Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “automatic adaptivity”

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 199 records · Page 11
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

A Combined Computer Vision and Deep Learning Approach for Rapid Drone-Based Optical Characterization of Parabolic Troughs

Optical accuracy is a primary driver of parabolic trough concentrating solar power (CSP) plant performance, but can be damaged by wind loads, gravity, error during installation, and regular plant operation. Collecting and analyzing optical measurements over an entire operating parabolic trough plant is difficult, given the large scale of typical installations. Distant Observer, a software tool developed at the National Renewable Energy Laboratory, uses images of the absorber tube reflected in the collector mirror to measure both surface slope in the parabolic mirror and offset of the absorber tube from the ideal focal point. This technology has been adapted for fast data collection using low-cost commercial drones, but until recently still required substantial human labor to process large amounts of data. A new method leveraging advanced deep learning and computer vision tools can drastically reduce the time required to process images. This new method addresses the primary analysis bottleneck, identifying featureless, reflective mirror corner points to a high degree of accuracy. Recent work has shown promising results using computer vision methods. The combined deep learning and computer vision approach presented here proved highly effective and has the potential to further automate data collection and analysis, making the tool more robust. The method presented in this paper automatically identified 74.3% of mirror corners within 2 pixels of their manually marked counterparts and 91.9% within 3 pixels. This level of accuracy is sufficient for practical Distant Observer analysis within a target uncertainty. A commercial drone collected video of over 100 parabolic trough modules at an operating CSP plant to demonstrate the deep learning and computer vision method's usefulness in processing large amounts of data. These troughs were successfully analyzed using Distant Observer, paired with the new deep learning and computer vision algorithm, and can provide plant operators and trough designers with valuable insight about plant performance, operating strategies, and plant-wide optical error trends.

computer vision↗
DOE OSTI2025

Extending JuTrack’s capabilities to the FRIB accelerator to enhance online modeling

JuTrack is a Julia-based accelerator modeling and tracking package that utilizes compiler-level automatic differentiation (AD) to enable fast and accurate derivative calculations. While JuTrack provides a solid foundation for beam dynamics simulations, its capabilities must be extended to support the Facility for Rare Isotopes (FRIB) linac. This includes modeling heavy-ion linac accelerator components such as the liquid-lithium charge stripper, which facilitates efficient acceleration by remove electrons from heavy isotopes, and incorporating multi-charge state acceleration tracking, which allows for charge-dependent beam dynamics. These extensions address challenges such as the beam matching and optimization of multi charge state through various accelerating structures and beam-material interaction modeling while maintaining the auto differentiation capability. This work focuses on adapting JuTrack to incorporate these elements, enhancing its online modeling abilities. We present modifications to JuTrack’s framework and demonstrate their performance in FRIB simulations.

Accelerator Physics↗
DOE OSTI2024

Measuring stomatal and guard cell metrics for plant physiology and growth using StoManager1

Abstract Automated guard cell detection and measurement are vital for understanding plant physiological performance and ecological functioning in global water and carbon cycles. Most current methods for measuring guard cells and stomata are laborious, time-consuming, prone to bias, and limited in scale. We developed StoManager1, a high-throughput tool utilizing geometrical, mathematical algorithms, and convolutional neural networks to automatically detect, count, and measure over 30 guard cell and stomatal metrics, including guard cell and stomatal area, length, width, stomatal aperture area/guard cell area, orientation, stomatal evenness, divergence, and aggregation index. Combined with leaf functional traits, some of these StoManager1-measured guard cell and stomatal metrics explained 90% and 82% of tree biomass and intrinsic water use efficiency (iWUE) variances in hardwoods, making them substantial factors in leaf physiology and tree growth. StoManager1 demonstrated exceptional precision and recall (mAP@0.5 over 0.96), effectively capturing diverse stomatal properties across over 100 species. StoManager1 facilitates the automation of measuring leaf stomatal and guard cells, enabling broader exploration of stomatal control in plant growth and adaptation to environmental stress and climate change. This has implications for global gross primary productivity (GPP) modeling and estimation, as integrating stomatal metrics can enhance predictions of plant growth and resource usage worldwide. Easily accessible open-source code and standalone Windows executable applications are available on a GitHub repository (https://github.com/JiaxinWang123/StoManager1) and Zenodo (https://doi.org/10.5281/zenodo.7686022).

