Abstract: A Holographic Transformer Network (HTN) system and method ingests and enriches semantic knowledge graph (KG) data using a holographic encoder and at least one non-GNN holographic transformer to produce graph node encodings in a node-count-mutable way. The produced node encodings can be used by a downstream network whose architecture is not tied to the static node defined by a particular vignette. The same network can be utilized for decision making even if the number of nodes in the graph (or entities in the simulation) increases.
Type:
Application
Filed:
September 24, 2025
Publication date:
March 26, 2026
Applicant:
Leidos, Inc.
Inventors:
Joshua P. Wilson, Jackson D. Scott, Christian A. Clark, Bryce E. King, Olivia Galliker d’Aliberti
Abstract: A process and system for facilitating natural language interrogation of system engineering models is described. The process and system utilize multiple formatting translations of the original SysML model into a graph schema which can be queried by a user using LLM-backed Retrieval Augmented Generation (RAG).
Abstract: A framework for generating a systems engineering model for a legacy system includes identifying and extracting code from a specific legacy system or repository and/or codebase; translating the learned code by feeding it to a LLM that specializes in code understanding to elicit natural language descriptions of the functionality of each segment of code and generate a text based output useable by a SysML generator to generate a SysML model of the legacy system from the LLM output.
Type:
Application
Filed:
December 17, 2024
Publication date:
January 22, 2026
Applicant:
Leidos, Inc.
Inventors:
Erin A. Smith Crabb, Matthew T. Jones, Laura D. Stagnaro
Abstract: An attack path modeling and discovery system and process of the present combines comprehensive network data with machine learning to determine the greatest risk to a network environment by exposing the path of least resistance an attacker would likely take. This system and process considers both the likelihood of a threat agent to exploit a vulnerability, and the potential for loss when that threat occurs. The system utilizes two key models to accomplish this goal. First, Knowledge Graphs (KG) are leveraged to comprehensively model relationships across an environment, and second, Graph Neural Networks (GNNs) are used to predict the path of least resistance to the network's user-defined most valuable assets.
Type:
Application
Filed:
November 18, 2024
Publication date:
October 16, 2025
Applicant:
Leidos, Inc.
Inventors:
Cooper Linsky, J. Marc Brasher, Todd Jaspers, Robert Lindsay Allen
Abstract: A microstructure device includes a heterogenous fiber formed of two materials having different compositions, the heterogenous fiber including a junction between the first and second materials. The first material is silicon carbide (SiC) and the second material is carbon (C). The junction may form an angle between the first and second materials.
Abstract: A methodology for bias mitigation in large language models (LLMs), leverages correlations between linguistic feature evaluations and bias benchmarks. A CL framework is used to investigate potential relationships between model behaviors and biased outcomes, providing a deeper understanding of the mechanisms underlying bias in LLMs. These insights are applied in a multi-task learning framework to demonstrate a more generalizable bias mitigation approach, achieving measurable reductions in gender and age biases with minimal trade-offs in model performance.
Type:
Application
Filed:
February 24, 2025
Publication date:
August 28, 2025
Applicant:
Leidos, Inc.
Inventors:
Laura V. Benson, Roopa Vasan, Ahmet Okutan
Abstract: This disclosure provides peptides which have a strong affinity for the checkpoint receptor “programmed death 1” (PD-1). These peptides block the interaction of PD-1 with its ligand PD-L1 as well as the interaction of CTLA4 with CD86 and can therefore be used for various therapeutic purposes, such as inhibiting the progression of a hyperproliferative disorder, including cancer; treating infectious diseases; enhancing a response to vaccination; treating sepsis; and promoting hair re-pigmentation or lightening of pigmented skin lesions.
Type:
Grant
Filed:
May 12, 2021
Date of Patent:
July 29, 2025
Assignee:
Leidos, Inc.
Inventors:
Gabriel M. Gutierrez, Vinayaka Kotraiah, Timothy W. Phares, James Pannucci
Abstract: A low power mass spectrometer assembly includes at least an ionization component, an electrostatic analyzer, a lens assembly, a magnet assembly and at least one detector located in a same plane as the entrance to the magnet assembly for detecting the deflected sample ions and/or fragments of sample ions, including ions or ion fragments indicative of the Vitamin D metabolite within the sample.
Abstract: A temperature sensor capable of taking contact method temperature measurements in extreme temperature environments includes a heterogenous ceramic fiber formed of two materials having different compositions, the heterogenous fiber including a junction between the first and second materials.
Abstract: A process and systems for constructing arbitrarily large virtual arrays using two or more collection platforms (e.g. AUX and MOV systems) having differing velocity vectors. Referred to as Motion Extended Array Synthesis (MXAS), the resultant imaging system is comprised of the collection of baselines that are created between the two collection systems as a function of time. Because of the unequal velocity vectors, the process yields a continuum of baselines over some range, which constitutes an offset imaging system (OIS) in that the baselines engendered are similar to those for a real aperture of the same size as that swept out by the relative motion, but which are offset by some (potentially very large) distance.
