Search NASA⌕ Search

SEARCH · Search NASA

Results for “AI 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 865 records · Page 48
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

Global 3D Data Visualization and Analysis Platform With Advanced Machine Learning Capabilities in Support of Lunar Exploration

Introduction: The science goals for NASA’s Artemis program include: a) Understanding the character and origin of lunar polar volatiles, b) Conducting experimental science in the lunar environment and c) Investigating and mitigating exploration risks [1]. The permanently shadowed regions (PSRs) on the Lunar south pole are expected to host large quantities of water-ice and volatiles that are important for sustainable Lunar exploration [2]. There are several missions such as onboard Korea Pathfinder Lunar Orbiter (KPLO: Korean name Danuri) with onboard ShadowCam camera [3], Astrobotic Peregrine Mission One [4], and other efforts underway to obtain high resolution topographic, minerals, volatiles and other information on the moon. We envision a need in immediate future for platforms to integrate these data sets, provide rendering and visualization capabilities in the context of a 3D Lunar globe for easier information access and analysis. NASA's Celestial Mapping System (CMS) [5] is developed to address the need for 3D tools for planetary science investigations, mission planning, in-situ operations, in a 3D-first design constructed around a unified view of a planetary globe. At present CMS provides many critical functionalities that include: 1) equipment planning and optimized placement on Lunar surface 2) line-of-sight (LOS) analysis 3) powerful measurement tools based on 3D terrain with realistic 3D models to represent rovers, astronauts and equipment 4) visualization of derived mapping products (e.g. resource maps), and 5) a data engine for hosting new observations that are not available in other contemporary lunar data tools [5, 7]. Planetary Data Ingestion: CMS can consume and analyze data from locally hosted and external third party sources. It is compatible with Open Geospatial Consortium (OGC) data and file standards and currently integrates datasets from the Astrogeology Science Center of USGS. This includes global and local data acquired from NASA (LRO, Clementine, Lunar Orbiter) and JAXA (SELENE/Kaguya), with capability of integrating more datasets. In addition, users can specify other WMS-hosted data endpoints, which CMS can then query and stream data from automatically. To set-up an automated process for ingestion and accurate rendering, visualization and analysis of external 3rd party planetary datasets within CMS, we initiated the process of ingesting unique dataset of super-enhanced images of the permanently shadowed regions (PSRs) at the lunar poles which were produced by the Hyper-effective nOise Removal U-net Software (HORUS) tool [8]. This tool was developed to enhance the extremely low-light images of the interior of PSRs and provide the ability to see within these regions at and discern surface features (i.e. boulders and craters) down to 3 meters in size. We focused on the Nobile region on the Lunar south pole, selected site for VIPER mission and stitched several images to create a high-resolution map within one of the PSR of Nobile crater. Figure 1 shows the dark PSR zone form the original NAC layer of LRO as the base layer (left image) and the illuminated areas within that crater (center) which was created by ingesting and merging several of HORUS generated images. At present we employ a semi-automated process to ensure spatial accuracy and merger of several overlapping zones. However, we are in the process of completely automating this process by employing AI based techniques that would rank, sort, and stack the images based on their information density. The georectification of the images would employ selected features. Analysis on Ingested Planetary Datasets: Once an external planetary data-set is successfully ingested, georectified and merged seamlessly as a data-layer; CMS’ numerous analysis tools can be used on this data. A Line of Sight (LOS) tool has been developed for CMS which analyzes terrain profiles and obstructions to determine visibility for remote observers [5,6]. Figure 1 (right image) shows the viewshed analysis on the same PSR in the Nobile region. The yellow pin shows the observer location outside the PSR. The yellow area shows the visible part of PSR. The obstructed area with no visibility for the observer is shown in red. The Measurements tool allows the user to take area and distance measurements of features on the terrain using various shapes. Measurement type can be specified in a number of ways: Line, Path, Polygon, Circle, Ellipse, Square, Rectangle or Freehand. Once the shape is specified, elevation information can then be extracted along each of these shapes. Figure 2 (left) shows the measurements performed on a crater n illuminated PSR in Nobile region. The equipment placement tool allows the user to place a 3D equipment model at a desired location and analyze its coverage area. The equipment placement tool is coupled with LOS to determine the coverage. Figure 2 (right) shows an equipment placed on the Lunar terrain and it’s coverage area. The red rays are blocked sight lines and the green rays are non-obstructed sight lines with the cyan lines showing the point of intersection with the terrain. More details are provided in the video demonstrations in Reference 5. Overcoming Polar Distortions: 3D geospatial applications exhibit significant distortions in polar imagery due to several reasons: 1) distortions in the source imagery, 2) incompatible tessellation algorithms at the poles, and 3) map projections. We are leveraging new tessellation algorithms and reprojecting data using projections that are better suited for Lunar poles. The goal is to seamlessly switch to polar projections while maintaining 3D view and navigation.