Wang, Jiaxin (ORCID:0000000348085085)↗
DOE OSTI2026

Machine learning at the Spallation Neutron Source accelerator and target

We describe the ongoing efforts to apply Machine Learning techniques to improve the performance of our accelerator and target. Specially, we are looking to minimize halo beam losses in the absence of a proper physics model, automatically detect and log anomalies in the target support systems such as cooling, and detect and prevent errant beam pulses in the linac. We also describe the infrastructure we use to acquire and stream data to the GPU cluster for training, our code development cycle, and edge computing for model inference. To minimize halo beam losses, we use a Reinforcement Learning technique tested on a virtual accelerator. The target anomaly detection is trained on archived data using incomplete physics models and is made part of the existing target reporting system. The errant beam prevention analyzes beam current and beam phase waveforms as well as accelerator configuration data to predict errant pulses. We also develop continual learning to adapt to changes in the accelerator.

Accelerator Physics↗
DOE OSTI2024

Robust and Simple ADMM Penalty Parameter Selection

We present a new method for online selection of the penalty parameter for the alternating direction method of multipliers (ADMM) algorithm. ADMM is a widely used method for solving a range of optimization problems, including those that arise in signal and image processing. In its standard form, ADMM includes a scalar hyperparameter, known as the penalty parameter, which usually has to be tuned to achieve satisfactory empirical convergence. In this work, we develop a framework for analyzing the ADMM algorithm applied to a quadratic problem as an affine fixed point iteration. Using this framework, we develop a new method for automatically tuning the penalty parameter by detecting when it has become too large or small. We analyze this and several other methods with respect to their theoretical properties, i.e., robustness to problem transformations, and empirical performance on several optimization problems. Our proposed algorithm is based on a theoretical framework with clear, explicit assumptions and approximations, is theoretically covariant/invariant to problem transformations, is simple to implement, and exhibits competitive empirical performance.

42 ENGINEERING↗
DOE OSTI2020

Quantifying patterns in optical micrographs of one- and two-dimensional ellipsoidal particle assemblies

Current developments in colloidal science include the assembly of anisotropic colloids with broad geometric diversity. As the complexity of particle assemblies increases, the need for ubiquitous algorithms that quantitatively analyze images of the assemblies to deliver key information such as quantification of crystal structures becomes more urgent. This contribution describes algorithms capable of image analysis for classifying colloidal structures based on abstracted interparticle relationship information and quantitatively analyzing the abundance of each structure in mixed pattern assemblies. The algorithm parameters can be adjusted, allowing for the algorithms to be adapted for different image analyses. Three different ellipsoidal particle assembly images are presented to demonstrate the effectiveness of the algorithms: a one-dimensional (1D) particle chain assembly and two two-dimensional (2D) polymorphic crystals each consisting of assemblies of two distinct plane symmetry groups. Angle relationships between neighbouring particles are calculated and neighbour counts of each particle are determined. Combining these two parameters as rules for classification criteria allows for the labeling and quantification of each particle into a defined symmetry class within an assembly. The algorithms provide a labelled image comprising classification results and particle counts of each defined class. For multiple images or individual frames from a video, the script can be looped to achieve automatic processing. The yielded classification data allow for more in-depth image analysis of mixed pattern particle assemblies. Finally, we envision that these algorithms will have utility in quantitative analysis of images comprising ellipsoidal colloidal materials, nanoparticles, or biological matter.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

TChem-atm (v2.0.0): scalable performance-portable multiphase atmospheric chemistry

