Search NASA⌕ Search

SEARCH · Search NASA

Results for “interpretable machine learning”

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 811 records · Page 45
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

A physics-informed and hierarchically regularized data-driven model for predicting fluid flow through porous media

This paper presents a new deep learning data-driven model for predicting structure dependent pore-fluid velocity fields in rock. The model is based on a Convolutional Auto-Encoder (CAE) artificial neural network capable of learning from image data generated by direct numerical simulations of fluid flow through pore-structures, such as by Lattice Boltzmann or molecular dynamics methods. The main novelty of the model in comparison to previous CAE-based data-driven approaches consists of three parts. The first is a methodology for decomposing the full-domain of the porous media into sub-regions, or “sub-domains”, in order to reduce the overall size of the CAE, batch process the sub-domains in parallel, and enable the CAE to learn local and generalizable nonlinear mappings of pore-fluid velocities. The second consists of embedding the finite difference solutions of the incompressible Navier-Stokes and continuity equations into convolutional layers prior to the CAE in order to provide the CAE with knowledge of fluid dynamics physics (PhyFlow). The third main novelty is that the training of the CAE is regularized with a hierarchical loss function that encourages the learning of fluid flow patterns (in a way similar to ranked modes in principal component analysis), ranking from most to least important. This is shown to increase the stability in learning, reduce over-fitting, and promote interpretability of the CAE neural network layers (HierCAE). The comprehensive new data-driven model, which we call the PhyFlow-HierCAE model, is shown to exhibit improved accuracy and generalizability of flow field predictions over conventional CAE models, attributable to the embedded physical knowledge and the hierarchical regularization, as well as realize orders of magnitude speed-ups in computation times as a surrogate for the direct numerical simulations. Examples of training and forward predictions on unseen pore-structures are provided and evaluated for data from Lattice Boltzmann and molecular dynamics simulations of pore-fluid flow. The model is shown to be a fast and accurate emulator (or “surrogate”) for predicting effective permeability of unseen pore-structures based on learning from relatively small direct numerical simulation datasets.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2024

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

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

composite materials↗
DOE OSTI2026

From bulk to surface: Structure and dynamics of amorphous alumina from deep potential molecular dynamics

Understanding the atomic-scale structure and dynamics of amorphous oxide surfaces is essential for interpreting their chemical reactivity, mechanical stability, and interfacial behavior, yet direct experimental characterization remains challenging. We employ Deep Potential (DP) molecular dynamics to generate large-scale, ab initio -quality models of amorphous Al 2 O 3 bulk glasses and melt-quenched free surfaces, enabling a quantitative analysis of both structure and relaxation dynamics with statistical confidence inaccessible to direct ab initio simulation. The trained DP model reproduces experimental liquid and glass structure, captures the cooling-rate dependence of the bulk glass transition, and corrects systematic biases in the polyhedral populations predicted by widely used classical force fields. At the free surface, mass density recovers to bulk values over ~10 Å, while local coordination requires a slightly wider subsurface region to fully converge. The outermost layer is oxygen-enriched, exhibits altered polyhedral connectivity with contracted Al–O bonds, and hosts a broad population of under-coordinated motifs (notably AlO 3 and OAl 2 ) whose abundances are governed by glass stability. These under-coordinated surface motifs exhibit distinct vibrational signatures and occur as locally paired Lewis acid and Brønsted base sites consistent with bond-valence compensation, yet remain spatially dispersed rather than aggregating into extended clusters. Despite this pronounced structural heterogeneity, surface relaxation and the glass-transition temperature remain comparable to their bulk counterparts, suggesting that the disordered surface is kinetically stable once formed. Together, these results establish a molecular-level picture of amorphous alumina surfaces and demonstrate the capability of machine-learned potentials to resolve structure–property relationships in disordered oxide interfaces.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Deep learning uncertainty quantification for clinical text classification