Abstract: This disclosure provides nucleic acids encoding peptides which bind to LAG3 and can be used to block the interaction of LAG 3 with other molecules such as MHC-II, FGL1, and ?-synuclein. These peptides can be used for various therapeutic purposes, such as inhibiting the progression of a hyperproliferative disorder, including cancer, or inhibiting the progression of a synucleinopathy, inhibiting the progression of sepsis, inhibiting the progression of an infectious disease, and enhancing a response to a vaccine.
Type:
Grant
Filed:
June 30, 2022
Date of Patent:
April 8, 2025
Assignee:
Leidos, Inc.
Inventors:
Gabriel M. Gutierrez, Vinayaka Kotraiah, Timothy W. Phares, James Pannucci, Marc Mansour
Abstract: Object detection architectures for detecting and classifying objects in an image are modified to incorporate an extending Rapid Class Augmentation (XRCA) progressive learning algorithm with its defining aspect of memory built into its optimizer which allows joint optimization over both the old and the classes using just the new class data and eliminates the issues associated with catastrophic forgetting.
Abstract: A low power mass spectrometer assembly includes at least an ionization component, an electrostatic analyzer, a lens assembly, a magnet assembly and at least one detector located in a same plane as the entrance to the magnet assembly for detecting the deflected sample ions and/or fragments of sample ions, including ions or ion fragments indicative of the Vitamin D metabolite within the sample.
Abstract: An incremental learning algorithm, extending Rapid Class Augmentation (“XRCA”), implements an unconstrained, recursive least-squares (RLS) style of optimization that incorporates knowledge of all the past training examples into each optimization step by recursively computing an IFCM in a single multi-class prediction head. The single multi-class prediction head receives class token feature vectors from a pretrained, self-supervised transformer model and is able to achieve the same optimal performance as a non-incrementally trained classifier in a jointly optimal manner over a set of increasing classes.
Abstract: Ridge Regression for Rapid Class Augmentation (R3CA), a regularized version of the XRCA incremental learning algorithm, implements an unconstrained, recursive least-squares (RLS) style of optimization that incorporates knowledge of all the past training examples into each optimization step by recursively computing an IFCM in a single multi-class prediction head. The single multi-class prediction head receives class token feature vectors from a pretrained, self-supervised, vision transformer model and is able to achieve the same optimal performance as a non-incrementally trained classifier in a jointly optimal manner over a set of increasing classes. R3CA excels at low sample incremental learning applications.
Abstract: Ridge Regression for Rapid Class Augmentation (R3CA), a regularized version of the XRCA incremental learning algorithm, is applied to large language model classification tasks such as topic classification, e.g., given a text article, determining to which predetermined topic category it should be classified, and name-entity-recognition (NER), e.g., identifying new named-entities such as a word or word phrase representing a person, organization, geographical location, art-artifact, event or nationality.
Abstract: A multihull vessel includes four main beam assemblies connected a central hull with two outer hulls and a deployment/retraction system which operates on the principle of converting transverse width of the beams into the longitudinal dimension, within the overall length of the vessel. An overall length of the multihull vessel is approximately the same when the four main beam assemblies are deployed and when four main beam assemblies are retracted.
Abstract: Object detection architectures for detecting and classifying objects in an image are modified to incorporate an extending Rapid Class Augmentation (XRCA) progressive learning algorithm with its defining aspect of memory built into its optimizer which allows joint optimization over both the old and the classes using just the new class data and eliminates the issues associated with catastrophic forgetting.
Abstract: A process for the rapid fabrication of a sorbent hollow fiber membrane (HFM) with a very high metal organic framework (MOF) content as well as the apparatus to contain such fibers for the purposes of sequestering and separating chemicals is described. Herein we developed a process to rapidly prototype meters long HFM batches with a high MOF content for sequestering and filtration. The HFM produced herein can be tailored to precisely sequester chemical of a hazardous nature which may include chemical warfare agents (CWA) or toxic industrial chemicals (TIC). The HFM are comprised of a polymer-based material that includes a polymeric binder; and one or more porous active materials that adsorb, chemisorb, decompose, or a combination thereof, a hazardous chemical.
Type:
Application
Filed:
July 15, 2024
Publication date:
January 2, 2025
Applicants:
Leidos, Inc., The United States of America as Represented by the Secretary of the Army
Inventors:
John Landers, Sergio Garibay, Trenton Tovar, Jonathan Sampson, John Mahle, Gregory Peterson
Abstract: A horizontal acoustic vector sensor system described herein includes a housing which has a gimbal assembly therein which is attached to a sensor assembly which has multiple pairs of seismometers that arranged orthogonally to one or more neighboring pairs of seismometers, along an approximately horizontal axis. The gimbal assembly with sensor assembly are enclosed within the housing by an endcap which includes an electronics assembly. The multiple pairs of seismometers are wired to the electronics assembly through a slip-ring which allows for movement of the gimbal assembly without entangling the wires. The horizontal acoustic vector sensor system further includes at least one omni-directional hydrophone integrated into the endcap.