pith. sign in

arxiv: 2605.08396 · v1 · submitted 2026-05-08 · 💻 cs.CV

Delivering Science as a Service: Sci-Orchestra's Cloud-Native Approach to HPC

Harinarayan Krishnan , Shubhabrata Mukerjee , Jeffrey Donatelli , Daniela Ushizima This is my paper

Pith reviewed 2026-05-12 01:15 UTC · model grok-4.3

classification 💻 cs.CV
keywords Sci-OrchestraHPC orchestrationcloud-native computingKubernetes deploymentscientific workflowsAPI-driven interfaceautonomous marketplaceblack-box interoperability
0
0 comments X

The pith

Sci-Orchestra uses an API-driven interface over Kubernetes to automate HPC authentication, resources, and deployments so researchers can focus on science.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents Sci-Orchestra as a layered orchestration framework that automates experimental workflows in high-performance computing environments. By taking responsibility for secure authentication, resource management, and scalable deployment, the system frees scientists from infrastructure tasks. Its autonomous marketplace lets users select and share specialized services through simple selections without complex setups. The design also supports industry partnerships by allowing black-box use of proprietary tools that protects intellectual property while fitting into broader pipelines. If the approach works, it would accelerate the shift from lab prototypes to larger applications by lowering technical barriers to collaboration.

Core claim

Sci-Orchestra is a layered orchestration framework that fully automates experimental workflows by abstracting execution through an API-driven interface. The interface assumes responsibility for secure authentication, resource management, and scalable deployment across diverse high-performance computing environments using Kubernetes architectures. A central feature is its autonomous marketplace, which enables cross-institutional collaboration by letting researchers rapidly deploy and share specialized services via simple selections, while providing black-box interoperability that allows external collaborators to test proprietary tools without source-code exchange.

What carries the argument

The API-driven orchestration layer of Sci-Orchestra that manages authentication, resources, and Kubernetes-based deployment while operating an autonomous marketplace for service sharing.

If this is right

  • Researchers spend less time on infrastructure management and container deployments.
  • Specialized services become deployable through simple selections without complex installations.
  • Cross-institutional collaboration increases because services can be shared easily via the marketplace.
  • Industry partners can test and validate tools in secure environments without exchanging source code.
  • Workflows scale across different HPC setups while maintaining secure execution.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Smaller labs could gain access to advanced computing resources without needing dedicated IT staff.
  • The black-box model might encourage wider sharing of computational methods while still preserving competitive advantages.
  • Scientific software development could shift toward service-based distribution that emphasizes reproducibility over full code release.

Load-bearing premise

The framework can reliably deliver autonomous marketplace functionality and black-box interoperability that protects intellectual property while enabling seamless pipeline integration without source-code exchange or manual infrastructure work.

What would settle it

A test case in which deploying or sharing a service still requires manual Kubernetes configuration or source-code access would show that the automation and protection claims do not hold.

Figures

Figures reproduced from arXiv: 2605.08396 by Daniela Ushizima, Harinarayan Krishnan, Jeffrey Donatelli, Shubhabrata Mukerjee.

Figure 1
Figure 1. Figure 1: Sci-orchestra layered architecture and user interaction model enable seamless hardware and software [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: A detailed architectural diagram of the Sci-Orchestra Model Ecosystem, illustrating how user requests (simple [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Sci-Orchestra Kubernetes Orchestration and Service Delivery Scalability: architectural schematic illustrates [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: A new and seamless way of demonstrating software tools in HPC or the cloud: (a) login at NERSC, (b) [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Automated detection using Zenesis-web for zero-shot semantic and object segmentation executed via Sci [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
read the original abstract