Machine learning algorithms are expected to work side-by-side with humans in decision-making pipelines. Thus, the ability of classifiers to make reliable decisions is of paramount importance. Deep neural networks (DNNs) represent the state-of-the-art models to address real-world classification. Although the strength of activation in DNNs is often correlated with the network’s confidence, in-depth analyses are needed to establish whether they are well calibrated. In this paper, we demonstrate the use of DNN-based classification tools to benefit cancer registries by automating information extraction of disease at diagnosis and at surgery from electronic text pathology reports from the US National Cancer Institute (NCI) Surveillance, Epidemiology, and End Results (SEER) population-based cancer registries. In particular, we introduce multiple methods for selective classification to achieve a target level of accuracy on multiple classification tasks while minimizing the rejection amount—that is, the number of electronic pathology reports for which the model’s predictions are unreliable. We evaluate the proposed methods by comparing our approach with the current in-house deep learning-based abstaining classifier. Overall, all the proposed selective classification methods effectively allow for achieving the targeted level of accuracy or higher in a trade-off analysis aimed to minimize the rejection rate. On in-distribution validation and holdout test data, with all the proposed methods, we achieve on all tasks the required target level of accuracy with a lower rejection rate than the deep abstaining classifier (DAC). Interpreting the results for the out-of-distribution test data is more complex; nevertheless, in this case as well, the rejection rate from the best among the proposed methods achieving 97% accuracy or higher is lower than the rejection rate based on the DAC. We show that although both approaches can flag those samples that should be manually reviewed and labeled by human annotators, the newly proposed methods retain a larger fraction and do so without retraining—thus offering a reduced computational cost compared with the in-house deep learning-based abstaining classifier.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

The suitability of differentiable, physics-informed machine learning hydrologic models for ungauged regions and climate change impact assessment

As a genre of physics-informed machine learning, differentiable process-based hydrologic models (abbreviated as δ or delta models) with regionalized deep-network-based parameterization pipelines were recently shown to provide daily streamflow prediction performance closely approaching that of state-of-the-art long short-term memory (LSTM) deep networks. Meanwhile, δ models provide a full suite of diagnostic physical variables and guaranteed mass conservation. Here, we ran experiments to test (1) their ability to extrapolate to regions far from streamflow gauges and (2) their ability to make credible predictions of long-term (decadal-scale) change trends. We evaluated the models based on daily hydrograph metrics (Nash–Sutcliffe model efficiency coefficient, etc.) and predicted decadal streamflow trends. For prediction in ungauged basins (PUB; randomly sampled ungauged basins representing spatial interpolation), δ models either approached or surpassed the performance of LSTM in daily hydrograph metrics, depending on the meteorological forcing data used. They presented a comparable trend performance to LSTM for annual mean flow and high flow but worse trends for low flow. For prediction in ungauged regions (PUR; regional holdout test representing spatial extrapolation in a highly data-sparse scenario), δ models surpassed LSTM in daily hydrograph metrics, and their advantages in mean and high flow trends became prominent. In addition, an untrained variable, evapotranspiration, retained good seasonality even for extrapolated cases. The δ models' deep-network-based parameterization pipeline produced parameter fields that maintain remarkably stable spatial patterns even in highly data-scarce scenarios, which explains their robustness. Combined with their interpretability and ability to assimilate multi-source observations, the δ models are strong candidates for regional and global-scale hydrologic simulations and climate change impact assessment.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Data-driven prediction of geometry- and toolpath sequence-dependent intra-layer process conditions variations in laser powder bed fusion

Geometrical features and toolpath sequence are two important factors that cause process condition variations, such as variations in the meltpool temperature or meltpool size, that might lead to undesired material properties in the laser powder bed fusion (LPBF) process. Due to the high dynamics and complex physics of the LPBF process, it is difficult to predict variations in process conditions with simulations alone. Advances in measurement technology and computational technologies open up new possibilities for smart manufacturing. In this paper, a data-driven method to predict intra-layer variations in the processing conditions that source from the toolpath sequence and part geometry is presented. The approach is demonstrated using two-color on-axis pyrometer measurements. Three demonstration cases are presented in which it is demonstrated (1) how the trained predictive model can be used as a filter to ease the interpretation of process variations and discover patterns related to toolpath and part geometry, and (2) how to generate predictions that can be used for feedforward control, i.e., for adjusting laser power or scanning speed along the toolpath using a meltpool temperature prediction model generated based on on-axis measurements. Results show that the developed prediction model is able to meaningfully predict process variations resulted from toolpath sequence and geometry. Predictions are aligned with the results from the related work of others and for the case of 180° laser path turnarounds in our high-speed X-ray imaging experiments. In conclusion, the potential issues related to the current maturity status of the process and measuring equipment that could in practice affect the performance of the proposed solutions are also discussed.

