Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “execution model”

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 847 records · Page 47

Optimizing Tensor Contractions in CCSD(T) for Efficient Execution on GPUs

Tensor contractions are higher dimensional analogs of matrix multiplications, used in many computational contexts such as high order models in quantum chemistry, deep learning, nite element methods etc. In contrast to the wide availability of high-performance libraries for matrix multiplication on GPUs, the same is not true for tensor contractions. In this paper, we address the optimization of a set of symmetrized tensor contractions that form the computational bottleneck in the CCSD(T) coupled-cluster method in computational chemistry suites like NWChem. Some of the challenges in optimizing tensor contractions that arise in practice from the variety of dimensionalities and shapes for tensors include effective mapping of the high-dimensional iteration space to threads, choice of data buffering in shared-memory and registers, and tile sizes for multi-level tiling. Furthermore, in the case of symmetrized tensor contractions in CCSD(T), it is also a challenge to fuse contractions to reduce data movement cost by exploiting reuse of intermediate tensors. In this paper, we developed an efficient GPU implementation of the tensor contractions in CCSD(T) using shared-memory buffering, register tiling, loop fusion and register transpose. Experimental results demonstrate significant improvement over the current state-of-the-art.

Kim, Jinsung↗

Decovalex-2019 (Executive Summary)

The DECOVALEX Project is an on-going international research collaboration, established in 1992, to advance the understanding and modeling of coupled Thermal (T), Hydrological (H), Mechanical (M) and Chemical (C) processes in geological in geological systems. DECOVALEX was initially motivated by the recognition that prediction of these coupled effects is an essential part of the performance and safety assessment of geologic disposal systems for radioactive waste and spent nuclear fuel. Later it was realized that these processes also play a critical role in other subsurface engineering activities, such as subsurface CO 2 storage, enhanced geothermal systems, and unconventional oil and gas production through hydraulic fracturing. Research teams from many countries (e.g., Canada, China, Czech Republic, Finland, France, Germany, Japan, Republic of Korea, Spain, Sweden, Switzerland, Taiwan, United Kingdom, and the United States) various institutions have participated in the DECOVALEX Project over the years, providing a wide range of perspectives and solutions to these complex problems. These institutions represent radioactive waste management organizations, national research institutes, regulatory agencies, universities, as well as industry and consulting groups. The overall aim of DECOVALEX-2019 was to increase the understanding of various thermo-hydro-mechanical-chemical processes of importance for radionuclide release and transport from a repository to the biosphere and how they can be described and modelled using mathematical models. The scientific and technical objectives are: to increase the basic understanding of T-H-M-C coupled processes in fractured rocks (crystalline, sedimentary, argillaceous) and buffer materials; to investigate the predictive capabilities of different codes to field experiments and to perform verification of codes; to exchange experimental data, and improve the understanding of the constitutive behavior of crystalline and argillaceous rock masses and buffer materials; and to perform THMC calculations in a performance/safety assessment context.

58 GEOSCIENCES↗

Analyzing inference workloads for spatiotemporal modeling

Ensuring power grid resiliency, forecasting climate conditions, and optimization of transportation infrastructure are some of the many application areas where data is collected in both space and time. Spatiotemporal modeling is about modeling those patterns for forecasting future trends and carrying out critical decision-making by leveraging machine learning/deep learning. Once trained offline, field deployment of trained models for near real-time inference could be challenging because performance can vary significantly depending on the environment, available compute resources and tolerance to ambiguity in results. Users deploying spatiotemporal models for solving complex problems can benefit from analytical studies considering a plethora of system adaptations to understand the associated performance-quality trade-offs. To facilitate the co-design of next-generation hardware architectures for field deployment of trained models, it is critical to characterize the workloads of these deep learning (DL) applications during inference and assess their computational patterns at different levels of the execution stack. In this paper, we develop several variants of deep learning applications that use spatiotemporal data from dynamical systems. We study the associated computational patterns for inference workloads at different levels, considering relevant models (Long short-term Memory, Convolutional Neural Network and Spatio-Temporal Graph Convolution Network), DL frameworks (Tensorflow and PyTorch), precision (FP16, FP32, AMP, INT16 and INT8), inference runtime (ONNX and AI Template), post-training quantization (TensorRT) and platforms (Nvidia DGX A100 and Sambanova SN10 RDU). Overall, our findings indicate that although there is potential in mixed-precision models and post-training quantization for spatiotemporal modeling, extracting efficiency from contemporary GPU systems might be challenging. Instead, co-designing custom accelerators by leveraging optimized High Level Synthesis frameworks (such as SODA High-Level Synthesizer for customized FPGA/ASIC targets) can make workload-specific adjustments to enhance the efficiency.