Maps↗
DOE OSTI2025

Measurements and interpretations of W ± Z production cross-sections in pp collisions at $\sqrt{s}=13$ TeV with the ATLAS detector

Measurements of integrated and differential cross-sections for W ± Z production in proton-proton collisions are presented. The data collected by the ATLAS detector at the Large Hadron Collider from 2015 to 2018 at a centre-of-mass energy of $\sqrt{s}=13$ TeV are used, corresponding to an integrated luminosity of 140 fb −1 . The W ± Z candidate events are reconstructed using leptonic decay modes of the gauge bosons into electrons or muons. The integrated cross-section per lepton flavour for the production of W ± Z is measured in the detector fiducial region with a relative precision of 4%. The measured value is compared with the Standard Model prediction at a precision of up to next-to-next-to-leading-order in QCD and next-to-leading-order in electroweak. Cross-sections for W + Z and W − Z production and their ratio are presented. The W ± Z production is also measured differentially as functions of various kinematic variables, including new observables sensitive to CP-violation effects. All measurements are compared with state-of-the-art Standard Model predictions from fixed-order calculations or Monte Carlo generators based on next-to-leading-order matrix elements interfaced with parton showers. An effective field theory interpretation of the measurements is performed, considering both CP-conserving and CP-violating dimension-6 operators modifying the W ± Z production. In the absence of observed deviations from the Standard Model, limits on CP-conserving Wilson coefficients are extracted using the transverse mass of the W ± Z system. For CP-violating coefficients a machine learning approach is used to construct an observable with enhanced sensitivity to CP-violation effects.

hadron-hadron scattering↗
DOE OSTI2025

How Does Feedback Affect the Star Formation Histories of Galaxies?

Star formation in galaxies is regulated by the interplay of a range of processes that shape the multiphase gas in the interstellar and circumgalactic media. Using the Cosmology and Astrophysics with MachinE Learning Simulations (CAMELS) suite of cosmological simulations, we study the effects of varying feedback and cosmology on the average star formation histories (SFHs) of galaxies at z ∼ 0 across the IllustrisTNG, SIMBA, and ASTRID galaxy formation models. We find that galaxy SFHs in all three models are sensitive to changes in stellar feedback, which affect the efficiency of baryon cycling and the rates at which central black holes grow, whereas the effects of varying active galactic nucleus (AGN) feedback depend on model-specific implementations of black hole seeding, accretion, and feedback. We also find strong interaction terms that couple stellar and AGN feedback, usually by regulating the amount of gas available for the central black hole to accrete. Using a double power law to describe the average SFHs, we derive a general set of equations relating the shape of the SFHs to physical quantities like baryon fraction and black hole mass across all three models. We find that a single set of equations (albeit with different coefficients) can describe the SFHs across all three CAMELS models, with cosmology dominating the SFH at early times, followed by halo accretion, and feedback and baryon cycling at late times. Galaxy SFHs provide a novel, complementary probe to constrain cosmology and feedback, and can connect the observational constraints from current and upcoming galaxy surveys with the physical mechanisms responsible for regulating galaxy growth and quenching.