The increasing complexity of modern computational environments often burdens researchers with infrastructure management, authentication protocols, and container deployments. We present Sci-Orchestra, a layered orchestration framework designed to fully automate experimental workflows, allowing scientists to prioritize scientific discovery over backend operations. By abstracting execution through an API-driven interface, the system assumes responsibility for secure authentication, resource management, and scalable deployment across diverse high-performance computing environments using Kubernetes architectures. A key innovation of Sci-Orchestra is its autonomous marketplace, which serves as a catalyst for cross-institutional collaboration. Through an intuitive user interface, researchers can rapidly deploy and share specialized services via simple selections, eliminating the need for complex installations and technical setups. This modular infrastructure is specifically designed to facilitate industry partnerships as it provides a secure execution environment and allows external collaborators to test and validate proprietary tools without the need for source-code exchange. This ``black-box'' interoperability protects intellectual property while enabling seamless integration into broader scientific pipelines, ultimately accelerating the transition from laboratory prototypes to industrial-scale applications.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript presents Sci-Orchestra, a layered cloud-native orchestration framework built on Kubernetes that abstracts HPC execution through an API-driven interface. It claims to automate experimental workflows by assuming responsibility for secure authentication, resource management, and scalable deployment across heterogeneous environments, while introducing an autonomous marketplace that enables rapid deployment and sharing of specialized services. A central feature is black-box interoperability that allows external collaborators to test proprietary tools without source-code exchange, thereby protecting intellectual property and facilitating cross-institutional and industry partnerships.

Significance. If realized with the described capabilities, Sci-Orchestra could meaningfully reduce the infrastructure overhead on researchers and accelerate the transition of laboratory prototypes to production-scale applications by providing secure, modular access to HPC resources. The marketplace concept addresses a genuine need for simplified collaboration and IP-safe integration, but the absence of any implementation details, performance data, or validation experiments means the practical significance remains speculative at this stage.

major comments (2)
  1. [Abstract] Abstract: The central claim that the autonomous marketplace delivers secure 'black-box' interoperability protecting intellectual property without source-code exchange rests on an API-driven Kubernetes layer, yet the manuscript supplies no description of the required mechanisms (container isolation model, API abstraction boundaries for proprietary binaries, marketplace autonomy logic, or leakage-prevention controls). This omission is load-bearing for the assertion that the system can reliably assume responsibility for authentication, resource management, and cross-institutional collaboration.
  2. [Abstract] Abstract and system description: No implementation details, performance metrics, validation experiments, or error analysis are provided to support the benefits of automation or the marketplace functionality, leaving the soundness of the architecture unverified and preventing evaluation of whether the framework achieves its stated goals.
minor comments (1)
  1. The manuscript would benefit from explicit references to related work in Kubernetes-based workflow orchestration and HPC-as-a-service platforms to better situate the novelty of the proposed marketplace and abstraction layer.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive report. The comments correctly identify that the submitted manuscript provides only a high-level architectural overview. We will revise the paper to supply the missing technical descriptions and empirical evidence while preserving the original scope and claims.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim that the autonomous marketplace delivers secure 'black-box' interoperability protecting intellectual property without source-code exchange rests on an API-driven Kubernetes layer, yet the manuscript supplies no description of the required mechanisms (container isolation model, API abstraction boundaries for proprietary binaries, marketplace autonomy logic, or leakage-prevention controls). This omission is load-bearing for the assertion that the system can reliably assume responsibility for authentication, resource management, and cross-institutional collaboration.

    Authors: We agree that the current text does not elaborate the underlying mechanisms. In the revised manuscript we will add a new subsection under System Architecture that specifies: (1) the container isolation model based on Kubernetes namespaces, pod security policies, and seccomp profiles; (2) the API abstraction layer that wraps proprietary binaries inside signed, read-only container images with gRPC endpoints; (3) the marketplace autonomy logic implemented via a service registry and policy engine that handles discovery, deployment, and revocation without exposing source code; and (4) leakage-prevention controls including encrypted inter-pod communication, mandatory access control, and audit logging. These additions will directly support the claims about secure black-box operation. revision: yes

  2. Referee: [Abstract] Abstract and system description: No implementation details, performance metrics, validation experiments, or error analysis are provided to support the benefits of automation or the marketplace functionality, leaving the soundness of the architecture unverified and preventing evaluation of whether the framework achieves its stated goals.