97 MATHEMATICS AND COMPUTING↗

On measuring stimulated photon–photon scattering using multiple ultraintense lasers

Stimulated photon–photon scattering is a predicted consequence of quantum electrodynamics that has yet to be measured directly. Measuring the cross section for stimulated photon–photon scattering is the aim of a flagship experiment for NSF OPAL, a proposed laser user facility with two, 25-PW beamlines. We present optimized experimental designs for achieving this challenging and canonical measurement. A family of experimental geometries is identified that satisfies the momentum- and energy-matching conditions for two selected laser frequency options. Numerical models predict a maximum signal exceeding 1000 scattered photons per shot at the experimental conditions envisaged at NSF OPAL. Experimental requirements on collision geometry, polarization, cotiming and copointing, background suppression, and diagnostic technologies are investigated numerically. These results confirm that a beam cotiming shorter than the pulse duration and control of the copointing on a scale smaller than the shortest laser wavelength are needed to robustly scatter photons on a per-shot basis. Finally, we assess the bounds that a successful execution of this experiment may place on the mass scale of Born–Infeld nonlinear electrodynamics beyond the standard model of physics.

Beyond the Standard Model↗

Simulation toolkit for digital material characterization of large image-based microstructures

In this paper, an efficient image-based simulation toolkit for material characterization is presented, which is scalable to work from personal computers to workstations. The effective thermal conductivity, elasticity, and permeability are evaluated employing a computational homogenization framework based on the Finite Element Method (FEM). Two complementary open-source packages are presented: one developed in Python, which can convert digital images into voxel meshes (pyTomoviewer); the other developed in Julia, that can run numerical simulations to compute effective material properties (chpack). Also, a CUDA C version of chpack is provided (chfem_gpu). They were designed to deal with large multi-phase models, so strategies were devised to minimize their memory footprint, while avoiding a high toll on execution time. The voxel-based approach significantly simplifies the FEM meshes and allows efficient matrix-free implementations. In that sense, to handle large linear systems of equations, the element-by-element (EBE) technique is adopted, in conjunction with a low-memory implementation of the Preconditioned Conjugate Gradient (PCG) method. Finally, the code was thoroughly tested on an artificial geometry made of a square array of cylinders, for which analytical solutions exist, as well as on a real micro-tomographic reconstruction of FiberForm TM , a carbon preform commonly used in thermal protection systems.

36 MATERIALS SCIENCE↗

Scalable Programming Workflows for Validation of Quantum Computers

Hybrid quantum-classical workflows have become standard methods for executing variational algorithms and other quantum simulation techniques, which are key applications for noisy intermediate scale quantum (NISQ) computers. Validating these simulations is an important task which helps gauge the progress of quantum computer development, and classical simulation can serve as a tool to this end. Both exact and more scalable approximate methods with quantifiable error bounds can be used in validation tasks where the applicable metrics include the distance from a calculable ground truth, the quality of an error model fit to data, etc. Here we present a library extension that includes methods for validation of quantum simulations based on scalable hybrid workflows executable on high performance computers. We provide examples that use approximate methods based on tensor networks and stabilizer simulators to bound the error of quantum simulations on NISQ hardware.

Nguyen, Thien↗

Experimental design applications for modeling and assessing carbon dioxide sequestration in saline aquifers