We present TChem-atm, a performance-portable approach that enables efficient simulation of chemically detailed and multiphase atmospheric chemistry on modern heterogeneous computing architectures. Unlike previous efforts that rely on architecture-specific code or focus exclusively on gas-phase chemistry, TChem-atm supports fully coupled gas–aerosol systems with execution across CPUs, NVIDIA GPUs, and AMD GPUs through the Kokkos programming model. It integrates the flexible multiphase capabilities of the Community Atmospheric Model Chemistry Package (CAMP) with the high-performance kinetic routines of TChem, and includes automatic Jacobian construction with support for a range of stiff ODE solvers. In a proof-of-concept integration with the particle-resolved model PartMC, TChem-atm reproduces the existing PartMC–CAMP implementation within solver tolerances and delivers substantial GPU speedups, especially for large particle populations. Performance benchmarks reveal substantial speedups on GPU platforms, particularly for large particle populations, with consistent results across hardware backends. TChem-atm enables performance-portable execution across CPUs and GPUs, though optimal efficiency may require modest architecture-specific tuning (e.g., team and vector sizes), with up to a twofold improvement on the NVIDIA H100. It directly supports sectional and particle-resolved host models, while modal aerosol schemes require minor adaptation to provide particle-scale quantities such as representative diameters. By enabling chemically detailed, multiphase simulations with performance portability and host-model flexibility, TChem-atm facilitates the incorporation of advanced chemistry into atmospheric models.

Díaz-Ibarra, Oscar Homero [Sandia National Laborat↗
DOE OSTI2025

Multi-objective automatic discovery of optimized metamaterials for varying velocity impact protection

Mechanical metamaterials have demonstrated exceptional impact performance while remaining lightweight. Impact resistance has traditionally been investigated using quasi-static simulations, often with the assumption that performance will translate to high-velocity impact scenarios. However, critical crash protection parameters—such as peak stress and absorbed energy—are highly sensitive to impact velocity, leading to inconsistent performance under dynamic loading. To address this, we introduce a strain-rate-aware, active deep learning framework that enables multi-objective optimization of impact protection metrics across a wide range of impact velocities. Our framework captures the strain-rate sensitivity of architected lattices by learning to control spatial gradation in cellular metamaterials, resulting in over 200 % enhancement in impact protection relative to state-of-the-art designs such as Voronoi and re-entrant lattices. We demonstrate its practical utility by designing next-generation lattice structures for automotive bumper systems that satisfy multiple, velocity-specific safety criteria—capabilities beyond those of conventional designs. More than just a predictive tool, this framework marks the first step towards enabling adaptable impact-resistant structures across dynamic regimes.

Deep learning↗
DOE OSTI2021

High-Level Synthesis of Parallel Specifications Coupling Static and Dynamic Controllers

The increased need for efficient ways to implement domain-specific accelerators is driving design methodologies towards the use of abstractions higher than the Register Transfer Level (RTL). In this scenario, High Level Synthesis (HLS) plays a significant role by enabling the automatic generation of custom hardware accelerators starting from high level descriptions (e.g., C code). Conventional HLS tools exploit parallelism mostly at the Instruction Level (ILP). They statically schedule the input specifications, and build centralized Finite State Machine (FSM) controllers. However, aggressive exploitation of ILP in many applications has diminishing returns and, usually, centralized approaches do not efficiently exploit coarser parallelism because FSMs are inherently serial. In this paper we present a HLS framework able to synthesize applications that, beside ILP, also expose Task Level Parallelism (TLP). An application can expose TLP through annotations that identify the parallel functions (i.e., tasks). To generate accelerators that efficiently execute concur- rent tasks, we need to solve several issues: devise a mechanism to support concurrent execution flows, exploit memory parallelism, and manage synchronization. To support concurrent execution flows, we introduce a novel adaptive controller. The adaptive controller is composed of a set of interacting control elements that independently manage the execution of a single operation or function call. These control elements check dependencies and resource constraints at runtime, enabling as soon as possible execution. To support parallel access to shared memories and synchronization, we introduce a novel Hierarchical Memory Interface (HMI). With respect to previous solutions, the proposed interface supports multi-ported memories and atomic memory operations, which commonly occur in parallel programming. Our framework can generate the hardware implementation of C functions by employing two different approaches, depending on its characteristics. If a function exposes TLP, then the framework generates hardware implementations based on the adaptive controller. Otherwise, the framework implements the function by exploiting a more conventional FSM approach, which is optimized for ILP exploitation. We evaluate our framework on a set of parallel applications, and show substantial performance improvements (average speedup of 4.7) with limited area over- heads (average area increase of 5.48 times).