    Authors: The submitted version is intentionally concise and focuses on the design rationale. We accept that this leaves the practical claims unverified. The revised manuscript will include: (a) concrete implementation details of the Kubernetes-based deployment (Helm charts, custom operators, and CI/CD pipelines), (b) performance metrics collected from our internal testbed (deployment latency, resource overhead, and scaling behavior across CPU/GPU nodes), (c) validation experiments demonstrating end-to-end workflow automation and marketplace service sharing, and (d) error analysis from observed failure modes and recovery mechanisms. These results will be presented with figures and tables to allow independent assessment. revision: yes

Circularity Check

0 steps flagged

No circularity: purely descriptive architecture proposal

full rationale

The paper is a descriptive account of a proposed Kubernetes-based orchestration system and its marketplace feature. It contains no equations, no fitted parameters, no derivation chains, and no self-citations that bear load on any claim. All statements are forward-looking architectural assertions rather than reductions of outputs to inputs by construction. The reader's assessment of zero circularity is confirmed; the skeptic's concerns address missing implementation details and validation, which fall under evidentiary weakness rather than circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 2 invented entities

The central claim rests on the functionality of a newly described software system and assumptions about underlying cloud technologies; no free parameters or mathematical axioms are involved.

axioms (1)
  • domain assumption Kubernetes architectures support secure authentication, resource management, and scalable deployment across diverse HPC environments
    Invoked in the abstract as the basis for the system's backend operations.
invented entities (2)
  • Sci-Orchestra no independent evidence
    purpose: Layered orchestration framework to automate workflows and provide API-driven access
    New system presented as the core contribution of the paper.
  • autonomous marketplace no independent evidence
    purpose: Catalyst for cross-institutional collaboration and service sharing
    Described as a key innovation enabling simple selections and black-box interoperability.

pith-pipeline@v0.9.0 · 5485 in / 1252 out tokens · 80042 ms · 2026-05-12T01:15:00.767995+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

40 extracted references · 40 canonical work pages

  1. [1]

    Towards a bes light source wide event-triggered tomography data analysis pipeline using a sustainable software stack.Microscopy and Microanalysis, 26(S2):3092–3094, Aug

    Hari Krishnan, Tom Caswell, and Dan Allan. Towards a bes light source wide event-triggered tomography data analysis pipeline using a sustainable software stack.Microscopy and Microanalysis, 26(S2):3092–3094, Aug

    work page

  2. [2]

    doi:10.1017/s143192762002379x

    ISSN 1431-9276. doi:10.1017/s143192762002379x

    work page doi:10.1017/s143192762002379x

  3. [3]

    pycbir: Content-based image retrieval, 2022

    Daniela Ushizima, Flavio Araujo, and Romuere Silva. pycbir: Content-based image retrieval, 2022. URL https://github.com/lbl-camera/pycbir0.1.1. Software repository

    work page 2022

  4. [4]

    Holman, Harinarayan Krishnan, Derek R

    Elizabeth A. Holman, Harinarayan Krishnan, Derek R. Holman, Hoi-Ying N. Holman, and Paul W. Sternberg. Toward implementing autonomous adaptive data acquisition for scanning hyperspectral imaging of biological systems.Applied Physics Reviews, 10(1):011319, 03 2023. ISSN 1931-9401. doi:10.1063/5.0123278. URL https://doi.org/10.1063/5.0123278

    work page doi:10.1063/5.0123278 2023

  5. [5]

    tsuchinoko

    Ronald Pandolfi and Marcus Noack. tsuchinoko. https://github.com/lbl-camera/tsuchinoko, 2026. Center for Advanced Mathematics for Energy Research Applications (CAMERA) Software repository. Accessed: 2026-05-05

    work page 2026

  6. [6]

    Flávio H. D. Araújo, Romuere Rôdrigues Veloso e Silva, Fátima N. S. de Medeiros, Dula Parkinson, Alexander Hexemer, Cláudia M. Carneiro, and Daniela Mayumi Ushizima. Reverse image search for scientific data within and beyond the visible spectrum.Expert Syst. Appl., 109:35–48, 2018. URL https://doi.org/10.1016/j. eswa.2018.05.015

    work page doi:10.1016/j 2018

  7. [7]

    Rhizonet segments plant roots to assess biomass and growth for enabling self-driving labs.Scientific Reports, 14(1):12907, 2024

    Zineb Sordo, Peter Andeer, James Sethian, Trent Northen, and Daniela Ushizima. Rhizonet segments plant roots to assess biomass and growth for enabling self-driving labs.Scientific Reports, 14(1):12907, 2024. doi:10.1038/s41598-024-63497-8. URLhttps://doi.org/10.1038/s41598-024-63497-8

    work page doi:10.1038/s41598-024-63497-8 2024

  8. [8]