Iyer, Kartheik G. [Columbia Univ., New York, NY (U↗
DOE OSTI2023

Measurement of the properties of Higgs boson production at $\sqrt{s}$ = 13 TeV in the H → γγ channel using 139 fb -1 of pp collision data with the ATLAS experiment

Measurements of Higgs boson production cross-sections are carried out in the diphoton decay channel using 139 fb -1 of pp collision data at $\sqrt{s}$ = 13 TeV collected by the ATLAS experiment at the LHC. The analysis is based on the definition of 101 distinct signal regions using machine-learning techniques. The inclusive Higgs boson signal strength in the diphoton channel is measured to be $1.04^{+0.10}_{-0.09}$. Cross-sections for gluon-gluon fusion, vector-boson fusion, associated production with a W or Z boson, and top associated production processes are reported. An upper limit of 10 times the Standard Model prediction is set for the associated production process of a Higgs boson with a single top quark, which has a unique sensitivity to the sign of the top quark Yukawa coupling. Higgs boson production is further characterized through measurements of Simplified Template Cross-Sections (STXS). In total, cross-sections of 28 STXS regions are measured. The measured STXS cross-sections are compatible with their Standard Model predictions, with a p-value of 93%. The measurements are also used to set constraints on Higgs boson coupling strengths, as well as on new interactions beyond the Standard Model in an effective field theory approach. No significant deviations from the Standard Model predictions are observed in these measurements, which provide significant sensitivity improvements compared to the previous ATLAS results.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2026

Pre-training Vision Models for the Classification of Alerts from Wide-field Time-domain Surveys

Modern wide-field time-domain surveys facilitate the study of transient, variable and moving phenomena by conducting image differencing and relaying alerts to their communities. Machine learning tools have been used on data from these surveys and their precursors for more than a decade, and convolutional neural networks (CNNs), which make predictions directly from input images, saw particularly broad adoption through the 2010s. Since then, continually rapid advances in computer vision have transformed the standard practices around using such models. It is now commonplace to use standardized architectures pre-trained on large corpora of everyday images (e.g., ImageNet). In contrast, time-domain astronomy studies still typically design custom CNN architectures and train them from scratch. Here, we explore the effects of adopting various pre-training regimens and standardized model architectures on the performance of alert classification. We find that the resulting models match or outperform a custom, specialized CNN like what is typically used for filtering alerts. Moreover, our results show that pre-training on galaxy images from Galaxy Zoo tends to yield better performance than pre-training on ImageNet or training from scratch. We observe that the design of standardized architectures are much better optimized than the custom CNN baseline, requiring significantly less time and memory for inference despite having more trainable parameters. On the eve of the Legacy Survey of Space and Time and other image-differencing surveys, these findings advocate for a paradigm shift in the creation of vision models for alerts, demonstrating that greater performance and efficiency, in time and in data, can be achieved by adopting the latest practices from the computer vision field.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Search for squarks and gluinos in pp collisions at $\sqrt{s} = 13$ TeV and 13.6 TeV in events with $\tau$-leptons, jets and missing transverse momentum using the ATLAS detector

A search for R-parity-conserving supersymmetry in events with large missing transverse momentum, jets and at least one hadronically decaying $\tau$-lepton is presented. Both gluino and squark pair production are considered, with the cascade decay of each gluino or squark producing either a $\tau$-slepton or a $\tau$-sneutrino. Three channels are examined, requiring either exactly one hadronically decaying $\tau$-lepton and no other leptons, exactly one hadronically decaying $\tau$-lepton and at least one other lepton, or two or more hadronically decaying $\tau$-leptons. Analyses in the three channels are optimised independently and combined statistically. Two separate analysis strategies, either a cut-and-count or machine-learning approach, are used. The search uses 140 and 51.8 of pp collision data recorded by the ATLAS detector at the Large Hadron Collider during 2015–2018 at TeV and 2022–2023 at TeV, respectively. Gluino masses below 2.25 TeV and squark masses up to 1.7 TeV are excluded

Aad, G. [CNRS/IN2P3] (ORCID:0000000266654934)↗
DOE OSTI2025

Atomate2: modular workflows for materials science

High-throughput density functional theory (DFT) calculations have become a vital element of computational materials science, enabling materials screening, property database generation, and training of “universal” machine learning models. While several software frameworks have emerged to support these computational efforts, new developments such as machine learned force fields have increased demands for more flexible and programmable workflow solutions. This manuscript introduces atomate2, a comprehensive evolution of our original atomate framework, designed to address existing limitations in computational materials research infrastructure. Key features include the support for multiple electronic structure packages and interoperability between them, along with generalizable workflows that can be written in an abstract form irrespective of the DFT package or machine learning force field used within them. Our hope is that atomate2's improved usability and extensibility can reduce technical barriers for high-throughput research workflows and facilitate the rapid adoption of emerging methods in computational material science.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

