token communication. Within a restricted budget, PHAT-JeT achieves state-of-the-art accuracy and background rejection among all resource-constrained jet tagging models on four benchmarks (\textsc{hls4ml}, JetClass, Top Tagging, and Quark--Gluon). Our code is available at https://github.com/aaronw5/PHAT-JeT.
The imminent era of error-corrected quantum computing demands robust methods to characterize quantum state complexity from limited, noisy measurements. We introduce the Quantum Attention Network (QuAN), a classical artificial intelligence (AI) framework leveraging attention mechanisms tailored for learning quantum complexity. Inspired by large language models, QuAN treats measurement snapshots as tokens while respecting permutation invariance. Combined with our parameter-efficient miniset self-attention block, this enables QuAN to access high-order moments of bit-string distributions and preferentially attend to less noisy snapshots. We test QuAN across three quantum simulation settings: driven hard-core Bose-Hubbard model, random quantum circuits, and toric code under coherent and incoherent noise. QuAN directly learns entanglement and state complexity growth from experimental computational basis measurements, including complexity growth in random circuits from noisy data. In regimes inaccessible to existing theory, QuAN unveils the complete phase diagram for noisy toric code data as a function of both noise types, highlighting AI’s transformative potential for assisting quantum hardware.
In the evolving digital landscape, fake news is a significant challenge, influencing public perception and decision-making. Traditional detection approaches focus on single-modal data or simple multimodal fusion, often overlooking deeper interactions and news credibility. We propose a novel model addressing these limitations by introducing rotational attention and news domain information as a feature. Unlike static attention mechanisms, our rotational attention dynamically shifts query, key, and value roles across text and image inputs, enabling richer cross-modal interaction. Incorporating news domain information further enhances the model’s reliability by associating news posts with top domains extracted from Google search results, reducing false detections. This approach assesses both the content and the broader web context in which the news is discussed. Our model outperforms existing state-of-the-art methods by providing deeper, layered multimodal integration and domain information analysis, resulting in a more robust and adaptive fake news detection system.
Additive manufacturing (AM), while commonly used for rapid prototyping and creating components with complex geometries, has not been widely adopted for critical applications across the aerospace, automotive, defense, energy, and medical industries. This is, in part, due to the challenges of controlling flaws and uncertainty in the mechanical behavior of additively manufactured components. In recent years, there has been an increase in research aimed at predicting the final mechanical properties of additively manufactured components during the printing process. To address these issues, a 3D-CNN model was trained using low-cost in situ visible-light camera data, anomaly classifications, and the chosen process parameters to predict the ultimate tensile strength (UTS), yield strength (YS), total elongation (TE), and uniform elongation (UE). The 3D-CNN layers of the model employed attention mechanisms to prioritize features in the data, thereby improving prediction accuracy. Furthermore, the effect of each process parameter and anomaly class is investigated using attention-based dynamic sigmoid weighted gates to interpret the influence each class has on the final prediction. Different combinations of the in situ data were fed into the 3D-CNN, with varying amounts of image layers, to determine the ideal combination for predicting mechanical properties in situ. Here, the 3D-CNN model achieved mean absolute percentage errors (MAPE) below 5% for both UTS and YS while using only a single camera input and under half of the available image layers.
Accurate day-ahead load forecasting is essential for reliable power system operations and market efficiency. System operators such as the Midcontinent Independent System Operator (MISO) rely on forecasts from multiple vendors, yet combining them effectively remains a persistent challenge due to vendor-specific biases. This paper presents a novel LSTM-Attention Fusion Network with Error Representation (LAF-Net) that enhances day-ahead hourly load forecasting through deep residual learning and multi-modal cross-attention. The proposed model builds a historical error memory from past vendor performance and dynamically queries it with future hour context to generate adaptive, hour-specific trust weights for each vendor. A bounded residual correction further refines forecasts by mitigating systematic and temporally localized errors. Tested on real MISO LBA data with multi-vendor forecasts, LAF-Net consistently outperforms the best vendor baseline across all 38 LBAs, achieving more than a 40% reduction in system-level mean absolute error (MAE) during peak load hours relative to the best vendor baseline.
Magnetic particle imaging (MPI) is an emerging noninvasive molecular imaging modality with high sensitivity and specificity, exceptional linear quantitative ability, and potential for successful applications in clinical settings. Computed tomography (CT) is typically combined with the MPI image to obtain more anatomical information. Herein, a deep learning‐based approach for MPI‐CT image segmentation is presented. The dataset utilized in training the proposed deep learning model is obtained from a transgenic mouse model of breast cancer following administration of indocyanine green (ICG)‐conjugated superparamagnetic iron oxide nanoworms (NWs‐ICG) as the tracer. The NWs‐ICG particles progressively accumulate in tumors due to the enhanced permeability and retention (EPR) effect. The proposed deep learning model exploits the advantages of the multihead attention mechanism and the U‐Net model to perform segmentation on the MPI‐CT images, showing superb results. In addition, the model is characterized with a different number of attention heads to explore the optimal number for our custom MPI‐CT dataset.
Identifying molecular structure based on spectroscopic readings is a key task in a variety of chemical and biological applications. Common spectroscopy techniques, such as Infrared (IR) Spectroscopy and Mass Spectrometry (MS), provide detailed information on the structure of molecular compounds but nonetheless require expert-level knowledge to decode. Machine learning has emerged as a potential solution for automating structure prediction from chemical spectra; however, current approaches generally focus on single sensor modalities, neglecting to leverage the complementary information contained within differing spectra. In this paper, we introduce Peak2Patch, a novel approach to fusion-enhanced prediction of functional groups from IR and mass spectra. First, we perform a detailed comparison of backbone networks for encoding both sparse mass spectra and dense IR spectra and demonstrate the superior performance of transformer neural networks over current state-of-the-art convolutional neural networks. Second, we evaluate three broad categories of fusion: early (raw feature), middle (deep feature), and late (decision) fusion, demonstrating the potential of a deep feature fusion-based approach. Lastly, we present Peak2Patch, our attention-based fusion scheme, which leverages cross-attention to mix features between encoded tokens of the two modalities. We validate our approach on a publicly available multimodal spectroscopic data set of 790k simulated molecules, demonstrating a large improvement in functional group prediction over both the previous state-of-the-art and our own strong single-modal baselines.
Machine learning (ML) offers considerable promise for the design of new molecules and materials. In real-world applications, the design problem is often domain-specific, and suffers from insufficient data, particularly labeled data, for ML training. In this study, we report a data-efficient, deep-learning framework for molecular discovery that integrates a coarse-grained functional-group representation with a self-attention mechanism to capture intricate chemical interactions. Our approach exploits group-contribution concepts to create a graph-based intermediate representation of molecules, serving as a low-dimensional embedding that substantially reduces the data demands typically required for training. Using a self-attention mechanism to learn the subtle but highly relevant chemical context of functional groups, the method proposed here consistently outperforms existing approaches for predictions of multiple thermophysical properties. In a case study focused on adhesive polymer monomers, we train on a limited dataset comprising only 6,000 unlabeled and 600 labeled monomers. The resulting chemistry prediction model achieves over 92% accuracy in forecasting properties directly from SMILES strings, exceeding the performance of current state-of-the-art techniques. Furthermore, the latent molecular embedding is invertible, enabling the design pipeline to automatically generate new monomers from the learned chemical subspace. We illustrate this functionality by targeting several properties, including high and low glass transition temperatures (Tg), and demonstrate that our model can identify new candidates with values that surpass those in the training set. The ease with which the proposed framework navigates both chemical diversity and data scarcity offers a promising route to accelerate and broaden the search for functional materials.