SOLUTIONS
Novel Processor Architectures for Probabilistic Computing
Novel Processor Architectures for Probabilistic Computing
Novel Processor Architectures for Probabilistic Computing
Neoskye’s Unary Computing (UC) architecture leverages its flagship product (ML-enabled SoC), an energy efficient design in a small form-factor for critical AI/ML applications.
The product can be used for computer visions, imaging systems, video processing, telecom systems.
Keywords: Novel Processor, Unary Computing, Computer Architecture, Probabilistic Computing, Neuromorphic, Artificial Intelligence, Machine Learning, Deep Neural Network (DNN)
Advanced GPS-Based IED Detection/Clearance System
Novel Processor Architectures for Probabilistic Computing
Novel Processor Architectures for Probabilistic Computing
Neoskye’s ML-enabled hardware can analyze data from various sensors to identify different types of explosives and geo-tag their locations.
Neoskye's Integrated AI Mine Detector can find application in landmine detection, emergency response services, and other high-risk applications areas.
Keywords: Digitized Threat Map, Mine Detection, Data Fusion, Object Detection, Terrain Mapping, Machine Learning, Artificial Intelligence
Behavioristic Electromagnetic Spectrum Assessment General Learning Engine
Behavioristic Electromagnetic Spectrum Assessment General Learning Engine
Behavioristic Electromagnetic Spectrum Assessment General Learning Engine
The automated RF emitter characterization environment solution identifies and classifies RF signals within an area of interest by analyzing the emitter position, navigation, timing (PNT), and modulation.
The approach establishes a state variable emitter registration method, for which modulation features include symbol rates, deviation ratios, carrier suppression, orthogonality, duty cycle, wide-band modulation gain, frequency hopping, and periodicity.
Keywords: RF Signal, Signal Classification, Machine Learning, Artificial Intelligence, Modulation Classification, RF Signal Trending, RF Interference
Sensor Synthetic Data Generation
Behavioristic Electromagnetic Spectrum Assessment General Learning Engine
Behavioristic Electromagnetic Spectrum Assessment General Learning Engine
A Federated Generative Adversarial Networks (F-GAN) architecture based on RNN (recurrent neural network) and LSTM (long-short-term-memory).
This project incorporated algorithm development, software implementation, performance evaluation of data quality and integrity, algorithmic analysis, impacts on classification, resilience against false data injection, and adversarial learning.
Keywords: Federated Generative Adversarial Networks with Bayesian Network (FMB-GAN), Multivariate Long Short-Term Memory (MLSTM), Long Short-Term Memory (LSTM), Sensors, Machine Learning, Synthetic Data, AI/ML, Artificial Intelligence.