42 ENGINEERING↗
DOE OSTI2025

Coincident learning for beam-based rf station fault identification using phase information at the SLAC linac coherent light source

Anomalies in radio-frequency (rf) stations can result in unplanned downtime and performance degradation in linear accelerators such as SLAC’s Linac Coherent Light Source (LCLS). Detecting these anomalies is challenging due to the complexity of accelerator systems, high data volume, and scarcity of labeled fault data. Prior work identified faults using beam-based detection, combining rf amplitude and beam position monitor data. Due to the simplicity of the rf amplitude data, classical methods are sufficient to identify faults, but the recall is constrained by the low-frequency and asynchronous characteristics of the data. In this work, we leverage high-frequency, time-synchronous rf phase data to enhance anomaly detection in the LCLS accelerator. Due to the complexity of phase data, classical methods fail, and we instead train deep neural networks within the Coincident Anomaly Detection (CoAD) framework. We find that applying CoAD to phase data detects nearly 3 times as many anomalies as when applied to amplitude data, while achieving broader coverage across rf stations. Furthermore, the rich structure of phase data enables us to cluster anomalies into distinct physical categories. Through the integration of auxiliary system status bits, we link clusters to specific fault signatures, providing additional granularity for uncovering the root cause of faults. We also investigate interpretability via Shapley values, confirming that the learned models focus on the most informative regions of the data and providing insight for cases where the model makes mistakes. This work demonstrates that phase-based anomaly detection for rf stations improves both diagnostic coverage and root cause analysis in accelerator systems and that deep neural networks are essential for effective analysis.

Accelerator Physics (physics.acc-ph)↗
DOE OSTI2025

Using the optimal combined index weight ratio to improve the probability of anomaly detection in big area additive manufacturing

Big Area Additive Manufacturing (BAAM) of composites requires significant time, energy, and material, so it is critical to reduce production inefficiencies to make functional parts without multiple iterations. Statistical process control coupled with Principal Component Analysis (PCA) is a powerful technique that provides a quick, computationally inexpensive, and intuitive way for operators to detect defects that form in a manufacturing process without massive datasets. Recently, a combined index that is a weighted sum of the Hotelling's T 2 and squared residual error statistics has been proposed that can be monitored in one chart, improving interpretation accuracy and simplicity. However, the literature does not offer a formal method to optimise the weights. Here, we introduce two new approaches to the traditional weight selection approach using simulated and BAAM image data. Approach 1 uses a theoretically motivated optimum inspired by probabilistic principal component analysis. Approach 2 systematically varies the ratio of the weights to find the optimum. We show that approach 1 delivers optimal anomaly detection performance in select cases while approach 2 fares better in practice. Surprisingly, we also show that choosing a more complex PCA model has a minimal negative impact on anomaly detection performance compared to a more simplistic model.

3-dimensional printing↗
DOE OSTI2024

Jet classification using high-level features from anatomy of top jets

Recent advancements in deep learning models have significantly enhanced jet classification performance by analyzing low-level features (LLFs). However, this approach often leads to less interpretable models, emphasizing the need to understand the decision-making process and to identify the high-level features (HLFs) crucial for explaining jet classification. To address this, we consider the top jet tagging problems and introduce an analysis model (AM) that analyzes selected HLFs designed to capture important features of top jets. Our AM mainly consists of the following three modules: a relation network analyzing two-point energy correlations, mathematical morphology and Minkowski functionals for generalizing jet constituent multiplicities, and a recursive neural network analyzing subjet constituent multiplicity to enhance sensitivity to subjet color charges. We demonstrate that our AM achieves performance comparable to the Particle Transformer (ParT) while requiring fewer computational resources in a comparison of top jet tagging using jets simulated at the hadronic calorimeter angular resolution scale. Furthermore, as a more constrained architecture than ParT, the AM exhibits smaller training uncertainties because of the bias-variance tradeoff. We also compare the information content of AM and ParT by decorrelating the features already learned by AM. Lastly, we briefly comment on the results of AM with finer angular resolution inputs.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2023

Statistical Learning for Nonlinear Model Reduction from Local Simulations of Stochastic and Particle- and Agent-Based Systems