Castellana, Vito G.↗
DOE OSTI2021

High-Level Synthesis of Parallel Specifications Coupling Static and Dynamic Controllers

The increased need for efficient ways to implement domain-specific accelerators is driving design methodologies towards the use of abstractions higher than the Register Transfer Level (RTL). In this scenario, High Level Synthesis (HLS) plays a significant role by enabling the automatic generation of custom hardware accelerators starting from high level descriptions (e.g., C code). Conventional HLS tools exploit parallelism mostly at the Instruction Level (ILP). They statically schedule the input specifications, and build centralized Finite State Machine (FSM) controllers. However, aggressive exploitation of ILP in many applications has diminishing returns and, usually, centralized approaches do not efficiently exploit coarser parallelism because FSMs are inherently serial. In this paper we present a HLS framework able to synthesize applications that, beside ILP, also expose Task Level Parallelism (TLP). An application can expose TLP through annotations that identify the parallel functions (i.e., tasks). To generate accelerators that efficiently execute concur- rent tasks, we need to solve several issues: devise a mechanism to support concurrent execution flows, exploit memory parallelism, and manage synchronization. To support concurrent execution flows, we introduce a novel adaptive controller. The adaptive controller is composed of a set of interacting control elements that independently manage the execution of a single operation or function call. These control elements check dependencies and resource constraints at runtime, enabling as soon as possible execution. To support parallel access to shared memories and synchronization, we introduce a novel Hierarchical Memory Interface (HMI). With respect to previous solutions, the proposed interface supports multi-ported memories and atomic memory operations, which commonly occur in parallel programming. Our framework can generate the hardware implementation of C functions by employing two different approaches, depending on its characteristics. If a function exposes TLP, then the framework generates hardware implementations based on the adaptive controller. Otherwise, the framework implements the function by exploiting a more conventional FSM approach, which is optimized for ILP exploitation. We evaluate our framework on a set of parallel applications, and show substantial performance improvements (average speedup of 4.7) with limited area over- heads (average area increase of 5.48 times).

Castellana, Vito G.↗
DOE OSTI2022

Mitigate: An Adaptive Network Data Anonymization Tool Using Condensation-Based Differential Privacy

Modern network devices collect a large amount of data that can be analyzed to identify bottlenecks, anomalies, cyber-attacks, etc. Therefore, there is often a need to analyze such collections of network data quite often by an external expert or by the research community. However, these collections of data contain sensitive, proprietary information. In order for the network data to be shared, it must first be anonymized. The overall objective of this project is to develop an innovative privacy management tool to anonymize network data and achieve sufficient privacy, acceptable data utility, and efficient data analysis at the same time. No existing anonymization methods can achieve all of these at the same time. The core of this technology is a differential private clustering algorithm that provides strong privacy protection, preserves data properties important for subsequent analysis, and allows the party receiving the anonymized data to conduct analysis directly on anonymized data without the need of decryption or any extra processing. The research carried out was to design, implement and verify a solution to this problem by completing the following tasks: 1) developing the core technology; 2) developing a context based method that automatically recommends fields that must be anonymized; 3) conducted experiments showing superior results using our approach compared to existing tools, and 4) developed an intuitive but basic user interface. The research that was conducted generated novel algorithmic techniques that utilize state-of-the-art methods such as condensation, differential privacy preservation, clustering, automated tuning based on contextual awareness, and recommendation techniques to specify columns to users for anonymization leading to optimal privacy that allows research analysis on the dataset. Experiments were conducted to evaluate the efficacy of these novel algorithmic techniques by performing analysis on original non-anonymized datasets, then conducting analysis on the same yet anonymized datasets and comparing the results of the analyses. Overall, the anonymized analysis results were within 1% of the original results, verifying that the generated technology not only guarantees a high level of privacy but also enables research analysis as if it were conducted on the original dataset. Potential applications of this technology include anonymization of any type of structured network datasets that contain sensitive identifiers, such as IP addresses, that can be used in multiple applications. For example, to create an AI or machine learning model for cyber security, e.g., to detect attacks, or for performance analysis, e.g., identify bottlenecks or predict performance. In addition, a market analysis that was conducted for potential applications of this technology identified a broader range of applications of our anonymization technology beyond the network sector that includes healthcare, banking, insurance, securities, finance (FISB), data brokering, cloud services, ad sales, and government.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Towards Automatically Matching Security Advisories to CPEs: String Similarity-based Vendor Matching