This project was a computer modeling effort to couple reservoir simulation and ED/RSM using Sensitivity Analysis, Uncertainty Analysis, and Optimization Methods, to assess geologic, geochemical, geomechanical, and rock-fluid effects and factors on CO 2 injectivity, capacity, and plume migration. The project objective was to develop proxy models to simplify the highly complex coupled geochemical and geomechanical models in the utilization and storage of CO 2 in the subsurface. The goals were to investigate and prove the feasibility of the ED/RSM processes and engineering development, and bridge the gaps regarding the uncertainty and unknowns of the many geochemical and geomechanical interacting parameters in the development and operation of anthropogenic CO 2 sequestration and storage sites. The bottleneck in this workflow is the high computational effort of reactive transport simulation models and large number of input variables to optimize with ED/RSM techniques. The project was not to develop the reactive transport, geomechanical, or ED/RSM software, but was to use what was commercially and/or publically available as a proof of concept to generate proxy or surrogate models. A detailed geologic and petrographic mineral assemblage and geologic structure of the doubly plunging anticline was defined using the USDOE RMOTC formations of interest data (e.g., Lower Sundance, Crow Mountain, Alcova Limestone, and Red Peak). The assemblage of 23 minerals was primarily developed from literature data and petrophysical (well log) analysis. The assemblage and structure was input into a commercial reactive transport simulator to predict the effects of CO 2 injection and complex reactions with the reservoir rock. Significant impediments were encountered during the execution phase of the project. The only known commercial reactive transport simulator was incapable of simulating complex geochemistry modeled in this project. Significant effort and project funding was expended to determine the limitations of both the commercial simulator and the Lawrence Berkeley National Laboratory (LBNL) R&D simulator, TOUGHREACT available to the project. A simplified layer cake model approximating the volume of the RMOTC targeted reservoirs was defined with 1-3 minerals eventually modeled with limited success. Modeling reactive transport in porous media requires significant computational power. In this project, up to 24 processors were used to model a limited mineral set of 1-3 minerals. In addition, geomechanical aspects of injecting CO 2 into closed, semi-open, and open systems in various well completion methods was simulated. Enhanced Oil Recovery (EOR) as a storage method was not modeled. A robust and stable simulation dataset or base case was developed and used to create a master dataset with embedded instructions for input to the ED/RSM software. Little success was achieved toward the objective of the project using the commercial simulator or the LBNL simulator versions available during the time of this project. Several hundred realizations were run with the commercial simulator and ED/RSM software, most having convergence problems and terminating prematurely. A proxy model for full field CO 2 injection sequestration utilization and storage was not capable of being developed with software available for this project. Though the chemistry is reasonably known and understood, based on the amount of effort and huge computational time required, predicting CO 2 sequestration storage capacity in geologic formations to within the program goals of ±30% proved unsuccessful.

54 ENVIRONMENTAL SCIENCES↗

DFSynthesizer: Dataflow-based Synthesis of Spiking Neural Networks to Neuromorphic Hardware

Spiking Neural Networks (SNNs) are an emerging computation model that uses event-driven activation and bio-inspired learning algorithms. SNN-based machine learning programs are typically executed on tile-based neuromorphic hardware platforms, where each tile consists of a computation unit called a crossbar, which maps neurons and synapses of the program. However, synthesizing such programs on an off-the-shelf neuromorphic hardware is challenging. This is because of the inherent resource and latency limitations of the hardware, which impact both model performance, e.g., accuracy, and hardware performance, e.g., throughput. We propose DFSynthesizer, an end-to-end framework for synthesizing SNN-based machine learning programs to neuromorphic hardware. The proposed framework works in four steps. First, it analyzes a machine learning program and generates SNN workload using representative data. Second, it partitions the SNN workload and generates clusters that fit on crossbars of the target neuromorphic hardware. Third, it exploits the rich semantics of the Synchronous Dataflow Graph (SDFG) to represent a clustered SNN program, allowing for performance analysis in terms of key hardware constraints such as number of crossbars, dimension of each crossbar, buffer space on tiles, and tile communication bandwidth. Finally, it uses a novel scheduling algorithm to execute clusters on crossbars of the hardware, guaranteeing hardware performance. We evaluate DFSynthesizer with 10 commonly used machine learning programs. Our results demonstrate that DFSynthesizer provides a much tighter performance guarantee compared to current mapping approaches.

Computer Science↗

Deep Learning for Full Waveform Inversion of Elastic Active-Source Seismic Data to Estimate P-Wave Velocity Models

Seismic imaging methods are critical for Global Security and Energy & Homeland Security missions and activities that rely on subsurface characterization, but traditional methods remain computationally expensive and require significant labor hours and expertise to execute. Within the past few years, machine learning (ML), namely deep learning (DL), has been used to develop data-driven end-to-end full waveform inversion (FWI) methods to estimate 2D P-wave velocity (Vp) models in a fraction of the time as conventional FWI. These methods, however, are trained on simplistic acoustic wave seismic data and Vp models that are not realistic nor representative of real-world observations, leaving a large gap between the state-of-the-art and deployable, feasible, and practical DL FWI methods. Here, we generate a synthetic active-source, 3D, elastic wave seismic data set and a variety of Vp models with realistic geologic structure for training DL FWI methods. We evaluate six different methods that have performed well for acoustic DL FWI or medical imaging tasks using our more realistic dataset. We find that these six trained models do not match the performance of published acoustic end-to-end DL FWI methods, indicating more training data may be needed, physics may need to be incorporated to achieve good accuracy at the sacrifice of the end-to-end advantage, and/or novel methods need to be developed to enable end-to-end DL FWI methods to perform well for real-world seismic data.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Machine Learning-Based Model Predictive Control of Two-Time-Scale Systems