Stochastic physical systems across the sciences that have very high-dimensional state spaces, with a large number of fast degrees of freedom that force direct simulators to proceed by integration steps that are orders of magnitude smaller than events of interests (e.g., particle collisions). Examples range from molecular motion to dynamics of large populations of cells. A grand challenge in the simulation and understanding of such systems is the systematic construction of accurate, interpretable, reduced models, enabling faster simulations, revealing fundamental properties of the dynamics, and predicting phenomena of interest that the original simulator could not reached with sufficient accuracy or within a given computational budget. In this projected we developed novel statistical estimation/machine learning techniques for analyzing and building empirical reduced models for important families of high-dimensional stochastic systems, in particular: - we developed techniques for estimating interaction kernels in interacting particle- and agent-based systems, which are ubiquitous in Physics, Biology and many other sciences, given observed trajectories of the system; - we developed techniques for nonlinear model reduction for high-dimensional stochastic systems that have a small number of unknown, nonlinear slow variables, and a large number of fast modes, that are possibly of large magnitude, given observed short trajectories of the system in the form of bursts of trajectories from different initial conditions; - we developed novel techniques for estimating linear dynamical systems on graphs when both the dynamics and the underlying graph are unknown, and we have a sparse set of space-time observations; - we considered the problem of estimating an unknown nonlinear observation function of a standard process (e.g. Brownian motion), so that we can recognized if an observed dynamics is "just" a nonlinear version of a known dynamics; we also developed benchmarks for learning algorithms aimed at learning and classifying diffusion processes.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

Nonnegative tensor factorization for contaminant source identification

Unsupervised Machine Learning (ML) is becoming increasingly popular for solving various types of data analytics problems including feature extraction, blind source separation, exploratory analyses, model diagnostics, etc. In this work, we have developed a new unsupervised ML method based on Nonnegative Tensor Factorization (NTF) for identification of the original groundwater types (including contaminant sources) present in geochemical mixtures observed in an aquifer. Frequently, groundwater types with different geochemical signatures are related to different background and/or contamination sources. The characterization of groundwater mixing processes is a challenging but very important task critical for any environmental management project aiming to characterize the fate and transport of contaminants in the subsurface and perform contaminant remediation. This task typically requires solving complex inverse models representing groundwater flow and geochemical transport in the aquifer, where the inverse analysis accounts for available site data. Usually, the model is calibrated against the available data characterizing the spatial and temporal distribution of the observed geochemical types. Numerous different geochemical constituents and processes may need to be simulated in these models which further complicates the analyses. Additionally, the application of inverse methods may introduce biases in the analyses through the assumptions made in the model development process. Here, we substitute the model inversion with unsupervised ML analysis. The ML analysis does not make any assumptions about underlying physical and geochemical processes occurring in the aquifer. Our ML methodology, called NTFk, is capable of identifying (1) the unknown number of groundwater types (contaminant sources) present in the aquifer, (2) the original geochemical concentrations (signatures) of these groundwater types and (3) spatial and temporal dynamics in the mixing of these groundwater types. These results are obtained only from the measured geochemical data without any additional site information. In general, the NTFk methodology allows for interpretation of large high-dimensional datasets representing diverse spatial and temporal components such as state variables and velocities. NTFk has been tested on synthetic and real-world site three-dimensional datasets. Finally, the NTFk algorithm is designed to work with geochemical data represented in the form of concentrations, ratios (of two constituents; for example, isotope ratios), and delta notations (standard normalized stable isotope ratios).

58 GEOSCIENCES↗
DOE OSTI2022

Advances in Machine and Deep Learning for Modeling and Real-time Detection of Multi-Messenger Sources

We live in momentous times. The science community is empowered with an arsenal of cosmic messengers to study the universe in unprecedented detail. Gravitational waves, electromagnetic waves, neutrinos, and cosmic rays cover a wide range of wavelengths and timescales. Combining and processing these datasets that vary in volume, speed, and dimensionality requires new modes of instrument coordination, funding, and international collaboration with a specialized human and technological infrastructure. In tandem with the advent of large-scale scientific facilities, the last decade has experienced an unprecedented transformation in computing and signal-processing algorithms. The combination of graphics processing units, deep learning, and the availability of open source, high-quality datasets has powered the rise of artificial intelligence. This digital revolution now powers a multibillion dollar industry, with far-reaching implications in technology and society. In this chapter, we describe pioneering efforts to adapt artificial intelligence algorithms to address computational grand challenges in multi-messenger astrophysics. We review the rapid evolution of these disruptive algorithms, from the first class of algorithms introduced in early 2017 to the sophisticated algorithms that now incorporate domain expertise in their architectural design and optimization schemes. We discuss the importance of scientific visualization and extreme-scale computing in reducing time-to-insight and obtaining new knowledge from the interplay between models and data.