AI-Accelerated Design of Targeted Covalent Inhibitors for SARS-CoV-2

Direct-acting antivirals for the treatment of the COVID-19 pandemic caused by the SARS-CoV-2 virus are needed to complement vaccination efforts. Given the ongoing emergence of new variants, automated experimentation, and active learning based fast workflows for antiviral lead discovery remain critical to our ability to address the pandemic’s evolution in a timely manner. While several such pipelines have been introduced to discover candidates with noncovalent interactions with the main protease (M pro ), here we developed a closed-loop artificial intelligence pipeline to design electrophilic warhead-based covalent candidates. Here, this work introduces a deep learning-assisted automated computational workflow to introduce linkers and an electrophilic “warhead” to design covalent candidates and incorporates cutting-edge experimental techniques for validation. Using this process, promising candidates in the library were screened, and several potential hits were identified and tested experimentally using native mass spectrometry and fluorescence resonance energy transfer (FRET)-based screening assays. We identified four chloroacetamide-based covalent inhibitors of M pro with micromolar affinities (K I of 5.27 μM) using our pipeline. Experimentally resolved binding modes for each compound were determined using room-temperature X-ray crystallography, which is consistent with the predicted poses. The induced conformational changes based on molecular dynamics simulations further suggest that the dynamics may be an important factor to further improve selectivity, thereby effectively lowering KI and reducing toxicity. These results demonstrate the utility of our modular and data-driven approach for potent and selective covalent inhibitor discovery and provide a platform to apply it to other emerging targets.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

Search for pair production of higgsinos in events with two Higgs bosons and missing transverse momentum in $\sqrt{𝑠}$ = 13 TeV 𝑝⁢𝑝 collisions at the ATLAS experiment

This paper presents a search for pair production of higgsinos, the supersymmetric partners of the Higgs bosons, in scenarios with gauge-mediated supersymmetry breaking. Each higgsino is assumed to decay into a Higgs boson and a nearly massless gravitino. The search targets events where each Higgs boson decays into $b\bar{b}$, leading to a reconstructed final state with at least three energetic 𝑏-jets and missing transverse momentum. Two complementary analysis channels are used, with each channel specifically targeting either low or high values of the higgsino mass. The low-mass (high-mass) channel exploits 126 (139) fb −1 of $\sqrt{𝑠}$ = 13 TeV data collected by the ATLAS detector during Run 2 of the Large Hadron Collider. No significant excess above the Standard Model prediction is found. At 95% confidence level, masses between 130 GeV and 940 GeV are excluded for higgsinos decaying exclusively into Higgs bosons and gravitinos. Exclusion limits as a function of the higgsino decay branching ratio to a Higgs boson are also reported.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
NASA NTRSDate not supplied

Deep Learning-Based Negotiation Strategy Selection for Cooperative Conflict Resolution in Urban Air Mobility

This paper presents a collaborative conflict resolution technique using deep neural network-based intelligent search of the solution space. This approach offers a rapid convergence to a mutually acceptable solution for real-time conflict resolution, suitable for urban air mobility operations. Furthermore, the presented technique allows operational flexibility to the urban air mobility agents where these agents can collaboratively devise the solution via integrative negotiation, based on their local utility functions, as long as such a solution does not violate the global safety thresholds. The presented machine-to-machine negotiation method is built on our prior work on holistic assessment of the airspace and potential conflict detection implemented at-the-edge, onboard the unmanned aircraft systems. This paper extends the prior work to augment decision-making at-the-edge, thereby, promising a true distributed control architecture for urban air mobility. In this approach, each agent (a) builds a potential in-flight conflict map, (b) identifies the conflicting agents, (c) dynamically prepares a list of alternatives based on its current utility functions, (d) negotiates with the conflicting agents to pick one of these alternatives, and (e) implements the negotiated alternative to mutually resolve the conflict. Note that such an approach does not require a contingency plan to be made pre-flight, as the conflict resolution strategies are decided and negotiated in real time based on the present state of the agent. The contingency plan, if available, can serve as an input to the real-time conflict resolution strategy formulation, and also can be used as a fallback plan in case the negotiation fails and the impacted agents need to switch to a rule-based/supervisory resolution mode from the discussed distributed resolution mode. The presented collaborative negotiation-based conflict resolution technique incorporates a time-dependent reward function to catalyze collaborative resolution by incentivizing the agents with local and global rewards beneficial to their business operations.