When a vulnerability is reported by the National Vulnerability Database (NVD), affected products are listed in the structured Common Platform Enumeration (CPE) format. Unfortunately, if the vulnerability is in a software library (e.g., Log4j), it will not include CPEs for each product containing that library. In these cases, security operators need to manually read the vendor's or third-party security advisories to see if their product is affected. However, these advisories do not report affected products in a structured format, which prevents automated processing, This paper makes the first effort towards automatically constructing structured CPEs for the vulnerable products in a non-NVD security advisory from the unstructured data in the advisory. Since this is a very challenging problem, this paper specifically focuses on the initial but key step of matching the un-structured vendor names in security advisories to the structured vendor representations in the standard CPE format. We explore the feasibility of using string similarity to solve the problem. The basic idea is to compare a vendor name from the non-NVD advisory with each vendor in the official CPE dictionary. The CPE vendor with the highest similarity score to the advisory's vendor will be considered as the match. We first conduct an experimental, comparative study of multiple mainstream string similarity metrics for this matching problem. To improve the performance, we then design a new string similarity metric that is adapted from an existing metric by weighing different tokens in the advisory's vendor name differently.

McClanahan, Kylie↗
DOE OSTI2024

Reductive Analysis with Compiler-Guided Large Language Models for Input-Centric Code Optimizations

Input-centric program optimization aims to optimize code by considering the relations between program inputs and program behaviors. Despite its promise, a long-standing barrier for its adoption is the difficulty of automatically identifying critical features of complex inputs. This paper introduces a novel technique, reductive analysis through compiler-guided Large Language Models (LLMs), to solve the problem through a synergy between compilers and LLMs. It uses a reductive approach to overcome the scalability and other limitations of LLMs in program code analysis. The solution, for the first time, automates the identification of critical input features without heavy instrumentation or profiling, cutting the time needed for input identification by 44× (or 450× for local LLMs), reduced from 9.6 hours to 13 minutes (with remote LLMs) or 77 seconds (with local LLMs) on average, making input characterization possible to be integrated into the workflow of program compilations. Optimizations on those identified input features show similar or even better results than those identified by previous profiling-based methods, leading to optimizations that yield 92.6% accuracy in selecting the appropriate adaptive OpenMP parallelization decisions, and 20-30% performance improvement of serverless computing while reducing resource usage by 50-60%.

Input-Centric Optimization↗
DOE OSTI2022

ACES: Infrastructure As Code. Model Optimization and Performance Capability

Infrastructure as Code (IaC) refers to managing infrastructure (networks, physical/virtual machines, storage, and connection topology) in a descriptive model/language, rather than configuring it manually or using interactive configuration tools. Just like source code can be compiled to generate the same binary code, IaC enables generating the same environment every time it is applied. IaC is a key DevOps practice and is generally used in conjunction with continuous integration (CI) and continuous delivery (CD). In CI, all code changes are merged into a mainline branch and validated multiple times a day as developers check in their changes to the source code. In CD on the other hand, code changes are automatically packaged for a new release-to-production on a regular basis. This typically enables teams to deliver software changes much more quickly and often. Developing and managing the ACES platform using (IaC) is vital for the robust deployment and continued sustainment of this foundational computing capability. IaC and DevOps practices will help us solve many of the common challenges often encountered in developing and maintaining compute infrastructure. First, it will make the provisioning, deployment, and maintenance of the compute infrastructure across multiple environments much more efficient. Second, these processes help make the overall system much more stable by continuously testing new changes as they are introduced to the system. Third, it allows us to be much more confident of the security controls in place since they can be tested as part of the CI process and all new changes can be audited and tracked. Finally, IaC enables the ACES Platform to be adaptable to the emerging technologies due to its ability to spin up different test beds to evaluate and incorporate these technologies. This document addresses common infrastructure-management challenges, describes what happens if they are not addressed, and highlights the value of utilizing IaC to tackle them. Finally, we will provide a high-level overview of the IaC and DevOps practices being utilized by the ACES Platform team.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Iterative Stability Enforcement in Adaptive Antoulas–Anderson Algorithms for \({\boldsymbol{\mathcal{H}_2}}\) Model Reduction