In this study, we present a general form of nonlinear two-time-scale systems, where singular perturbation analysis is used to separate the dynamics of the slow and fast subsystems. Machine learning techniques are utilized to approximate the dynamics of both subsystems. Specifically, a recurrent neural network (RNN) and a feedforward neural network (FNN) are used to predict the slow and fast state vectors, respectively. Moreover, we investigate the generalization error bounds for these machine learning models approximating the dynamics of two-time-scale systems. Next, under the assumption that the fast states are asymptotically stable, our focus shifts toward designing a Lyapunov-based model predictive control (LMPC) scheme that exclusively employs the RNN to predict the dynamics of the slow states. Additionally, we derive sufficient conditions to guarantee the closed-loop stability of the system under the sample-and-hold implementation of the controller. A nonlinear chemical process example is used to demonstrate the theory. In particular, two RNN models are constructed: one to model the full two-time-scale system and the other to predict solely the slow state vector. Both models are integrated within the LMPC scheme, and we compare their closed-loop performance while assessing the computational time required to execute the LMPC optimization problem.

97 MATHEMATICS AND COMPUTING↗

Risk Model for EM Decision Support Toolsets

The Government Office of Accountability (GAO) has published several reports identifying the need for the Department of Energy (DOE) Office of Environmental Management (EM) to address mounting costs for DoE's cleanup program. DOEEM could greatly benefit from independent decision tool-sets/models that would allow them to evaluate options and inform business decision at the enterprise level considering site-specific life cycle costs and system plans. Program goals: Develop a tool-set that enables EM to independently evaluate alternatives, assess outcomes from different contracting strategies, and inform critical decisions for the enterprise. Project goals: Adaption of a Risk Model for integration with complimentary tool-sets for project level decision making. Operational Events: Discrete event model that evaluates operational variables (e.g. capacity, throughput, maintenance constraints) and identify bottlenecks. Identifying and Bounding Project Risk: Identify risks that impact confidence in meeting goals (e.g. production). Life cycle Cost: Evaluate impacts of staffing levels, inventory, and capital investments on life cycle costs. Methods and approach: Monte Carlo Analysis is being executed to generate results: Input derives from Risk Register Data; Tied to Projected and Target Schedules; Incorporates float duration within the model. Assumes associated risk mitigation actions being completed within a timeline of five fiscal years. Metrics include: Confidence in Meeting Production Goal; Confidence in Safety Standards; Confidence in Continuous Operation; Other Metrics can be added as appropriate regarding specific site needs. The adapted risk model can be used as a stand alone decision tool or can be integrated complementary tool-sets (i.e. process and cost models) for project-specific decisions. These support tool-sets can then be integrated with others for site- and complex- level evaluations. Future work includes designing an adaptable and modular framework that would allow integration of multiple tool-sets for holistic and/or targeted evaluation of alternative strategies for decision making that could lead to risk and cost reduction across the enterprise.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Machine Learning Using a Simple Feature for Detecting Multiple Types of Events From PMU Data

This paper describes simple and efficient machine learning (ML) methods for efficiently detecting multiple types of power system events captured by PMUs scarcely placed in a large power grid. It uses a single feature from each PMU based on a rectangle area enclosing the event in a given data window. This single feature is sufficient to enable commonly used ML models to detect different types of events quickly and accurately. The feature is used by five ML models on four different data-window sizes. The results indicated a tradeoff between the execution speed and detection accuracy in variety of data-window size choices. Here, the proposed method is insensitive to most data quality issues typical for data from field PMUs, and thus it does not require major data cleansing efforts prior to feature extraction.