Artificial Intelligence↗
DOE OSTI2024

Designing workflows for materials characterization

Experimental science is enabled by the combination of synthesis, imaging, and functional characterization organized into evolving discovery loop. Synthesis of new material is typically followed by a set of characterization steps aiming to provide feedback for optimization or discover fundamental mechanisms. However, the sequence of synthesis and characterization methods and their interpretation, or research workflow, has traditionally been driven by human intuition and is highly domain specific. Here, we explore concepts of scientific workflows that emerge at the interface between theory, characterization, and imaging. In this study, we discuss the criteria by which these workflows can be constructed for special cases of multiresolution structural imaging and functional characterization, as a part of more general material synthesis workflows. Some considerations for theory–experiment workflows are provided. We further pose that the emergence of user facilities and cloud labs disrupts the classical progression from ideation, orchestration, and execution stages of workflow development. To accelerate this transition, we propose the framework for workflow design, including universal hyperlanguages describing laboratory operation, ontological domain matching, reward functions and their integration between domains, and policy development for workflow optimization. These tools will enable knowledge-based workflow optimization; enable lateral instrumental networks, sequential and parallel orchestration of characterization between dissimilar facilities; and empower distributed research.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Perspectives for self-driving labs in synthetic biology

Self-driving labs (SDLs) combine fully automated experiments and data collection with artificial intelligence (AI) and control algorithms that decide not only the set of parameters for the next experiment, but also potentially which scientific hypotheses to test. Taken to their ultimate expression, SDLs could usher a new paradigm of scientific research, where the world is probed, interpreted, and explained by machines for human benefit. Whereas there are functioning SDLs in the fields of chemistry and materials science, we contend that synthetic biology provides a unique opportunity since the genome provides a single, easily accessible, target for affecting the incredibly wide repertoire of biological cell behavior. Since they can provide large amounts of high-quality data, SDLs can be a platform for AI to develop approaches to systematically convert data into scientific knowledge systems. These knowledge systems can be used both to understand the biological world and to design bioengineered systems to fit a desired specification (inverse design). However, the level of investment required for the creation of biological SDLs is only warranted if directed towards solving difficult and enabling biological questions. Here, we discuss challenges and opportunities in creating SDLs for synthetic biology.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Artificial Intelligence-Assisted Daytime Video Monitoring for Bird, Insect, and Other Wildlife Interactions with Photovoltaic Solar Energy Facilities

Studying bird, insect, and other wildlife interactions with photovoltaic (PV) solar energy facilities is difficult due to limited multi-season, multi-site data. Researchers can address such data gaps by combining passive monitoring and artificial intelligence (AI). As a part of the development of AI-enabled avian–solar monitoring software, we collected over 19,000 h of daytime videos at five PV sites across three U.S. regions between 2019 and 2024. We applied a moving object detection and tracking (MODT Version 1) AI model we developed earlier to 4373 h of the footage to extract moving objects in video frames, and human reviewers interpreted the model output and identified 68,646 bird, 25,968 insect, and 169 other wildlife instances to generate the training/validation dataset. We analyzed the data by site, region, and season, considering ground cover and landscapes. Songbirds were most common, with raptors as the next most frequent group. Most notably, no bird collisions were confirmed in our observations collected from the videos. Birds most often flew over or near panels, with the highest observations in the Midwest and Northeast (approximately 30 observations per hour on average) and fewer in the desert Southwest. Other behaviors included perching, foraging, and nesting. Bird abundance peaked during breeding and migration seasons. AI-assisted video monitoring proved effective for non-invasively studying flying wildlife at solar facilities to inform ecologically mindful energy development.

avian mortality↗
DOE OSTI2023

High-dimensional data analytics in civil engineering: A review on matrix and tensor decomposition