Advanced Air Mobility↗
DOE OSTI2026

Explainable tokamak-agnostic forecasting of fusion plasma instability via megahertz turbulent fluctuations

Scientific applications of artificial intelligence (AI) often remain limited by device-specific training and unexplained “black-box” approaches, creating fundamental barriers to cross-system generalization. This challenge is critical for nuclear fusion, where future reactors will have limited operational data for AI training. Here, we demonstrate that our neural network, trained solely on megahertz-scale turbulence measurements from one machine (DIII-D), forecasts Type-I edge localized mode (ELM) onsets in a different tokamak (KSTAR) through zero-shot weight transfer following physics-consistent preprocessing without device-specific retraining. Through an explainable AI framework combining gradient-weighted class activation mapping with physics validation, we reveal that our network can internalize physics relationships governing the ELM instabilities rather than memorizing device-specific patterns. The network perceives spatiotemporal features that correlate consistently with independently calculated instability growth rates, magnetohydrodynamic stability limits, and pedestal structure dynamics. Statistical analyses of dimensionally-reduced saliency features reveal the identical triangular features between the saliency representations, instability growth rates, and prediction probability across tokamaks, providing evidence that our forecasting system can show tokamak-agnostic generalization. This work contributes to a foundation for explainable scientific AI systems, where cross-system developments are essential for transcending traditional domain-specific constraints.

AI↗
DOE OSTI2025

Search for new physics in final states with semivisible jets or anomalous signatures using the ATLAS detector

A search is presented for hadronic signatures of beyond the Standard Model (BSM) physics, with an emphasis on signatures of a strongly coupled hidden dark sector accessed via resonant production of a 𝑍′ mediator. The ATLAS experiment dataset collected at the Large Hadron Collider from 2015 to 2018 is used, consisting of proton-proton collisions at $\sqrt{𝑠}$ = 13 TeV and corresponding to an integrated luminosity of 140 fb −1 . The 𝑍′ mediator is considered to decay to two dark quarks, which each hadronize and decay to showers containing both dark and Standard Model particles, producing a topology of interacting and noninteracting particles within a jet known as “semivisible.” Machine learning methods are used to select these dark showers and reject the dominant background of mismeasured multijet events, including an anomaly detection approach to preserve broad sensitivity to a variety of BSM topologies. A resonance search is performed by fitting the transverse mass spectrum based on a functional form background estimation. No significant excess over the expected background is observed. Results are presented as limits on the production cross section of semivisible jet signals, parametrized by the fraction of invisible particles in the decay and the 𝑍′ mass, and by quantifying the significance of any generic Gaussian-shaped mass peak in the anomaly region.

particle dark matter↗
DOE OSTI2026

Sequence-based generative AI design of versatile tryptophan synthases

Enzymes are powerful and sustainable catalysts, but their widespread application is limited by the difficulty of identifying functional starting points for optimization, creating a major bottleneck in early- stage biocatalyst discovery. Designing libraries of such starting enzymes remains particularly challenging. Here, we use the GenSLM protein language model to generate novel β-subunit of tryptophan synthase (TrpB) enzymes that express in Escherichia coli and are both stable and catalytically active. Many generated TrpBs also display significant substrate promiscuity, outperforming their natural counterparts on non-native substrates. Some even surpass laboratory-evolved TrpBs. Comparison of the most-active and most-promiscuous generated TrpB to its closest natural homolog confirms that the enhanced versatility is absent from the natural enzyme, highlighting the creative potential of generative models. These results demonstrate that the generated TrpBs not only preserve natural structure and function but also acquire non-natural properties, establishing generative models as powerful tools for biocatalyst discovery and engineering.

biocatalysis↗
DOE OSTI2025

Mind the gap: Bridging the divide between AI aspirations and the reality of autonomous microscopy

What does materials science look like in the “Age of Artificial Intelligence?” Each material’s domain—synthesis, characterization, and modeling—has a different answer to this question, motivated by unique challenges and constraints. This work focuses on the tremendous potential of autonomous characterization within electron microscopy. We present our recent advancements in developing domain-aware, multimodal models for microscopy analysis capable of describing complex atomic systems. We then address the critical gap between the theoretical promise of autonomous microscopy and its current practical limitations, showcasing recent successes while highlighting the necessary developments to achieve robust, real-world autonomy.