Big data↗

Evolvix BEST Names for semantic reproducibility across code2brain interfaces

Names in programming are vital for understanding the meaning of code and big data. We define code2brain (C2B) interfaces as maps in compilers and brains between meaning and naming syntax, which help to understand executable code. While working toward an Evolvix syntax for general‐purpose programming that makes accurate modeling easy for biologists, we observed how names affect C2B quality. To protect learning and coding investments, C2B interfaces require long‐term backward compatibility and semantic reproducibility (accurate reproduction of computational meaning from coder‐brains to reader‐brains by code alone). Semantic reproducibility is often assumed until confusing synonyms degrade modeling in biology to deciphering exercises. We highlight empirical naming priorities from diverse individuals and roles of names in different modes of computing to show how naming easily becomes impossibly difficult. We present the Evolvix BEST (Brief, Explicit, Summarizing, Technical) Names concept for reducing naming priority conflicts, test it on a real challenge by naming subfolders for the Project Organization Stabilizing Tool system, and provide naming questionnaires designed to facilitate C2B debugging by improving names used as keywords in a stabilizing programming language. Our experiences inspired us to develop Evolvix using a flipped programming language design approach with some unexpected features and BEST Names at its core.

97 MATHEMATICS AND COMPUTING↗

Initial results from wave tank test of closed-loop WEC control

This report describes the set up, execution, and some initial results from a series of wave tank tests of a model-scale wave energy converter (WEC) completed in May 2018 at the Navy's Maneuvering and Sea Keeping (MASK) basin. The purpose of these tests was to investigate the implementation and performance of a series of closed-loop WEC power take-off (PTO) controllers, intended to increase energy absorption/generation.

16 TIDAL AND WAVE POWER↗

A Single-Feature Machine Learning Method for Detecting Multiple Types of Events from PMU Data

This paper describes simple and efficient machine learning (ML) methods for efficiently detecting multiple types of power system events captured by PMUs scarcely placed in a large power grid. It uses a single feature from each PMU based on a rectangle area enclosing the event in a given data window. This single feature is sufficient to enable commonly used ML models to detect different types of events quickly and accurately. The feature is used by five ML models on four different data-window sizes. The results indicated a tradeoff between the execution speed and detection accuracy in variety of data-window size choices. The proposed method is insensitive to most data quality issues typical for data from field PMUs, and thus it does not require major data cleansing efforts prior to feature extraction.

Dokic, Tatjana↗

Fault Tolerant Communication Runtime Support for Data-Centric Programming Models

The largest supercomputers in the world today consist of hundreds of thousands of processing cores and many more other hardware components. At such scales, hardware faults are a common place necessitating the need for fault-resilient software systems. While different fault resilient models are available today, most of them focus on allowing the computational processes to survive faults. On the other hand, we have recently started investigating fault resilience techniques for data-centric programming models such as the partitioned global address space (PGAS) models. The primary difference in data-centric models is the decoupling of computation and data locality. That is, data placement is decoupled from the executing processes allowing us to view process failure (a physical node hosting a process is dead) separately from data failure (a physical node hosting data is dead). In this paper, we take a first step towards data centric fault resilience by designing and implementing a fault resilient one-sided communication runtime framework using Global Arrays and its communication system, ARMCI. The framework consists of a fault resilient process manager, low overhead and network assisted remote node fault detection module, non-data moving collective communication primitives and providing failure semantics and error codes for one-sided communication runtime systems. Our performance evaluation indicates that our framework incurs little overhead compared to state of the art designs and provides a fundamental framework of fault resiliency for PGAS models. Keywords

Vishnu, Abhinav↗

Profiling the BLAST bioinformatics application for load balancing on high-performance computing clusters