    Mukherjee, J

    S. Mukherjee, J. Lang, O. Kwon, I. Zenyuk, V . Brogden, A. Weber, and D. Ushizima. Foundation models for zero-shot segmentation of scientific images without AI-ready data. InProceedings of the 54th International Conference on Parallel Processing (ICPP-DRAI), 2025. doi:10.1145/3750720.3757283. URL https://doi. org/10.1145/3750720.3757283

    work page doi:10.1145/3750720.3757283 2025

  9. [9]

    Benchmarking topic models on scientific articles using berteley.Natural Language Processing Journal, 6:100044, 2024

    Eric Chagnon, Ronald Pandolfi, Jeffrey Donatelli, and Daniela Ushizima. Benchmarking topic models on scientific articles using berteley.Natural Language Processing Journal, 6:100044, 2024. ISSN 2949-7191. doi:https://doi.org/10.1016/j.nlp.2023.100044. URL https://www.sciencedirect.com/science/article/ pii/S2949719123000419

    work page doi:10.1016/j.nlp.2023.100044 2024

  10. [10]

    Nextflow enables reproducible computational workflows.Nature biotechnology, 35(4):316–319, 2017

    Paolo Di Tommaso, Maria Chatzou, Evan W Floden, Pablo Prieto Barja, Emilio Palumbo, and Cedric Notredame. Nextflow enables reproducible computational workflows.Nature biotechnology, 35(4):316–319, 2017

    work page 2017

  11. [11]

    Snakemake—a scalable bioinformatics workflow engine.Bioinformatics, 28(19):2520–2522, 2012

    Johannes Köster and Sven Rahmann. Snakemake—a scalable bioinformatics workflow engine.Bioinformatics, 28(19):2520–2522, 2012

    work page 2012

  12. [12]

    Pegasus, a workflow management system for science automation

    Ewa Deelman, Karan Vahi, Gideon Juve, Mats Rynge, Scott Callaghan, Philip J Maechling, Rajiv Mayani, Weiwei Chen, Rafael Ferreira Da Silva, and Miron Livny. Pegasus, a workflow management system for science automation. Future Generation Computer Systems, 46:17–35, 2015

    work page 2015

  13. [13]

    Jupyter notebooks-a publishing format for reproducible computational workflows

    Thomas Kluyver, Benjamin Ragan-Kelley, Fernando Pérez, Brian Granger, Matthias Bussonnier, Jonathan Frederic, Kyle Kelley, Jessica Hamrick, Jason Grout, and Sylvain Corlay. Jupyter notebooks-a publishing format for reproducible computational workflows. InPositioning and Power in Academic Publishing: Players, Agents and Agendas, pages 87–90. IOS Press, 2016

    work page 2016

  14. [14]

    Binder 2.0 - reproducible, interactive, sharable environments for science at scale

    Project Jupyter et al. Binder 2.0 - reproducible, interactive, sharable environments for science at scale. In Proceedings of the 17th Python in Science Conference, volume 113, page 120, 2018. 12 Sci-orchestraTECHNICALREPORT

    work page 2018

  15. [15]

    Portable learning environments for hands-on computational instruction: Using container- and cloud-based technology to teach data science

    Chris Holdgraf, Aaron Culich, Ariel Rokem, Fatma Deniz, Maryana Alegro, and Dani Ushizima. Portable learning environments for hands-on computational instruction: Using container- and cloud-based technology to teach data science. InPractice and Experience in Advanced Research Computing 2017 on Sustainability, Success and Impact, PEARC ’17, pages 1–9, New Y...

    work page doi:10.1145/3093338.3093370 2017

  16. [16]

    Materials data science for microstructural characterization of archaeological concrete.MRS Advancements - special issue: Materials Data Science, pages 1–14, 2020

    Daniela Ushizima, Ke Xu, and Paulo Monteiro. Materials data science for microstructural characterization of archaeological concrete.MRS Advancements - special issue: Materials Data Science, pages 1–14, 2020

    work page 2020

  17. [17]

    Workflow orchestration engines: A comprehensive survey.Journal of Cloud Computing, 10 (1):1–22, 2021

    Leandro Silva et al. Workflow orchestration engines: A comprehensive survey.Journal of Cloud Computing, 10 (1):1–22, 2021

    work page 2021

  18. [18]

    ACM SIGOPS Operating Systems Review49(1), 71–79 (2015) https://doi.org/10.1145/2723872.2723882