2D materials↗
DOE OSTI2026

Microsecond-latency feedback at a particle accelerator by online reinforcement learning on hardware

The commissioning and operation of future large-scale scientific experiments will challenge current tuning and control methods. Reinforcement learning (RL) algorithms are a promising solution due to their ability to dynamically adapt to changing environments and consider delayed consequences. In many real-world applications, RL policies must produce actions in real time, often within microseconds to milliseconds, imposing significant constraints on system latency and computational overhead that conventional machine learning libraries are not designed to handle. To control phenomena in real time at these timescales, RL needs to be deployed on-the-edge, namely on dedicated hardware located near the system it controls, without relying on a host CPU or cloud-based inference. In this work we present the design and deployment of an experience accumulator system in a particle accelerator. In this system, deep-RL algorithms run using hardware acceleration and act within a few microseconds, enabling the use of RL for control of phenomena like beam instabilities. The training uses the collected data offline to reduce the number of operations carried out on the acceleration hardware. The proposed architecture was tested in real experimental conditions at the Karlsruhe research accelerator, a synchrotron light source, where the system was used to control artificially induced horizontal betatron oscillations in real-time, with a control loop period of just 2.7 μs. The results showed a performance comparable to the commercial feedback system available at the accelerator, demonstrating the viability and potential of this approach. Due to the self-learning and reconfiguration capability of this implementation, a seamless application to other control problems is possible. Applications range from particle accelerators to large-scale research and industrial facilities.

FPGA↗
DOE OSTI2025

Weakly supervised anomaly detection for resonant new physics in the dijet final state using proton-proton collisions at $\sqrt{s}$ = 13 TeV with the ATLAS detector

An anomaly detection search for narrow-width resonances beyond the Standard Model that decay into a pair of jets is presented. The search is based on 139 fb −1 of proton-proton collisions at $\sqrt{s}$ = 13 TeV recorded during 2015–2018 with the ATLAS detector at the Large Hadron Collider. The analysis is optimized without a particular signal model and aims to be sensitive to a broad range of new physics. It uses two different machine learning strategies to estimate the background in different signal regions. In each region, a weakly supervised classifier is trained to distinguish this background model from data. The analysis focuses on events with high transverse momentum jets reconstructed as large-radius jets. The mass and substructure of these jets are used as inputs to the classifiers. After a classifier-based selection, the distribution of the invariant mass of the two jets is used to search for potential local excesses. The model-independent results of both the anomaly detection methods show no signs of significant local excesses. In addition to model-independent results, a representative set of signal models is injected into the data, and the sensitivity of the methods to these scenarios is reported.

Aad, G. [Aix-Marseille Université] (ORCID:00000002↗
DOE OSTI2025

Deformable phrase level attention: A flexible approach for improving AI based medical coding

Objective: Improving the AI-driven automated medical encoding of clinical text plays a vital role in gathering information on the occurrence of diseases to improve population-level health. This work presents a novel attention mechanism designed to enhance text classification models and ensure appropriate classification of medical concepts in unstructured electronic health records. Materials and Methods: We developed a deformable, phrase-level attention mechanism to identify important lexical word-level and contextual phrase-level information from clinical text documents. We evaluated conventional and transformer-based deep learning models that we extended with our attention mechanism on the extraction of critical cancer information (e.g., site, subsite, laterality, histology, behavior) from 629,908 electronic pathology reports and on the automated medical encoding of 52,722 hospital discharge summaries. Results: Transformer-based models with the deformable, phrase-level attention mechanism achieved the best performance on the extraction of critical cancer information from pathology reports. Conventional- and transformer-based models show similar or better performance than their baseline counterparts on the automated medical encoding of clinical documents. Discussion: The addition of phrase-level information allowed models extended with our proposed method to outperform standard word-level attention. Our method showed favorable properties for the real-world application in terms of model robustness and phenotyping. These results indicate that our method is promising for automated data harmonization for common data models. Conclusion: This work proposes a novel deformable, phrase-level attention mechanism that enhances text classification models in the extraction of medical concepts from clinical text documents. We demonstrate strong performances on two clinical text datasets and showcase real-world deployability of our method.

Automated medical encoding↗