Recent developments in sensing and monitoring techniques have led to the generation of high-dimensional data in the field of civil engineering. High-dimensional data analytics methods have thus been developed to interpret such complex data. Among the different high-dimensional data analytics techniques, matrix and tensor decomposition methods have acquired a notable interest in the civil engineering community over the past decade. Due to their unique ability to deal with highly redundant and correlated data, these methods are establishing themselves as promising and efficient tools to analyze high-dimensional data in the civil engineering arena. In this paper, high-dimensional data is referred to as a data set in which the number of features is comparable or larger than the number of observations. This review paper aims to summarize the applications of matrix and tensor decomposition methods in civil engineering over the last decade. The survey begins with a general overview of matrix and tensor decomposition followed by highlighting their significance in the field. Afterward, various applications of these high-dimensional data analytics methods in civil engineering are presented, while the advantages offered by these methods are discussed. Lastly, challenges and potential research avenues for employing matrix and tensor decomposition and future emerging trends for their novel use are highlighted.

42 ENGINEERING↗
NASA NTRS2009

Reliability of Maximal Strength Testing in Novice Weightlifters

The one repetition maximum (1RM) is a criterion measure of muscle strength. However, the reliability of 1RM testing in novice subjects has received little attention. Understanding this information is crucial to accurately interpret changes in muscle strength. To evaluate the test-retest reliability of a squat (SQ), heel raise (HR), and deadlift (DL) 1RM in novice subjects. Twenty healthy males (31 plus or minus 5 y, 179.1 plus or minus 6.1 cm, 81.4 plus or minus 10.6 kg) with no weight training experience in the previous six months participated in four 1RM testing sessions, with each session separated by 5-7 days. SQ and HR 1RM were conducted using a smith machine; DL 1RM was assessed using free weights. Session 1 was considered a familiarization and was not included in the statistical analyses. Repeated measures analysis of variance with Tukey fs post-hoc tests were used to detect between-session differences in 1RM (p.0.05). Test-retest reliability was evaluated by intraclass correlation coefficients (ICC). During Session 2, the SQ and DL 1RM (SQ: 90.2 }4.3, DL: 75.9 }3.3 kg) were less than Session 3 (SQ: 95.3 }4.1, DL: 81.5 plus or minus 3.5 kg) and Session 4 (SQ: 96.6 }4.0, DL: 82.4 }3.9 kg), but there were no differences between Session 3 and Session 4. HR 1RM measured during Session 2 (150.1 }3.7 kg) and Session 3 (152.5 }3.9 kg) were not different from one another, but both were less than Session 4 (157.5 }3.8 kg). The reliability (ICC) of 1RM measures for Sessions 2-4 were 0.88, 0.83, and 0.87, for SQ, HR, and DL, respectively. When considering only Sessions 3 and 4, the reliability was 0.93, 0.91, and 0.86 for SQ, HR, and DL, respectively. One familiarization session and 2 test sessions (for SQ and DL) were required to obtain excellent reliability (ICC greater than or equal to 0.90) in 1RM values with novice subjects. We were unable to attain this level of reliability following 3 HR testing sessions therefore additional sessions may be required to obtain an ICC of greater than or equal to 0.90. Future resistive exercise studies should consider the reliability of specific measures to ensure that changes in strength with training are attributable to training and not learning effects associated with 1RM testing.

Loehr, James A.↗
DOE OSTI2022

MetaText: Compositional Generalization in Deep Language Models

Compositional generalization, the ability to infer and reason using novel combinations of previously encountered entities and structures, is a trait of great utility across a variety of deep learning tasks; it implies a certain ability to reason using consistent and therefore interpretable rules, to understand the constituent parts of its input at multiple levels of granularity, to be robust to spurious differences between semantically equivalent inputs, and more. We seek to understand the degree to which existing natural language models achieve or fall short of compositional generalization across a variety of tasks, whether there are distinct types of compositional generalization in practice, and to identify avenues of intervention through which we can improve models’ compositional generalization ability. After exploring these questions using an array of models, tasks, and potential interventions we see that large, pretrained language models have encountered sufficient training data to account for a variety of fine-grained compositional behavior but still struggle to reason at the level of phrases or larger language structures. Non-intrusive, data-based interventions in the form of augmenting individual sequences with compressed versions of themselves or deriving new examples from induced grammars prove insufficient to encourage greater levels of compositional reasoning, indicating that future work might benefit most from focusing on changes to a model’s inductive bias at the architectural or loss level, or by integrating compositionality-boosting data interventions into the large-scale pretraining process itself.

97 MATHEMATICS AND COMPUTING↗