NAML 2026

March 2–5, 2026 | NIWC Pacific in San Diego, CA

Kitware is proud to participate in the Tenth Annual Workshop on Naval Applications of Machine Learning (NAML), the premier annual event showcasing current machine learning research relevant to Naval applications.

Anthony Hoogs, Ph.D., Keith Fieldhouse, and Jason Parham, Ph.D., are attending the unclassified and restricted days, and present two restricted lightning talks on Thursday, March 5th.

Lightning Talks

Establishing Trust in Maritime Detection Models with the Natural Robustness Toolkit

Restricted Session Lightning Talk | March 5th, 2026 (~11:30 AM – Track 2)
Presenter: Keith Fieldhouse
Authors: Brandon RichardWebster, Connor Hashemi, Emily Veenhuis, Bharadwaj Ravichandran, Stephen Crowell, Connor Anderson, Brian Hu, Anthony Hoogs, Keith Fieldhouse

Deployment of artificial intelligence (AI) in operational contexts is outpacing our ability to validate reliability under real-world conditions. In many cases, collecting representative test data is prohibitively expensive; in others it may be ethically or practically impossible. During model development, data augmentation has become a standard method for improving robustness and calibration. However, comparatively little attention has been paid to the test and evaluation (T&E) phase, where the T&E engineer must verify performance claims without visibility into the model or training data.

The Natural Robustness Toolkit (NRTK), developed as part of the Chief Digital and Artificial Intelligence Office (CDAO) Joint AI Test Infrastructure Capability (JATIC) program, addresses this gap through a framework for synthetic perturbation testing that enables rigorous, repeatable robustness evaluation. NRTK enables T&E engineers to explore how computer vision models respond to operationally relevant perturbations traditional augmentation libraries often don’t provide. These perturbations allow engineers to generate controlled data sweeps, identify boundaries of reliable model performance, and visualize degradation under specific operational conditions, all without the prohibitive cost of collecting new field data.

We will demonstrate the use of NRTK with operational data to address specific operational conditions and the resulting evaluation results.

See the open source NRTK repository on GitHub for more details.

Hierarchical Pretraining for Few-Shot Learning in Underwater Acoustic Classification

Restricted Session Lightning Talk | March 5th, 2026 (~11:20 AM – Track 1)
Presenter: Jason Parham
Authors: Jason Parham, Trevor Stout, Connor Hashemi, Anthony Hoogs

The underwater acoustic environment presents unique challenges for machine learning systems, including complex propagation effects, expensive data collection, and limited labeled datasets. This presentation demonstrates how hierarchical pretraining, a technique originally developed for consumer imaging sensors, can be effectively adapted to underwater acoustic classification in the 10Hz-10kHz regime.

Our approach progressively builds knowledge by starting with out-of-domain data and sequentially training through increasingly relevant domains, enabling few-shot learning for underwater acoustic classification. We validate our approach on a challenging real-world underwater dataset for marine mammal classification.

Our results demonstrate that hierarchical pretraining leverages the abundance of freely available & unlabeled acoustic data on the internet to improve few-shot learning performance for this problem. Our findings suggest that this approach could be valuable for other fields where labeled data is scarce but unlabeled out-of-domain or adjacent-domain data is plentiful.

AI/ML at Kitware

Kitware is a leader in leveraging artificial intelligence and machine learning for computer vision. Our core technical areas include:

• AI Test, Evaluation, and Assurance
• Generative AI
• Interactive AI and Human-Machine Teaming
• Explainable & Responsible AI
• 3D Reconstruction & Photogrammetry
• Multimedia Integrity Assurance
• Activity & Threat Detection
• Object Detection, Classification, & Tracking
• Cyber-Physical Systems
• Data Annotation & Curation
• Edge Computing
• Geospatial Analytics & Remote Sensing
• Semantic Segmentation
• Super Resolution
• Event-based Imaging

Kitware’s commitment to continuous exploration and participation in other research and development areas is unwavering. We are always ready to apply our technologies and tools across all domains, from undersea to space, to meet our customers’ needs.

We recognize the value of leveraging our advanced computer vision and deep learning capabilities to support academia, commercial organizations, industry, and the DoD and intelligence communities. We work with various government agencies, such as the Defense Advanced Research Project Agency (DARPA), Air Force Research Laboratory (AFRL), the Office of Naval Research (ONR), National Oceanic and Atmospheric Administration (NOAA), the National Geospatial Intelligence Agency (NGA), U.S. Army and the U.S. Air Force. We also partner with prestigious academic institutions on government contracts.

Kitware can help you solve your challenging commercial and federal computer vision problems using our software R&D expertise. Contact our team to learn more about how we can partner with you.

Kitware is hiring!

We are a small company that has a big impact on the world. Join our team to advance science and technology, empower global innovation, and solve the world’s challenges. Kitware offers numerous benefits, including:

• Flexible work schedules
• Generous PTO and benefits package
• Employee ownership
• Positive and diverse work environment
• Support to publish your novel work
• Tuition reimbursement
• The ability to attend professional and technical conferences
• And more!

Visit our Careers page to see all of our benefits and see what it’s like to work at Kitware.