This paper presents an extension of the Adaptive-Antoulas-Anderson (AAA) algorithm for rational modelling. Specifically, our new stable multi-input multi-output AAA (smiAAA) algorithm builds rational approximations of multi-input signals with a common set of stable poles. A new methodology is presented for iteratively enforcing stability constraints on the poles. We demonstrate the strengths of this approach compared to the stability enforcement in the FastAAA algorithm. Results using the smiAAA algorithm are compared with the commonly used Vector Fitting algorithm and the more recently published RKFIT algorithm. Vector Fitting and RKFIT both require the user to input the number of poles to use in the approximations. If the final approximation is not accurate enough, the user must re-start Vector Fitting or RKFIT with a larger number of poles and/or a new starting location for the poles. In contrast, the smiAAA algorithm is designed to allow the user to simply input the desired accuracy of the approximations, and the necessary number of poles is detected automatically. This permits users to produce approximations of a desired accuracy with no knowledge about the underlying order of the system being approximated, preventing the algorithm from ever needing to be rerun. An additional feature for preventing extraneous poles from being returned by AAA is also discussed. The cause of these extraneous poles is efficiently detected and removed by our presented methodology. In conclusion, the examples presented demonstrate that smiAAA can efficiently produce approximations of similar or better accuracy than Vector Fitting and RKFIT while requiring less input from the user.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Seismicity-constrained fault detection and characterization with a multitask machine learning model

Geological fault detection and characterization are crucial for understanding subsurface dynamics across scales. While methods for fault delineation based on either seismicity location analysis or seismic image reflector discontinuity are well-established, a systematic approach that integrates both data types remains absent. We develop a novel machine learning model that unifies seismic reflector images and seismicity location information to automatically identify geological faults and characterize their geometrical properties. The model encodes a seismic image and a seismicity location image separately, and fuses the encoded features with a spatial-channel attention fusion module to improve the learning of important features in both inputs. We design an automated strategy to generate high-quality synthetic training data and labels. To improve the realism of the seismicity location image, we include random seismicity noise and missing seismicity location associated with some of the faults. We validate the model’s efficacy and accuracy using synthetic data examples and two field data examples. Moreover, we show that fine-tuning the trained model with a small, domain-specific dataset enhances its fidelity for field data applications. The results demonstrate that integrating seismicity location and seismic images into a unified framework allows the end-to-end neural network to achieve higher fidelity and accuracy in delineating subsurface faults and their geometrical properties compared with image-only fault detection methods. Our approach offers an adaptive data-driven tool for geological fault characterization and seismic hazard mitigation, bridging the gap between seismicity location and image-based fault detection methods.

58 GEOSCIENCES↗
DOE OSTI2023

WiP Abstract: Edge-Based Privacy of Naturalistic Driving Data Collection

Collecting large driving datasets is important for data-driven transportation research and studied in naturalistic driving [3]. Due to the standard implementation of a Controller Area Network (CAN) bus for a vehicle’s inter-module communication, many off-the-shelf devices can easily transform a vehicle into a rich data collection utility [2]. Vehicles with Adaptive Cruise Control (ACC) are an example of a feature resulting in emergent traffic behavior when scaled [4]. While these utilities were designed with particular data use cases, data may be publicly shared to benefit other researchers through online tools like CyVerse [1]. However, such data should only be shared when any private information is removed. This private information may exist as a set of GPS coordinates, since the start and end points of a trip may designate a driver’s place of residence or work. When considering larger continuously-collected data sets with a focus on naturalistic driving, many drivers are needed to make data collection feasible. Removal of private information becomes much more of a challenge since every driver may uniquely define their geographic privacy. This project aims to build upon the foundation of libpanda [2] by adding features of edge-based data privacy enforcement. In it’s current form, libpanda uses a GPS in conjunction with a CAN interface to record data. Libpanda feature s a set of startup and shutdown scripts to perform automatic data collection and upload. With additional support hardware on a Raspberry Pi, the Pi can maintain power on vehicle shutdown to automatically upload data before shutdown.