    Carl Boettiger. An introduction to docker for reproducible research.ACM SIGOPS Operating Systems Review, 49 (1):71–79, 2015. doi:10.1145/2723872.2723882

    work page doi:10.1145/2723872.2723882 2015

  19. [19]

    Docker: lightweight linux containers for consistent development and deployment.Linux journal, 2014(239):2, 2014

    Dirk Merkel. Docker: lightweight linux containers for consistent development and deployment.Linux journal, 2014(239):2, 2014

    work page 2014

  20. [20]

    O’Reilly Media, Inc., Sebastopol, CA, 2nd edition, 2020

    Kief Morris.Infrastructure as Code: Dynamic Systems for the Cloud Age. O’Reilly Media, Inc., Sebastopol, CA, 2nd edition, 2020

    work page 2020

  21. [21]

    Breno B. N. de Farias et al. Infrastructure as code: A systematic mapping study.ACM Computing Surveys, 55(11): 1–38, 2023. doi:10.1145/3579993

    work page doi:10.1145/3579993 2023

  22. [22]

    Richer, P

    J. Richer, P. Jones, J. Bradley, M. Machulak, and P. Hunt. OAuth 2.0 Dynamic Client Registration Protocol. RFC 7591, Internet Engineering Task Force (IETF), July 2015. URLhttps://datatracker.ietf.org/doc/html/ rfc7591

    work page 2015

  23. [23]

    Authenticating mcp oauth clients with spiffe/spire

    Christian Posta. Authenticating mcp oauth clients with spiffe/spire. LinkedIn Pulse, 2024. URL https://www.linkedin.com/pulse/ authenticating-mcp-oauth-clients-spiffe-spire-christian-posta-juvsc . Accessed: 2026- 04-23

    work page 2024

  24. [24]

    The national research platform (nrp)

    National Research Platform. The national research platform (nrp). https://nationalresearchplatform. org/, 2026. Supported by the National Science Foundation under Award No. 2112167

    work page 2026

  25. [25]

    The pacific research platform: A regional, national, and international science dmz.IEEE/ACM Innovating the Network for Data-Intensive Science (INDIS), pages 28–41, 2018

    Larry Smarr, Thomas DeFanti, Dmitry Mishin, Phil Hutton, Frank Wuerthwein, et al. The pacific research platform: A regional, national, and international science dmz.IEEE/ACM Innovating the Network for Data-Intensive Science (INDIS), pages 28–41, 2018

    work page 2018

  26. [26]

    Experiment tracking with weights and biases

    Lukas Biewald. Experiment tracking with weights and biases. https://www.wandb.com/, 2020. Software available from wandb.com

    work page 2020

  27. [27]

    Exact gaussian processes for massive datasets via non-stationary sparsity-discovering kernels.Scientific Reports, 13(1):3155, Jan 2023

    MM Noack, H Krishnan, MD Risser, and KG Reyes. Exact gaussian processes for massive datasets via non-stationary sparsity-discovering kernels.Scientific Reports, 13(1):3155, Jan 2023. ISSN 2045-2322. doi:10.1038/s41598-023-30062-8

    work page doi:10.1038/s41598-023-30062-8 2023

  28. [28]

    Noack, Petrus H

    Marcus M. Noack, Petrus H. Zwart, Daniela M. Ushizima, Masafumi Fukuto, Kevin G. Yager, Katherine C. Elbert, Christopher B. Murray, Aaron Stein, Gregory S. Doerk, Esther H. R. Tsai, Ruipeng Li, Guillaume Freychet, Mikhail Zhernenkov, Hoi-Ying N. Holman, Steven Lee, Liang Chen, Eli Rotenberg, Tobias Weber, Yannick Le Goc, Martin Boehm, Paul Steffens, Paolo...

    work page doi:10.1038/s42254-021-00345-y 2021

  29. [29]

    Foundation models for zero-shot segmentation of scientific images without ai-ready data, 2025

    Shubhabrata Mukherjee, Jack Lang, Obeen Kwon, Iryna Zenyuk, Valerie Brogden, Adam Weber, and Daniela Ushizima. Foundation models for zero-shot segmentation of scientific images without ai-ready data, 2025. URL https://arxiv.org/abs/2506.24039

    work page arXiv 2025

  30. [30]

    The state-of-the-art in container technologies: Application, or- chestration and security.Concurrency and Computation: Practice and Experience, 32(17):e5668, 2020