Abstract Background The Basic Local Alignment Search Tool (BLAST) is a suite of commonly used algorithms for identifying matches between biological sequences. The user supplies a database file and query file of sequences for BLAST to find identical sequences between the two. The typical millions of database and query sequences make BLAST computationally challenging but also well suited for parallelization on high-performance computing clusters. The efficacy of parallelization depends on the data partitioning, where the optimal data partitioning relies on an accurate performance model. In previous studies, a BLAST job was sped up by 27 times by partitioning the database and query among thousands of processor nodes. However, the optimality of the partitioning method was not studied. Unlike BLAST performance models proposed in the literature that usually have problem size and hardware configuration as the only variables, the execution time of a BLAST job is a function of database size, query size, and hardware capability. In this work, the nucleotide BLAST application BLASTN was profiled using three methods: shell-level profiling with the Unix “time” command, code-level profiling with the built-in “profiler” module, and system-level profiling with the Unix “gprof” program. The runtimes were measured for six node types, using six different database files and 15 query files, on a heterogeneous HPC cluster with 500+ nodes. The empirical measurement data were fitted with quadratic functions to develop performance models that were used to guide the data parallelization for BLASTN jobs. Results Profiling results showed that BLASTN contains more than 34,500 different functions, but a single function, RunMTBySplitDB, takes 99.12% of the total runtime. Among its 53 child functions, five core functions were identified to make up 92.12% of the overall BLASTN runtime. Based on the performance models, static load balancing algorithms can be applied to the BLASTN input data to minimize the runtime of the longest job on an HPC cluster. Four test cases being run on homogeneous and heterogeneous clusters were tested. Experiment results showed that the runtime can be reduced by 81% on a homogeneous cluster and by 20% on a heterogeneous cluster by re-distributing the workload. Discussion Optimal data partitioning can improve BLASTN’s overall runtime 5.4-fold in comparison with dividing the database and query into the same number of fragments. The proposed methodology can be used in the other applications in the BLAST+ suite or any other application as long as source code is available.

59 BASIC BIOLOGICAL SCIENCES↗

PRIMED for the Future: Purposing Raw Intake for Machine Learning-Enabled Detection (Final Report)

The COVID-19 pandemic demonstrated how a novel, elusive, and diffuse biological threat can engender uncertainty and misinformation, and it underscored the need for flexible analytical modalities agnostic to the identity of biological material. Yet even before the pandemic, recognition of the limitations of the current, list-based approach, which focuses on known pathogens and biotoxins, and of the importance of agent-agnostic biodetection was growing within the biosecurity community. In a 2018 report on “Biodefense in the Age of Synthetic Biology,” for example, the National Academy of Sciences stated that “an overreliance on the Select Agent List is a systematic weakness affecting many aspects of the United States’ current biodefense mitigation capability”. More recently, a group of biodefense researchers proposed the identification and adoption of “bioagent-agnostic signatures (BASs)” as a way of detecting and characterizing not only existing agents but also novel ones, an approach they believe will “enable a more flexible and resilient biodefense posture”. Indeed, the future of biodetection requires us to begin developing novel analytics that can identify anomalies and/or characteristics that indicate a potential threat, whether known or unknown, without looking for a specific signature that has been identified previously. To assess potential threats more rapidly, it is critical to develop agnostic artificial intelligence (AI)/machine learning (ML) systems that can be employed for real-time assessment of the nature and source of a perturbation. Such systems should be multi scale and multi-dimensional, integrating sensor data from a range of biological, chemical, and physical application spaces. Emerging deep learning (DL) models demonstrate exceptional promise for identification of discriminatory features within multi-dimensional datasets. DL models have the capacity to recognize and encode highly complex patterns in a wide range of input data modalities, including images, text, and biological/chemical/physical spectra. As such, they can execute a wide range of assessments and determinations that have traditionally required a human operator. The promise of advances in DL is apparent in the realm of human health and medicine. DL models have been validated for evaluating a variety of clinical threats to human health in a range of contexts, including infection and cancer, and they demonstrated improved performance in predicting stroke relative to human neurologists in some categories of data. Continuously evolving advances in AI/ML are expected to support more efficient evaluation of raw sequence, spectroscopy, and spectrometry data. For instance, recent advances and deployment of large language models (LLM) such as Generative Pre training Transformer (GPT) and Bidirectional Encoder Representations from Transformers (BERT) have already motivated application of these models for biological function prediction. As frameworks such as LLMs become larger and more complex in their representations, their capacity to serve as pre-trained models that can be fine-tuned for biological/biodetection purposes will similarly be amplified. While existing and emerging AI/ML have found broad applicability and use cases in the clinical sciences, development for environmental evaluation and biodetection has been limited. Functionalizing such capabilities for this purpose requires an understanding of the existing technical landscape and how the respective tools and algorithms are currently being employed. This landscape awareness then allows an assessment of the current practical capabilities of existing models and the anticipated requirements and development efforts that will be needed to adapt available algorithms for biodetection applications relevant to DHS. Leveraging expertise in biodetection, ML, and operational biodetection, the effort described in this report is comprised of a systematic landscape assessment (Subtask 2.1), comparative evaluation (Subtask 2.2), and formulation of a value proposition (Subtask 2.3) for the prospect of ML-enabled, agnostic biodetection from raw, or minimally-processed, datasets.

59 BASIC BIOLOGICAL SCIENCES↗