Abstract: Methods and systems for processing sensor data for a plurality of vehicles are described. A first kinematic dataset is received from a sensor on a first vehicle, wherein the first kinematic dataset comprises a first location data of the first vehicle and a first ranging distance data for a first pair of vehicles comprising the first vehicle. A second kinematic dataset describing a plurality of vehicles is received. Data from the first kinematic dataset and dataset from the second kinematic dataset is fused, to generate a fused dataset, wherein the fused dataset comprises a fused location data of the first vehicle and a fused ranging distance data from the first pair.

Abstract: A genetic algorithm system generates a set of computer programs and executes a process for assessment and conditional modification of the set, repeating the process over a plurality of generations to mutate the population of solutions over time. At each generation, the system scores each program in the set to generate a respective primary score adjustment, a respective secondary score adjustment, and a respective current score. If a current score for a program is less than or equal to a first threshold, the system removes the computer program from the set. If the current score is greater than or equal to a second threshold, the system modifies the computer program to generate one or more offspring programs for use in subsequent generations. If a primary score adjustment for a program is greater than or equal to a third threshold, the system selects the computer program for performance of a task.

Abstract: In some embodiments, a method of multimodal ground truth sampling for creating synthetic multimodal training data is provided, the method performed by one or more processors, the method comprising: selecting a source object from a dataset; determining a valid pose transformation from a set of proposed pose transformations; applying the valid pose transformation to the source object to create a transformed object; generating synthetic image data based on the transformed object and a destination image; generating synthetic point cloud data based on the transformed object and a destination point cloud; and training a computer vision machine learning model from synthetic multimodal training data comprising the synthetic image data and the synthetic point cloud data.

Abstract: Systems and methods for simulating cyber-physical systems are disclosed. A plurality of geographic simulation layers representing respective infrastructure sectors of a real-world environment may be generated, and the layers may be linked together with one another to create a multi-layer simulation. The associations between the layers of the simulation may be adjusted, and characteristics of the simulation layers themselves may be adjusted, to ensure that the simulation conforms to characteristics of the real-world environment being simulated. In some embodiments, a multi-user simulation system allows users at separate terminals to execute attack inputs and defense inputs against the simulation to try to destabilize and stabilize the simulation, respectively. Results of the attack inputs and defense inputs may be simultaneously displayed on a plurality of terminals.

Abstract: A method for generating training data for a computer vision model can comprise providing an AI language model with first prompt data indicating visual scenarios to be evaluated by the computer vision model, generating, using the AI language model, based on the first prompt data and a prompting policy, second prompt data configured to cause an AI text-to-image model to generate images associated with the visual scenarios, generating the images using the second prompt data and the AI text-to-image model, applying the computer vision model to each image to generate, for each of the images, respective object detection data, and generating, for each image, performance data characterizing an effectiveness of the computer vision model, updating the prompting policy based on the performance data, and generating updated second prompt data based on the updated prompting policy.

Abstract: Methods for determining a genetic degeneracy score are described. The methods may comprise, for example, determining a window within the nucleotide sequence; determining one or more amino acids corresponding to one or more codons in the window; determining one or more degeneracy values for the one or more amino acids in the window; and combining the one or more degeneracy values in the window to determine the genetic degeneracy score. The methods may further comprise sliding the window by at least one nucleotide across the nucleotide sequence, to generate a plurality of genetic degeneracy scores. The methods may further comprise combining the plurality of genetic degeneracy scores into a final genetic degeneracy score.

Abstract: Provided are systems and methods for removing per- and polyfluoroalkyl substances (PFAS) from a contaminated stream comprising: collecting a contaminated stream comprising one or more PFAS; concentrating the one or more PFAS of the contaminated stream to achieve a concentrated stream having greater than or equal to 0.01 wt. % PFAS; and removing the one or more PFAS of the concentrated stream by heating the concentrated stream in the presence of calcium oxide to produce calcium fluoride.

Abstract: Data investigations are performed by querying a plurality of data sources. A system receives an investigation input and queries a plurality of data sources in accordance with the received input. The system receives, in response to the querying, response data from the plurality of data sources, and generates and stores a data structure representing relationships between the first investigation input and the first response data. The data structure may be in the form of a knowledge graph. The system may generate and display a visualization of the data structure. The system may generate and store a record of investigation steps used to generate the data structure, such that the investigation steps may be applied in future instances, for example using different inputs, to generate new data structures.

Abstract: A bioelectrical sensor for detecting one or more pathogens in a fluid sample is provided. The bioelectrical sensor receives a fluid sample comprising one or more pathogens, and detects the one or more pathogens using a series of chemical reactions. The series of chemical reactions include a detection step in which a detector organism detects a pathogen upon coming into contact with and/or to within a certain proximity of the pathogen, and a reporting step in which a reporter organism responds to the detection by generating an electrical signal comprising information about the detected pathogen. The electrical signal may then be transmitted to a computing device, which may identify the pathogen by mapping the generated electrical signal to a known pathogen.

Abstract: A system for automatic annotation of multimodal data comprises: an object of interest; a mobile data collection rig comprising an image sensor and a LiDAR sensor; a plurality of mobile beacons located on the object of interest and on the mobile data collection rig; one or more processors and memory. A method for automatic annotation of multimodal data comprises: receiving ultrasonic time of flight data from a plurality of mobile beacons; estimating position data for an object of interest and position data for a mobile data collection rig based on the ultrasonic time of flight data; computing pose information for the object of interest, the mobile data collection rig, an image sensor, and a LiDAR sensor; collecting image data from the image sensor and LiDAR data from the LiDAR sensor; an automatically annotating the image and LiDAR data based on the computed pose information.