    Emiliano Casalicchio and Stefano Iannucci. The state-of-the-art in container technologies: Application, or- chestration and security.Concurrency and Computation: Practice and Experience, 32(17):e5668, 2020. doi:10.1002/cpe.5668

    work page doi:10.1002/cpe.5668 2020

  31. [31]

    Containers in research: Hpc and beyond.Computing in Science & Engineering, 23(2): 66–70, 2021

    Abdulrahman Azab et al. Containers in research: Hpc and beyond.Computing in Science & Engineering, 23(2): 66–70, 2021. doi:10.1109/MCSE.2021.3060938

    work page doi:10.1109/mcse.2021.3060938 2021

  32. [32]

    doi: 10.3390/jimaging11080252

    Zineb Sordo, Eric Chagnon, Zixi Hu, Jeffrey J. Donatelli, Peter Andeer, Peter S. Nico, Trent Northen, and Daniela Ushizima. Synthetic scientific image generation with vae, gan, and diffusion model architectures.Journal of Imaging, 11(8), 2025. ISSN 2313-433X. doi:10.3390/jimaging11080252. URL https://www.mdpi.com/ 2313-433X/11/8/252. 13 Sci-orchestraTECHN...

    work page doi:10.3390/jimaging11080252 2025

  33. [33]

    Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen

    Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. Lora: Low-rank adaptation of large language models. InInternational Conference on Learning Representa- tions (ICLR), 2022. URLhttps://openreview.net/forum?id=nZeVKeeFYf9

    work page 2022

  34. [34]

    Real-time person detection in wooded areas using thermal images from an aerial perspective.Sensors, 23(22), 2023

    Oscar Ramírez-Ayala, Iván González-Hernández, Sergio Salazar, Jonathan Flores, and Rogelio Lozano. Real-time person detection in wooded areas using thermal images from an aerial perspective.Sensors, 23(22), 2023. ISSN 1424-8220. doi:10.3390/s23229216. URLhttps://www.mdpi.com/1424-8220/23/22/9216

    work page doi:10.3390/s23229216 2023

  35. [35]

    Data set of thermal images of people in forested areas.Figshare, 11 2023

    Oscar Ramirez. Data set of thermal images of people in forested areas.Figshare, 11 2023. doi:10.6084/m9.figshare.24473002.v1. URL https://figshare.com/articles/dataset/Data_set_of_ thermal_images_of_people_in_forested_areas_/24473002

    work page doi:10.6084/m9.figshare.24473002.v1 2023

  36. [36]

    E. R. Keydel, S. W. Lee, and J. T. Moore. MSTAR extended operating conditions: a tutorial. InAlgorithms for Synthetic Aperture Radar Imagery III, volume 2757, pages 228–242. SPIE, 1996

    work page 1996

  37. [37]

    Review of recent advances in AI/ML using the MSTAR data

    Erik Blasch, Uttam Majumder, Edmund Zelnio, and Vincent Velten. Review of recent advances in AI/ML using the MSTAR data. InAlgorithms for Synthetic Aperture Radar Imagery XXVII, volume 11393, page 113930C. SPIE, 2020. doi:10.1117/12.2559035. SPIE Defense + Commercial Sensing, 2020, Online Only

    work page doi:10.1117/12.2559035 2020

  38. [38]

    MSTAR public dataset

    Air Force Research Laboratory and Defense Advanced Research Projects Agency. MSTAR public dataset. Sensor Data Management System (SDMS), 1997. URL https://www.sdms.afrl.af.mil/index.php? collection=mstar. Accessed via the Air Force Research Laboratory Sensor Data Management System

    work page 1997

  39. [39]

    APPFL: Advanced Privacy-Preserving Federated Learning.https://appfl.ai/, 2026

    APPFL Team. APPFL: Advanced Privacy-Preserving Federated Learning.https://appfl.ai/, 2026. Accessed: 2026-04-17

    work page 2026

  40. [40]

    Berkeley Synchrotron Infrared Structural Biology Imaging Program.https://bsisb.lbl.gov/, 2026

    Berkeley Synchrotron Infrared Structural Biology (BSISB). Berkeley Synchrotron Infrared Structural Biology Imaging Program.https://bsisb.lbl.gov/, 2026. 14

    work page 2026