Bunting, Matt↗
DOE OSTI2025

Knowledge Oriented Graph Unified Transformer (KOGUT) v0.1

KOGUT — Knowledge Oriented Graph Unified Transformer KOGUT implements the Relational Graph Transformer (RelGT) architecture for knowledge graph link prediction in biological domains, with a primary focus on microbial growth media prediction. While the original RelGT (arXiv:2505.10960) targets relational tables, time series, and multi-table databases, KOGUT adapts this architecture for heterogeneous biological knowledge graphs, providing first-in-class AI predictive models for microbial cultivation. Key Adaptations Beyond Original RelGT: - Knowledge Graph Focus: Applied to biological KGs with semantic node types (taxa, chemicals, media, phenotypes, environments) versus generic relational database tables, trained on the KG-Microbe knowledge graph (1.3M entities, 2.9M edges, 24 relation types). - Multimodal Node Encoding: Integrates node labels, categories, descriptions, and synonyms from KG metadata through learned embedding layers—adapting relational column features to graph node attributes with textual semantics. - Extended K-Hop Subgraph Strategy: Optimized neighborhood sampling (3-hop default, configurable up to 200 nodes) tuned for sparse biological networks, building on the original local-global attention framework with biological relation preservation. - Biolink Predicate Preservation: Type-specific transformations for 24 biological edge semantics (occurs_in, consumes, produces, has_phenotype, subclass_of) beyond standard relational foreign keys, enabling multi-relation link prediction. - Inductive Learning Support: Enables zero-shot predictions for novel taxa through feature-based embeddings (temperature, oxygen requirements, gram stain, cell shape), extending the original transductive relational benchmark scope to uncultured microorganisms. CheapSOTA Performance Optimizations (This Distribution): - VQ-EMA Centroid Attention: Vector quantization with exponential moving average for improved global context modeling (+5-10% MRR improvement). - HDF5 Precomputed Data Loading: One-time preprocessing of k-hop subgraphs to eliminate redundant graph traversals (2-5× training speedup). - Distributed Data Parallel Training: Multi-GPU support for scaling to larger knowledge graphs (tested on 4× NVIDIA A100 GPUs at NERSC Perlmutter). - Mixed Precision Training: Automatic mixed precision (AMP) for memory efficiency and faster training. Advantages Over Standard Knowledge Graph Embedding Models: Combines RelGT's proven multi-element tokenization (features, type, hop, structure) with graph-native biological representations, enabling interpretable link prediction across heterogeneous entities that standard embedding models (TransE, RotatE, ComplEx) and table-based transformers cannot directly model. Achieves near-perfect performance on microbial growth media prediction (MRR: 0.9966, Precision@1: 0.9932, Hit@10: 1.0000) while maintaining explainability through attention-based reasoning over biological pathways. Training Data: - KG-Microbe merged knowledge graph: 1,379,337 nodes, 2,960,472 edges - 24 biological relation types including taxonomic hierarchies, metabolic interactions, phenotype associations, and environmental relationships - Primary prediction task: Growth media suitability for microbial taxa (biolink:occurs_in, 50K edges) - Multi-relation capability: Predicts links for any of the 24 relation types, including chemical consumption/production, phenotype associations, and taxonomic classification Citation: Original RelGT Architecture: Dwivedi et al., "Relational Graph Transformer", arXiv:2505.10960, 2025 KOGUT Implementation: Knowledge Oriented Graph Unified Transformer for Microbial Growth Media Prediction Developed at Lawrence Berkeley National Laboratory (LBNL) Trained on NERSC Perlmutter supercomputer

Joachimiak, Marcin [Lawrence Berkeley National Lab↗