Abstract: A system for evaluating trained models comprises one or more processors configured to cause the system to: receive a trained model; receive test data comprising a plurality of data objects; receive baseline classification data assigning each data object to a class; apply one or more perturbation operations to the test data to generate, for each data object, a respective plurality of perturbed data objects; apply the trained model to each perturbed data object to generate post-perturbation classification data, wherein the post-perturbation classification data indicates classification of the respective perturbed data object into at least one class and an associated confidence level of the trained model with respect to the classification; determine, for each perturbed data object, whether the post-perturbation classification data indicates a misclassification as compared to the baseline classification data; and generate and display a visualization based on the determination of whether the post-perturbation class

Abstract: Methods and systems for processing sensor data for a plurality of vehicles are described. A first kinematic dataset is received from a sensor on a first vehicle, wherein the first kinematic dataset comprises a first location data of the first vehicle and a first ranging distance data for a first pair of vehicles comprising the first vehicle. A second kinematic dataset describing a plurality of vehicles is received. Data from the first kinematic dataset and dataset from the second kinematic dataset is fused, to generate a fused dataset, wherein the fused dataset comprises a fused location data of the first vehicle and a fused ranging distance data from the first pair.

Abstract: Disclosed herein are systems and methods for optimizing air traffic control managing using a Quantum Annealing-based iterative path planning technique and algorithm that involves both classical and quantum computation components. The classical component can calculate the distances between aircraft and the target destination from a set of new, possible properties, such as aircraft location. The quantum component can select from the new, possible properties to minimize the distance of the aircraft to the target destination while ensuring adequate separation between aircraft. The algorithm can utilize qubits to represent maneuverability options for aircraft. The maneuverability options may be partitioned into a set of multiple qubits per aircraft. Each set may include a plurality of qubits that are representative of the sub options. The algorithm can utilize Quadratic Unconstrained Boolean Optimization (QUBO) to find the lowest cost-energy maneuverability option.

Abstract: Disclosed herein are systems and methods for path planning for UAVs. A set of UAVs is logically arranged into a plurality of swarms, each having a swarm-leader UAV. A first path planning algorithm is applied to determine navigation instructions to steer the swarm-leader UAVs towards respective destinations for the swarms and to deconflict the swarm-leader UAVs from one another. A set of second path planning algorithms is applied to determine navigation instructions to steer non-swarm-leader UAVs in each swarm toward their respective swarm leaders and to deconflict the UAVs from other UAVs in the swarm. Separate QUBO path planning algorithms may be used for the first path planning algorithm and the set of second path planning algorithms. If merging criteria for combining two swarms are met, a single QUBO may be used to control all non-swarm-leader UAVs in merged swarms.

Abstract: Provided are systems and methods for removing per- and polyfluoroalkyl substances (PFAS) from a contaminated stream comprising: collecting a contaminated stream comprising one or more PFAS; concentrating the one or more PFAS of the contaminated stream to achieve a concentrated stream having greater than or equal to 0.01 wt. % PFAS; and removing the one or more PFAS of the concentrated stream by heating the concentrated stream in the presence of calcium oxide to produce calcium fluoride.

Abstract: An adversarial reinforcement learning system is used to simulate a spatial environment. The system includes a simulation engine configured to simulate a spatial environment and various objects therein. The system further includes a first model configured to control objects in the simulation and a second model configured to control objects in the simulation. The first model generates a threat-mitigation input to control one or more objects in the simulation, and the second model generates a threat input to control one or more objects in the simulation. The system then executes a first portion of the simulation based at least in part of the threat mitigation input and the threat input.

Abstract: Systems and methods for password discovery are provided. A system receives a first password data set comprising known passwords and applies a rule-set to the first data set to generate a second password data set comprising passwords that are believed to be likely to be human-generated. The system trains a generative adversarial network, for generating predicted passwords, using the second data set, for example by incentivizing the GAN to favor passwords in the second data set. The system applies the generative adversarial network to generate a third password data set comprising predicted passwords. The system compares the third password data set to a data corpus to identify a string in the data corpus determined to match one of the predicted passwords in the first plurality of predicted passwords. The identified string may thus be identified as a previously undiscovered password, which may be applied to unlock password-protected systems and/or to further improve password discovery systems.

Abstract: Systems, devices, and methods for training a machine learning model and classifying encoded data include an exemplary method that includes: receiving first training data, by a variational autoencoder (VAE) comprising a probabilistic encoder and a probabilistic decoder; and encoding the first training data, by the probabilistic encoder, to generate encoded data in an embedding space; decoding the encoded data, by the probabilistic decoder, to generate decoded data; computing a loss function, comprising a hinge-loss term, based on the decoded data and the encoded data; and adjusting one or more parameters of the VAE based on the computed loss function including the hinge-loss term. The exemplary method may include classifying, by a linear support vector machine (SVM), the encoded data in the embedding space wherein classifying comprises determining the hyperplane in the embedding space separating the first class of the encoded data from the second class of the encoded data.

Abstract: Described herein are systems and methods for removing halogenated compounds from a contaminated source. These removal systems and methods include concentrating and removing halogenated compounds using select alkaline earth metal oxides, as well as acidifying and removing halogenated compounds using select acids.

Abstract: A portable kit for nucleic acid sequencing is disclosed, the kit including a DNA extraction system configured to perform a DNA extraction protocol, a DNA sequencing preparation system configured to perform a DNA sequencing preparation protocol, a sequencer system, and a portable enclosure configured to house the DNA extraction system, the sequencer preparation system, and the sequencer system.